Official Transcript of Proceedings NUCLEAR REGULATORY COMMISSION

Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title:
Advisory Committee on Reactor Safeguards
Meeting - Open Session
Docket Number:
(none)
Location:
Rockville, Maryland
Date:
Thursday, August 21, 2008
Work Order No.:
NRC-2373
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433
Pages 1-319
1
1
UNITED STATES OF AMERICA
2
NUCLEAR REGULATORY COMMISSION
3
+ + + + +
4
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS (ACRS)
5
+ + + + +
6
SUBCOMMITTEE ON THE ESBWR COL APPLICATION
7
+ + + + +
8
MEETING
9
+ + + + +
10
THURSDAY,
11
AUGUST 21, 2008
12
+ + + + +
13
ROCKVILLE, MARYLAND
14
+ + + + +
15
OPEN SESSION
16
+ + + + +
The
17
Subcommittee
met
at
the
Nuclear
18
Regulatory Commission, Two White Flint North, Room
19
T2B3, 11545 Rockville Pike, at 8:30 a.m., DR. MICHAEL
20
L. CORRADINI, Chairman, presiding.
21
MEMBERS PRESENT:
22
MICHAEL L. CORRADINI, Chairman
23
SAID ABDEL-KHALIK
24
GEORGE APOSTOLAKIS
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
2
DENNIS BLEY
1
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
3
1
MEMBERS PRESENT (Continued):
2
CHARLES H. BROWN, JR.
3
WILLIAM SHACK
4
JOHN W. STETKAR
5
ACRS CONSULTANT PRESENT:
THOMAS S. KRESS
6
7
NRC STAFF PRESENT:
8
MARK CARUSO
9
DONALD DUBE
10
ROCKY FOSTER
11
ED FULLER
12
GARY MILLER
13
ERIC OESTERLE
14
HAROLD VANDER MOLLEN
15
ALSO PRESENT:
16
JUSTIN HOWE
17
MOHSEN KHATIB-RAHBAR
18
JONATHAN LI
19
GLEN SEEMAN
20
RICK WACHOWIAK
21
22
23
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
4
TABLE OF CONTENTS
1
2
AGENDA ITEM
PAGE
3
Opening Remarks and Objectives
4
Michael L. Corradini, ACRS
5
Presentation of Severe Accidents
6
7
8
9
10
Evaluation of Severe Accidents
130
NRO
Introduction
148
R. Wachowiak, GEH
Presentation of Selected PRA Accident
12
Sequences
13
GEH
15
6
GEH
11
14
4
Continued Presentation/Discussion
150
241
GEH
16
17
18
19
20
21
22
23
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
5
P-R-O-C-E-E-D-I-N-G-S
1
(8:34 a.m.)
2
OPENING REMARKS AND OBJECTIVES
3
CHAIRMAN
4
CORRADINI:
Let's
get
The meeting will come to order.
going
5
here.
6
meeting
on
Reactor
7
Safeguards, the Subcommittee on the ESBWR.
My name
8
is Mike Corradini, Chair of the Subcommittee.
of
the
Advisory
Committee
This is a
Subcommittee members in attendance today
9
10
are
Said
Abdel-Khalik;
11
perhaps
12
Apostolakis; and soon to be Charlie Brown.
not;
John
Bill
Shack;
John
Sieber,
Stetkar;
Dennis
Bley;
George
13
MEMBER BROWN:
Thank you.
14
CHAIRMAN CORRADINI:
And Tom Kress, a
15
consultant to the Committee.
16
meeting is to discuss the probabilistic risk analysis
17
and severe accident management strategies which form
18
the basis of chapter 19 of the SER with open items
19
associated
20
application.
with
the
ESBWR
The purpose of this
design
certification
21
The Subcommittee will hear presentations
22
by and hold discussions with representatives of the
23
NRC staff and the ESBWR applicant, General Electric
24
Hitachi Nuclear Energy, regarding these matters.
The
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
6
1
Subcommittee
will
2
relevant
3
positions and actions as appropriate for deliberation
4
by the full Committee.
issues
gather
and
facts,
information,
and
formulate
analyze
proposed
Harold Vander Mollen is the designated
5
6
federal official for this meeting.
The rules for
7
participation in today's meeting have been announced
8
as part of the notice of this meeting previously
9
published in the Federal Register on July 22nd, 2008.
10
Portions of this meeting may be closed to protect
11
information that is proprietary to General Electric
12
Hitachi Nuclear Energy and its contractors pursuant
13
to 5 USC 552(b)(C)(4).
14
And just a side note, if we're coming to
15
that, I would really appreciate if GEH tells us that
16
before we enter into it and have to backtrack.
17
we're coming to something that's proprietary, give us
18
fair warning so we can close the session.
So if
19
A transcript of the meeting is being kept
20
and will be made available as stated in the Federal
21
Register notice.
22
identify themselves and speak with sufficient clarity
23
and volume so that they can be readily heard.
24
have not received any requests from members of the
It is requested that speakers first
We
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
7
1
public to make oral statements or written comments.
So we'll proceed with the meeting.
2
And I
3
will call upon Rick Wachowiak of General Electric
4
Hitachi Nuclear Energy to lead us off.
PRESENTATION OF SEVERE ACCIDENTS
5
MR. WACHOWIAK:
6
All right.
7
everyone.
8
from General Electric Hitachi.
9
morning
Good morning,
To start with, my name is Rick Wachowiak
for
the
first
part
I've got with me this
of
the
session
10
Seeman, senior engineer in the ESBWR PRA group.
11
mainly
12
accident area.
works
on
thermal
hydraulics
in
the
Glen
He
severe
13
Just as a bit of background on this, on
14
the proprietary portion that you brought up a minute
15
ago, there are aspects of the BiMAC test report that
16
are proprietary.
17
And just so that you would all know, as
18
you brought up this morning, those things that in the
19
report are marked with double brackets and then a
20
superscript next to them, those are the proprietary
21
things.
22
associated
23
performance of the BiMAC, anything that you would
24
need to construct it essentially, then we would have
So
if
with
we
the
get
into
anything
dimensions
or
that
is
specific
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
8
1
to move to a proprietary session on that.
My slides that I have out here do not
2
3
contain any of that information.
4
on the slides is public.
CHAIRMAN CORRADINI:
5
Everything that's
So, just to make
6
sure that the members are all on the same page, so if
7
we're going to want to ask questions about that, if
8
we start going down that path, let's try to note
9
that.
And we'll close the session at the end if
10
necessary to get back to dimensional -- if there are
11
things that come down to dimensional questions.
12
MR. WACHOWIAK:
Right.
13
CHAIRMAN CORRADINI:
14
MEMBER SHACK:
Okay.
But the dimensions, for
15
example, the refractory layer, are not in brackets
16
because they are not really defined anywhere.
MR. WACHOWIAK:
17
It is probably not in
18
brackets because it's not defined yet.
19
MEMBER SHACK:
20
MR. WACHOWIAK:
21
So we can discuss that.
When it's defined, it
will be in brackets.
22
CHAIRMAN CORRADINI:
23
MR. WACHOWIAK:
24
Well said.
All right.
So where we
are with this is last I'm guessing three or four
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
9
1
months ago we had a meeting on --
2
CHAIRMAN CORRADINI:
3
MR. WACHOWIAK:
June 3rd.
June 3rd?
That's about
4
three months ago, three and a half or so.
We had a
5
meeting discussing the chapter 19 SER and draft SER
6
with open items.
7
didn't get to the severe accident area.
8
first session this morning is intended to pick up
9
where that left off.
And we ran short of time.
So we
And this
So in the severe accident area, we're
10
11
required to do two things.
12
the severe accident prevention and mitigation.
13
then, in addition, there's a section or a requirement
14
that we talk about alternatives in the design, where
15
we could reduce the risk of severe accidents.
The
16
previous
We're required to discuss
sessions
that
we
And
had
17
discussed the areas where we were recovering severe
18
accident prevent ATWS, SBO.
19
things, internal and external events.
Those are level 1 PRA
So we're not going to cover that in this
20
21
meeting.
What we're going to cover here are the
22
things that are addressed just in the severe accident
23
area,
24
coolability, high-pressure melt eject, that sort of
things
like
hydrogen
control,
debris
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
10
1
issue.
Everything that we have is contained in
2
3
one of three documents.
And, actually, there is a
4
fourth document here that I should mention since it's
5
referenced now.
6
DCD chapter 19 has the insights from the
7
severe accident analysis and a brief description of
8
what we did.
9
of that is the severe accident phenomena discussion.
10
NEDO-33306 is the SAMDA report, where we discuss the
11
cost-benefit analysis of adding additional features
12
to the design to eliminate severe accidents.
NEDO-33201 is the PRA.
And the one that I don't have on here is
13
14
the test report for the BiMAC.
15
the name of that is?
16
MR. SEEMAN:
17
MR.
18
And chapter 21
Do you remember what
I believe it's 33392.
WACHOWIAK:
Thirty-three,
three
ninety-two?
19
MR. SEEMAN:
Yes.
20
MR. WACHOWIAK:
Okay.
And I believe you
21
have that report already because the question came up
22
on what the funny markings were in there.
MEMBER
23
24
SHACK:
I
don't
think
we
have
33306, though.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
11
CHAIRMAN CORRADINI:
1
2
the
one
I
was
3
cost-benefit.
ask
about.
That's
your
WACHOWIAK:
Severe
accident
mitigation design alternatives.
CHAIRMAN CORRADINI:
6
7
to
That's
Can you repeat what that one is?
MR.
4
5
going
Yes, we do.
have that.
I don't think we
I have been looking as you were talking.
8
MR. WACHOWIAK:
That has been submitted.
9
MR. OESTERLE:
Yes.
10
I'll have to check
on that.
CHAIRMAN CORRADINI:
11
12
fine.
All
the
13
electronic form.
others
we
MR. WACHOWIAK:
14
That's fine.
have
in
That's
some
CD
or
And I'm not really going
15
to get into that a lot here.
16
as we can see through this, we anticipated a lot of
17
things that would go into the SAMDA report during the
18
design.
So
19
in
the
end,
It's in the SER.
the
answer,
are
But,
there
20
additional things that you can add to reduce severe
21
accident threats, comes out essentially to be no,
22
there
23
during the design, we pretty much got at --
are
no
other
cost-beneficial
MR. FULLER:
24
ones
because
This is Ed Fuller from the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
12
1
NRC staff.
Regarding
2
that,
particular
severe
3
accident mitigation design alternatives, it came in
4
as a response to an RAI we put out to NRC just about
5
a year ago.
6
it
7
otherwise
8
stand-alone topical report.
may
have
not
might
gotten
have
the
gotten
visibility
had
CHAIRMAN CORRADINI:
9
10
And because it was a response to an RAI,
with that later.
Okay.
in
as
a
We can deal
came
up
Okay.
is
One of the other
12
questions
13
reports, like the BiMAC report, there's a reference
14
to the severe accident treatment report, SAT report.
that
in
some
of
the
You don't have that as a separate report.
15
16
that
come
it
Go ahead.
MR. WACHOWIAK:
11
it
that
That was supplied as an RAI response.
However,
17
everything that is in the SAT report is in chapter 21
18
of NEDO-33201.
19
document
20
UCSB-formatted document.
21
is the formatting.
guidelines.
The
SAT
report
is
a
Really, the only difference
CHAIRMAN CORRADINI:
22
23
It's just reformatted to match the GE
The information is
there.
MR. WACHOWIAK:
24
The information is the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
13
1
same.
2
CHAIRMAN CORRADINI:
3
MR. WACHOWIAK:
how
we
addressed
cover
5
plant.
6
addressed in existing plants.
We
9
debris.
severe
accidents
in
this
And it's somewhat different than has been
7
affect
Thank you.
So let me quickly try to
4
8
Okay.
the
look
at
each
containment
threat
containing
for
the
what
core
can
melt
And we look for different failure modes and
10
basically
come
up
with
a
set
of
things
that
can
11
affect the containment from the different threats.
12
Then we analyze what mitigating features
13
we have or what mitigating features we need to add to
14
address those things.
15
that's fairly typical for doing a level 2 PRA.
Where
16
we
So where we depart -- and
depart,
then,
is
we
look
at
17
whether or not that particular mitigating feature can
18
be treated probablistically or not.
19
about
20
systems.
how
systems
will
perform,
We know a lot
reliability
of
That is a statistical treatment.
21
We know that we can apply failure rate
22
distributions to components and to systems in a fault
23
tree/event
24
pretty
tree
good
methodology
idea
of
how
that
will
reliability
give
those
us
a
will
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
14
1
perform.
But we have other things that are in there
2
that are just uncertain or unknown.
3
will come out of the core following a core damage
4
event is one example.
How much melt
5
There are ways we can calculate that.
6
There are different codes that show different amounts
7
of material that would exit the vessel over a given
8
time
9
parameters that control that.
period.
not
a
But
we
really
10
that's
probabilistic
11
treated probablistically.
12
those in a bounding manner.
don't
know
all
the
So we'll say that
something
that
can
be
And we'll try to treat
13
The key to that is if you try to treat
14
things that are just unknown probablistically, you
15
have the potential of diluting their impact.
16
say that 50 percent of the time there is a large
17
amount of core debris and 50 percent of the time
18
there is a small amount of core debris because we're
19
not sure which side it is in, well, then 50 percent
20
of your probability goes to the more benign side and
21
you don't know whether or not that was the important
22
side.
23
whatever
24
there's large fraction there, that could have a much
If
it
the
really
large
is
amount
80
of
percent,
core
90
If you
percent,
debris
is,
if
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
15
1
bigger impact on the results.
So in our analysis, we tried to make a
2
3
determination.
Do we know that this behaves in a
4
statistical manner?
5
We'll build a fault tree model.
6
criteria.
7
tree process.
If so, we treat it that way.
We'll set up success
We'll put it through the fault tree/event
8
If we don't know that, it's just things
9
that are associated with unknowns, we'll treat them
10
in a bounding manner.
11
bounding conditions.
So we'll assess what are the
Do we know in that example about how much
12
13
melt can come out.
Can it be more than X amount?
I
14
think in the end, what we determined for some of
15
these cases is it can't be more than 100 percent.
16
We're pretty sure that it can't be more than 100
17
percent of the melt comes out.
18
as a boundary condition.
So we would use that
So we set up boundary conditions on these
19
20
things
that
21
fragilities of the containment essentially is what
22
we're looking at there.
23
affect
24
containment as failed in that situation and try to
the
we
don't
containment?
know.
We
look
at
the
And can these bounding loads
If
so,
we'll
treat
the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
16
1
add
different
mitigative
features
2
whether or not that the chance of getting into that
3
situation is low enough or we'll add new mitigating
4
features.
5
the ESBWR.
MEMBER APOSTOLAKIS:
7
side of it?
8
more costly?
So what's the down
On the left, where you say "No," is it
MR. WACHOWIAK:
more
determine
That's kind of how we got the BiMAC into
6
9
or
costly
because
It has the potential to
10
be
we
11
associated with things that are unknown.
12
MEMBER APOSTOLAKIS:
13
MR. WACHOWIAK:
are
adding
features
And bounding.
And bounding, yes.
So we may not need that feature.
It's
14
bounding.
If this
15
phase -- you know, I'm not sure what we do about
16
that.
17
nuclear power design.
18
that you may not need but just in case.
You know, that's kind of where you end up in
So you end up adding features
19
The down side from the PRA, really, is
20
that once we have decided that we need something
21
based on a bounding analysis, it is really hard to
22
understand
23
because it has been determined a priori that it is
24
important.
what
the
importance
of
that
thing
is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
17
1
So the BiMAC if you looked at the risk
2
achievement worth in the PRA of the BiMAC, it comes
3
out to be huge because we didn't analyze what happens
4
if the BiMAC is not there.
CHAIRMAN CORRADINI:
5
Then that answers a
6
question I have been trying to dig out of all of the
7
reports.
8
that the BiMAC is not functioning and you go into a
9
condition that would look a lot like the ABWR, which
10
So you did not calculations on the branch
you essentially have core concrete attack?
MR. WACHOWIAK:
11
Right.
Now, we did have
12
a question early on "What happens if you get down
13
onto that branch?"
14
RAI, we provided an ABWR-like analysis that basically
15
it took all of the parameters that the ABWR used for
16
that
17
likelihood of a basemat breach or lower drywell wall
18
breach would be based on that.
same
And I believe in response to an
situation
and
calculated
what
the
19
Once again, that doesn't fall into this
20
category here because we don't know whether or not
21
the melt is going to behave in such a way that we can
22
just pour water on the top and, whatever it was, 90
23
percent of the time that the melt will be coolable.
MEMBER SHACK:
24
But I was looking for a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
18
1
commitment for example, to types of concrete that
2
would minimize your gas generation.
3
find that.
And I didn't
4
MR. WACHOWIAK:
Right.
5
MEMBER SHACK:
So, I mean, when you say
6
you didn't get down that path, you don't even have
7
that as sort of a defense-in-depth kind of a backup,
8
which you would have in the ABWR.
9
10
11
MR. WACHOWIAK:
Right.
MEMBER
Which
BLEY:
is
the
down
side
George was asking about.
MR. WACHOWIAK:
12
That's the down side.
13
That's the down side of doing it this way.
We added
14
a
of
15
expense, we didn't go down that path.
core
measure.
Maybe
MEMBER
16
to
absorb
APOSTOLAKIS:
some
that
I don't -I
17
questions.
18
achievement work of BiMAC, it's huge.
19
confused.
20
that you have not quantified?
21
BiMAC
have
two
You said that if you calculate the risk
I'm a bit
How do you calculate overall for something
is
there
as
a
22
measure for bounding things, right?
23
MR. SEEMAN:
24
MEMBER APOSTOLAKIS:
defense-in-depth
It has no meaning.
The second question
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
19
1
is I understand how you got on the left there how
2
you're trying to bound things, which, you know, we
3
have been doing this for decades.
But then ROAAM, the R in ROAAM stands for
4
5
risk.
So I don't understand how the methodology that
6
is risk-oriented is used in a bounding analysis.
7
there a trick there I don't see?
MR. WACHOWIAK:
8
9
way
that
I
organized
Yes.
this.
The trick is the
The
ROAAM
review
10
looking at those bounding things.
11
process is the entire page.
12
things
13
deterministic is part of the ROAAM process.
14
MEMBER APOSTOLAKIS:
15
are
probabilistic
Is
is
But the ROAAM
So it's deciding which
and
which
things
Okay.
are
So it should
have been on the title?
MR. WACHOWIAK:
16
Okay.
17
looking here at the specific --
18
MEMBER APOSTOLAKIS:
19
MR.
WACHOWIAK:
20
specific review that we did.
21
process.
But I was
You put risk and --
The
box
there
is
the
It's not the ROAAM
The ROAAM process -CHAIRMAN CORRADINI:
22
Yes.
I guess I don't want
23
to get into nitpicking, but back 25 years ago for the
24
containment loads working group, the cartoon would
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
20
1
look very similar.
2
Back in '83, when NRC staff was trying to
3
essentially understand all of this, it kind of came
4
down to, in fact, they invented some computer model
5
that I don't -- I think it was called event tree by
6
Vance Bier at Sandia, which said certain things are
7
systems.
8
criteria branch.
And you would then do essentially a success
MR. WACHOWIAK:
9
Right.
CHAIRMAN CORRADINI:
10
Certain things are
11
physical processes, which we think we know, we don't
12
really know, we have some physical feel for.
13
would take it down this path, where you would have to
14
come
15
conditions
16
accident.
up
with
and
ranges
then
of
see
initial
how
and
that
And you
boundary
evolves
the
17
So I am struggling a bit because that
18
cartoon really hasn't changed in some sense, at least
19
in my mind, for about 25 years.
20
MEMBER
21
What makes this --
APOSTOLAKIS:
Why
should
it
change, Mike?
CHAIRMAN CORRADINI:
22
I don't think it
23
should, but you said, you started this whole thing
24
off saying, this is unique.
And I'm struggling to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
21
1
think that -MR. WACHOWIAK:
2
3
unique.
4
done for level 2 in --
Okay.
It's different than what is currently being
5
CHAIRMAN CORRADINI:
6
MR. WACHOWIAK:
7
MR. WACHOWIAK:
That's
-- I wasn't around 25
years ago.
(Laughter.)
13
DR. KRESS:
14
MR. WACHOWIAK:
Neither was Mike.
Okay.
I deserve that.
I
deserve it.
CHAIRMAN CORRADINI:
16
17
That's fine.
fine.
12
15
Maybe it's unique to me
CHAIRMAN CORRADINI:
10
11
Okay.
because --
8
9
Maybe it's not
But you are saying
current operating plants won't look at it this way.
MR. WACHOWIAK:
18
Right.
And if you go
19
into the ABWR level 2 assessment, it doesn't look at
20
it this way.
21
every phenomenon.
It tries to assign split fractions to
22
CHAIRMAN CORRADINI:
23
MR. WACHOWIAK:
24
versus low?
Okay.
Okay.
And sometimes is it high
High versus low, 90 percent, 10
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
22
1
percent.
2
tries to assign a split fraction to everything.
And
3
what
the
4
distributions that way.
5
the quantification.
you
Do we just not know?
end
is
you
lose
tails
And it
of
They end up going away in
That's where they
use the expert opinion elicitation process.
8
MR. WACHOWIAK:
9
MEMBER APOSTOLAKIS:
10
the
MEMBER APOSTOLAKIS:
6
7
up
Fifty/fifty.
it.
Right.
They didn't just say
Anyway, we understand.
11
MR. WACHOWIAK:
12
MEMBER BROWN:
Okay.
Can I ask a question since
13
I am ignorant?
14
the
15
defense-in-depth the original concrete?
BiMAC
Somebody talked about why if you have
you
wouldn't
have
16
CHAIRMAN CORRADINI:
17
MEMBER BROWN:
as
additional
An analysis of it.
An analysis.
Were you
18
leaning towards if you're going to have the BiMAC,
19
you still ought to have the concrete anyway or is
20
there -MEMBER SHACK:
21
I was asking a question
22
for information at this point.
23
MEMBER
24
BROWN:
trading one for the other.
Well,
I
mean,
we
You've got concrete.
are
I
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
23
1
don't know whether everybody knows, but I presume
2
there was some analysis because being ignorant, I can
3
ask this kind of a question, what type of response we
4
would have expected back with the ABWR based on some
5
analysis or test that we did back then, 15 years ago
6
or whatever.
7
supposedly better that concrete.
8
pardon?
MR. WACHOWIAK:
9
10
And now we've got the BiMAC, which is
Yet, the report --
I was trying to think if
I had --
11
MEMBER BROWN:
Yes.
12
CHAIRMAN CORRADINI:
Well, the -Actually, he's going
13
down the path which I want you to finish this, which
14
is if you answer this to the staff, if the ESBWR
15
looked like an ABWR in performance, where there was
16
no
17
calculation of how it would perform.
BiMAC,
you
have
what
you
consider
a
bounding
18
MEMBER BROWN:
19
MR. WACHOWIAK:
You have an --
20
MEMBER BROWN:
You said you submitted to
MR. WACHOWIAK:
-- an accepted method of
21
the staff --
22
23
I lost that, Mike.
calculation of how it would perform.
MEMBER BROWN:
24
The concrete or the BiMAC?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
24
1
MR. WACHOWIAK:
2
MEMBER BROWN:
3
MR. WACHOWIAK:
4
The concrete.
Okay.
So that's --
And so since you are new
on this, the --
5
MEMBER BROWN:
Really new.
6
MR. WACHOWIAK:
-- relatively new on this
7
Committee, I don't have the slide on here.
8
used it before.
9
We have
It kind of shows where the BiMAC is.
In the ABWR, there is a block, if you will, of
10
sacrificial
11
that
12
ablated by the core.
13
being replaced with the BiMAC.
you
concrete
are
that
talking
is
about
the
low-gas
that
is
concrete
meant
to
be
And that area or that volume is
So when we did the analysis, we looked at
14
15
just
what
the
basemat
16
limestone.
17
looked at both there:
18
concrete.
would
be,
which
is
the
Well, actually, we looked at both.
We
the limestone and basaltic
19
And we can -- I don't know -- at the
20
break or maybe for tomorrow look up some of these
21
results, but --
22
MEMBER SHACK:
23
MR. WACHOWIAK:
24
this
is
that
the
That would be good.
ablation
But what I remember from
rate
with
the
low-gas
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
25
1
concrete was higher.
And we got to a basemat failure
2
faster, but less gas was generated.
3
cases, we generated enough gas to overpressurize the
4
containment requiring venting.
5
CHAIRMAN CORRADINI:
6
MR. WACHOWIAK:
7
Within 24 hours?
That's the part I don't
remember.
CHAIRMAN CORRADINI:
8
9
But in both
part I'm curious about.
That's the critical
Okay.
Thank you.
Because
10
Charlie's question, actually, is what I eventually
11
wanted to ask when you had mentioned that you had
12
submitted something to the staff.
MR. WACHOWIAK:
13
And from way back when --
14
and we'll look this up.
My recollection is that we
15
were okay for 24 hours.
It was after 24 hours that
16
you ended up needing to either vent or you would go
17
through the thickness of the basemat or the lower
18
drywell.
While that meets the regulations, you can
19
20
have
21
thought is that there is really nothing you can do
22
about it, though.
23
Okay.
24
the
intact
barrier
for
24
hours.
And
our
So then what?
So you can make it 24 hours and
then 24 in the first 25th hour, you have a problem.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
26
1
That's
maybe
you
can
do
things
with
2
sheltering/evacuation off site there, but why not go
3
through and put in something like the BiMAC that can
4
essentially eliminate the need for having to go down
5
that path?
That was what our thinking was.
6
And, as
7
such, we didn't necessarily put or we didn't actually
8
put it into the base analysis because the analysis of
9
the BiMAC basically got us to the scenario where the
10
release due to BiMAC failure was already a low enough
11
frequency that we didn't need to pursue that branch
12
further.
MEMBER
13
BROWN:
In
terms
of
the
BiMAC
14
relative to concrete, old stuff versus new stuff, the
15
reports talk about tests that you have run.
16
there are some results, which I guess are supposed to
17
be presented at some point.
Were there tests run on concrete?
18
have
a
comparison
don't
20
testing regimen in terms of the performance of old
21
stuff versus what you had.
DR. KRESS:
some
similar
So we
19
22
of
And
type
of
There is a large database on
23
tests run on concrete with melt that was done in the
24
past, mostly at Sandia but at other places.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
27
MEMBER BROWN:
1
In a manner which tried to
2
simulate the melt similar to what they talked about
3
in the report of electric heating burning everything
4
up or what have you?
5
DR. KRESS:
6
MEMBER BROWN:
7
MR. WACHOWIAK:
8
on
those
9
specifics.
older
tests,
Okay.
that.
And we are going to rely
rather
MEMBER BROWN:
10
11
Yes.
than
repeating
the
I wasn't advocating doing
I just wondered what was there.
CHAIRMAN CORRADINI:
12
We'll get back to
13
this later.
14
my mind is, is the ABWR analysis under the conditions
15
there
16
something about the BiMAC design that doesn't focus
17
the transient loading on the basemat such that you
18
could actually get to by some means, at this point by
19
some means, earlier failure because of some transient
20
effect on the BiMAC, where you focus the heat and get
21
it down to the basemat quicker?
is
I guess the only lingering question in
no
BiMAC
boundary?
MR. WACHOWIAK:
22
That
is,
is
there
In our estimation, the
23
answer to that should be no, but that's also part of
24
the detailed design, how we design that sacrificial
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
28
1
layer, which is --
2
MEMBER SHACK:
3
MR. WACHOWIAK:
TBD?
Which is TBD.
We are in
4
the process of doing that now.
We have recently
5
determined the space that we have for that layer is
6
now in -MEMBER APOSTOLAKIS:
7
8
you have a slide on the BiMAC.
9
take place then so --
Excuse me.
I see
Could that discussion
10
MR. WACHOWIAK:
11
MEMBER APOSTOLAKIS:
Your slide 7 --
12
CHAIRMAN CORRADINI:
Keep on going.
13
MEMBER
14
MR. WACHOWIAK:
--
has
layers
I was trying to remember
if I had that in there or not.
CHAIRMAN CORRADINI:
17
18
APOSTOLAKIS:
I'm sorry.
developed.
15
16
Yes, sir.
Which?
Slide 7?
Also number 8.
19
MEMBER APOSTOLAKIS:
20
MR.
WACHOWIAK:
Also number 8.
I
was
just
going
on
21
precedent from last time, where we spent 45 minutes
22
on the first slide.
23
(Laughter.)
24
CHAIRMAN CORRADINI:
At least you didn't
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
29
1
mention epistemic and aleatory, George.
2
have been on this thing for another 45 minutes.
MEMBER APOSTOLAKIS:
3
4
here.
It's all epistemic
You guys don't -CHAIRMAN CORRADINI:
5
6
We could
Well, they treat the
epistemic in one way and the aleatory in another.
7
MEMBER APOSTOLAKIS:
8
(Laughter.)
9
CHAIRMAN CORRADINI:
MR. WACHOWIAK:
10
the
specific
of
12
hydrogen generation and control.
13
have
14
overpressurization
15
noncondensible gases.
are
that
hydrogen
due
Onward.
So let me move on on some
11
there
things
Let's go.
we
have
discussed:
The issues that we
detonation
to
the
and
then
additional
16
So in terms of detonation, basically we
17
followed what 50.44 says is that you can have in
18
order to contain it.
19
during operation with nitrogen.
20
allowed oxygen that is fairly small during operation.
We
have
time
There is a band of
when
the
22
containment does not have the inert atmosphere.
And
23
that is right just prior to and just following a
24
refueling outage.
21
do
So we inert the containment
periods
So the way we treated that in the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
30
1
severe accident analysis and in the PRA was we said:
2
Okay.
There
3
during
4
performance of the containment whatsoever during that
5
time frame.
that
is
time
no
inerting
frame.
of
the
containment
We'll
not
claim
any
So it's a bomb-being scenario.
So
6
every
sequence
that
goes
to
core
7
damage in the one day prior to fuel, like a fuel
8
reload outage and one day following that, we add into
9
the containment bypass sequence.
10
Certainly it's not
that that's bound and you can't be more than bypass.
Once
11
or
again,
get
makes
it
13
importance of things is during that one day and what
14
the procedure should be because we have assumed no
15
performance
16
limitation of what we're doing.
17
get to this discussion later this afternoon, we think
18
that that is okay for design certification and to
19
determine what design features we need to add to the
20
plan.
containment.
of
So
what
to
understand
the
characterization
difficult
12
of
a
it
that's
the
a
Once again, as we
The overpressurization is another aspect
21
22
of this.
Basically what we have done there is we
23
have
24
containment.
looked
at
the
ultimate
strength
of
the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
31
There is a calculation that's presented
1
2
in chapter 19(c) of the DCD.
And it talks about what
3
the strength of the containment is.
4
And we do have an analysis that says how
5
much gas is added by reacting all of the zirconium
6
surrounding the fuel pellets.
7
percent of the reactant.
We compare those two.
8
9
It would be like 100
on the next page.
And I think that's
The pressure achieved by reacting
10
all the zirconium, even though we don't -- none of
11
our codes predict that we react that much, but all of
12
it is at the bottom end.
MEMBER
13
14
cursor?
APOSTOLAKIS:
Can
you
use
the
Yes.
15
MEMBER SHACK:
16
MR. WACHOWIAK:
Use the friendly hand.
The friendly hand.
Down
17
here it's -- I believe that's .906 was the pressure
18
that we calculate megapascal gauge.
19
rev. 5.
See, the friendly hand goes away if we
20
21
It's in DCD,
leave it.
22
CHAIRMAN CORRADINI:
23
MR. WACHOWIAK:
24
CHAIRMAN CORRADINI:
Sorry.
It's okay.
The friendly laser
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
32
1
is.
2
MEMBER BROWN:
3
MR. WACHOWIAK:
Here it's --
4
MEMBER BROWN:
Hold it.
5
CHAIRMAN
CORRADINI:
Ask
him.
Ten
atmospheres.
1.45 times 10-4 is the
MR. WACHOWIAK:
8
9
What's a mega?
I'm an English guy.
6
7
Megapascal and psig.
factor for --
10
(Laughter.)
11
CHAIRMAN CORRADINI:
Okay, Charlie?
12
MR.
One
13
CHAIRMAN
CORRADINI:
MEMBER BROWN:
Correct.
That's
A hundred and fifty psig?
It's 140 psi?
CHAIRMAN
18
19
is
right.
16
17
megapascal
approximately 10 atmospheres.
14
15
WACHOWIAK:
CORRADINI:
One
hundred
forty-five but close enough.
MEMBER BROWN:
20
14.7 psi for atmosphere,
21
right, if I remember that number?
22
though.
23
I lost it already.
So one megapascal is?
MR. WACHOWIAK:
24
I call it 147,
Tell me that again.
A hundred fifty psig.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
33
MEMBER BROWN:
1
2
roughly, megapascal.
3
this afternoon maybe.
Okay.
Okay.
I'll remember that until
MR. WACHOWIAK:
4
Ten atmospheres,
And we discussed in the
5
DCD what these various things are.
6
leakage happens to be the weak point in the steady
7
state analysis at 500 degrees.
Once again, that's probably bounding as
8
9
The drywell head
well because that drywell head is under water.
10
it's
not
necessarily
11
temperature.
going
to
that
at
that
It may be somewhat lower than that.
CHAIRMAN CORRADINI:
12
be
And
13
the blue curve?
14
chapter 21?
So can I ask about
All of these are static loads from
15
MR. WACHOWIAK:
16
CHAIRMAN
Yes.
CORRADINI:
So
there
is
an
17
analysis in chapter 21 about a dynamic load from an
18
ex-vessel steam explosion.
19
in the blue curve?
That is not encapsulated
20
MR. WACHOWIAK:
21
CHAIRMAN CORRADINI:
22
MR. WACHOWIAK:
23
and
we'll
talk
about
24
ex-vessel portion.
No.
Okay.
But, once again, this --
that
when
we
get
to
the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
34
CHAIRMAN CORRADINI:
1
2
I just wanted to
make sure what was in the blue curve.
MR. WACHOWIAK:
3
Okay.
I'm happy.
Yes.
This is the
4
static load because it's essentially looked at, what
5
happens to get to the overpressurization.
CHAIRMAN CORRADINI:
6
7
MEMBER
BROWN:
MR. WACHOWIAK:
11
MEMBER BROWN:
12
MR. WACHOWIAK:
13
MEMBER BROWN:
14
MR. WACHOWIAK:
MEMBER BLEY:
17
MEMBER BROWN:
the
pressure
The design pressure is -Forty-something?
Approximately 45.
Okay.
It's also a megapascal
About a third.
I was going to say it's
way down.
MEMBER APOSTOLAKIS:
19
20
is
number, but --
16
18
What
containment design?
10
15
Thank
you.
8
9
All right.
it's .31 megapascal, right?
The top line says
Is that what it means?
21
MR. WACHOWIAK:
Here?
22
MEMBER APOSTOLAKIS:
23
MEMBER BROWN:
24
MR. WACHOWIAK:
Yes.
That's the skin.
Yes, yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
35
MEMBER BROWN:
1
2
am
I
reading
3
overpressurizes.
MEMBER
overpressurizes.
BROWN:
that
happens
It's .31 megapascals
Anything
that
happens
Is that what --
MEMBER SHACK:
8
9
Anything
design.
6
7
wrong?
MEMBER APOSTOLAKIS:
4
5
this
Anything that happens --
No, no.
Severe accident,
design basis.
10
MEMBER BROWN:
11
accidents overpressurized.
12
MR. WACHOWIAK:
Yes.
Any of the severe
What this is saying is if
13
you overpressurize past the design, you need to get
14
to approximately three times the design before the
15
containment will start to yield.
16
won't fail until three times --
So the containment
17
MEMBER BROWN:
The design pressure.
18
MR. WACHOWIAK:
-- design pressure.
19
MEMBER APOSTOLAKIS:
20
CHAIRMAN
21
CORRADINI:
So
that
is
135
approximately.
MEMBER APOSTOLAKIS:
22
23
And why is that?
Could you explain
that to me?
MR. WACHOWIAK:
24
The design?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
36
MEMBER APOSTOLAKIS:
1
2
Why is it three and
not four?
MR. DUBE:
3
I think the x-scale is off by
4
one.
The origin should be one, trying to do the
5
division safety, the risk assessment.
MEMBER APOSTOLAKIS:
6
The origin should be
7
one, which means what, that it is zero?
8
should have gotten it into the left?
MR. WACHOWIAK:
9
Then you
These are --
I think that the scale is
10
okay.
The factor on the design pressure is less
11
important for this PRA than what the actual pressure
12
is.
MEMBER APOSTOLAKIS:
13
14
why there is a design value.
15
higher.
16
understand.
17
18
start
seeing
We go several times
a
failure.
I
don't
CHAIRMAN CORRADINI:
You start to see
MEMBER APOSTOLAKIS:
Some probability of
what?
19
20
We
I want to understand
failure.
DR. KRESS:
21
That's because the ASME codes
22
have a lot of factors of safety built into them.
23
the design conforms to the ASME code.
24
lot of factors of safety built into it.
And
There are a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
37
MEMBER
1
2
BROWN:
They
are
taking
the
factors of safety out.
3
MR. WACHOWIAK:
4
MEMBER BROWN:
Best estimate calculation.
5
MEMBER BROWN:
The design value is 45 if
6
you factor -- am I saying this right?
7
DR. KRESS:
8
MEMBER
9
Yes.
BROWN:
You
put
in
factors
I
give
you
of
safety?
10
DR. KRESS:
11
MEMBER
12
That is the way I --
Yes.
BROWN:
If
best
estimate, it's three or four times that.
MEMBER BLEY:
13
Tom, if I recall right,
14
back some years ago, Sandia tried to blow up some
15
vessels.
16
overdesigned they were.
And
that's
when
17
DR. KRESS:
18
MEMBER BLEY:
19
found
out
how
much
They actually --
You had to really pump it
way up before it -CHAIRMAN
20
21
Yes.
we
CORRADINI:
That's
a
fairly
typical result from the Sandia tests.
MR. WACHOWIAK:
22
Some of the things to get
23
your head around this to make it easier to see where
24
some of these are is that we're looking at yield here
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
38
1
at one in the design.
2
limits plus margin.
3
So that's probably where most of the origin comes
4
into play.
So that's nowhere near yield.
MEMBER
5
we
We meet everything on stress
APOSTOLAKIS:
were
told
In
here
that
the
Japanese
6
earthquake,
the
actual
7
horizontal ground acceleration was two and a half to
8
three times the safe shutdown earthquake.
9
damage was minor.
And the
It seems to be consistent with
10
this, but there is a factor of about two to three.
11
So that's good.
CHAIRMAN CORRADINI:
12
13
mechanical engineers.
back.
Thank God for those
Let's keep on going.
MR. WACHOWIAK:
14
15
That's another defense-in-depth.
I saw a hand go up in the
It went back down.
16
So in the PRA report, we don't discuss a
17
lot about hydrogen control, but there is that section
18
on the containment fragility that encompasses that.
19
That will get to the placement of the BiMAC.
20
we're
21
coolability.
looking
Some
22
at
is
of
the
ex-vessel
current
core
plants
What
debris
that
are
23
operating and some of the proposed new designs look
24
at
in-vessel
core
debris
coolability
so
that
if
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
39
1
they're raising the lower head, you can cool it from
2
the outside.
Once again, we have chosen not to take
3
4
that position for a couple of reasons.
One, the BWR
5
lower head is filled with penetrations for the CRD
6
tubes.
7
staying in place with molten debris, we would have to
8
have
9
holding those tubes in from the bottom.
And if we want to take credit for those as
some
mechanism,
fairly
robust
mechanism,
of
Due to maintenance concerns on the plant,
10
11
it is one issue.
We don't have that.
Basically
12
they're hanging from the lower head.
13
us away from being able to credit ex-vessel cooling
14
in this particular area.
So that takes
15
The other thing is, though, as we get to
16
the steam explosion part, the ESBWR is the design of
17
the lower drywell where the pedestal has essentially
18
communication to the outside.
19
a
20
containment when you get to the situation where maybe
21
the core would come through the vessel.
situation
where
So
22
we
where
it
We don't want to be in
have
a
started
partially
out
as
flooded
yes,
it's
23
going to be hard to design an outage if we have all
24
of
this
other
structural
material
in
the
lower
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
40
1
drywell, we have to maintain these CRD mechanisms, it
2
started out that way going away from the in-vessel
3
retention option.
4
would
5
implement
6
because
7
explosion versus keeping the lower head cool.
have
that
of
extremely
in
the
severe
difficult
accident
competing
CHAIRMAN
8
9
been
It turns out that it probably
to
procedure
mechanisms
CORRADINI:
actually
But
space
of
steam
that
is
no
different of a conclusion than ABWR came to, right?
MR. WACHOWIAK:
10
Yes, that is part of the
11
conclusion of ABWR.
12
in that the pedestal wall itself, the outside of the
13
pedestal wall is the suppression pool; whereas, in
14
ESBWR,
15
environment.
16
impact having been damaged.
the
of
the
pedestal
wall
is
the
So it's a little bit of a different
CORRADINI:
Thank
you.
I'd
forgotten.
MEMBER
19
20
outside
CHAIRMAN
17
18
But ABWR is slightly different
BROWN:
Can
you
explain
that
pedestal wall stuff when you get to the pictures?
21
MR. WACHOWIAK:
I can.
22
MEMBER BROWN:
I have no idea what the
23
pedestal is in here.
24
say.
I mean, there are no words that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
41
MEMBER
1
APOSTOLAKIS:
Let's
2
figure.
3
earlier in the context of the figure.
4
repeating that?
go
to
the
I would like to understand your questions
Would you mind
Are you ready to go to the figure?
5
MR. WACHOWIAK:
I'm almost there.
6
MEMBER APOSTOLAKIS:
7
MR.
WACHOWIAK:
Okay.
So,
once
again,
to
8
address this basemat melt penetration issue, we have
9
the large spreading area, just like ABWR.
It's the
10
same, essentially the same, spreading area.
We still
11
have to flood the drywell just like that system.
12
happen to have a more passive system for flooding the
13
lower drywell, but that is modeled based on a fault
14
tree.
We know how to model systems.
15
We
We have
16
the large spreading area.
Like mentioned earlier, at
17
the time we licensed the ABWR, everybody was sure
18
that the latest tests were going to show that water
19
on top of the debris was coolable.
20
quite go that way for us.
21
approach and added the BiMAC.
Well, it didn't
So we took a different
But once we add the BiMAC, though, we
22
23
have to look at failure modes in the BiMAC.
24
at
local
burnout,
water
depletion,
We look
local
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
42
1
melt-through.
And
to
address
2
addressed them different ways.
these
things,
we
The local burnout was addressed during
3
4
the confirmatory testing.
The water depletion, once
5
again that is a systematic thing.
6
fault tree model.
7
addressed in the design of the sacrificial layer,
8
which is still ongoing.
And then the local melt-through is
Now to try to get into the pictures.
9
10
MEMBER APOSTOLAKIS:
11
MR. WACHOWIAK:
12
Let me go back up to the
MEMBER
APOSTOLAKIS:
14
interesting acronym there.
15
MR. WACHOWIAK:
16
MEMBER APOSTOLAKIS:
17
MR. WACHOWIAK:
18
MEMBER
an
BiMAC?
Yes.
Okay.
APOSTOLAKIS:
MR.
WACHOWIAK:
Nobody
Basemat
remembers
internal
melt
arrest and coolability.
22
MEMBER APOSTOLAKIS:
23
MR. WACHOWIAK:
24
That's
that?
20
21
Good.
front picture here.
13
19
So we have a PCCS
I knew you knew.
The experiments in the
mat because it wasn't just panel.
It was just melt
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
43
1
arrest.
2
MEMBER APOSTOLAKIS:
3
MR. WACHOWIAK:
Okay.
So, just to point out, as
4
I was saying earlier, this is the area that we're
5
looking at here, is the lower drywell pedestal area,
6
reactor vessels up above that.
7
basemat.
8
the reactor building.
It's sitting on the
And these rooms on the outside here are in
These are the pedestal walls that we're
9
10
talking about here.
The floor is the basemat of the
11
reactor.
12
these rooms here are outside of the containment.
13
the
14
buildings.
This is all underground, by the way.
pedestal
wall
is
up
against
CHAIRMAN CORRADINI:
15
the
And
So
reactor
In difference to the
16
ABWR, where we would have found the suppression pool
17
out there?
MR. WACHOWIAK:
18
Our suppression pool is
19
raised so that it's above the core in this design,
20
rather than down on the basemat.
21
MEMBER APOSTOLAKIS:
22
MR. WACHOWIAK:
Very good.
So when we go back to
23
this picture, these walls are the pedestal walls I
24
was just talking about.
This floor here should be
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
44
1
extended
2
basemat.
across
and
throughout
CHAIRMAN CORRADINI:
3
here.
It's
the
And then, as we were
4
talking about, would the sacrificial layer be still
5
TBD relative to dimension?
6
been scoped as to what it's going to be in terms of
7
penetration, depth, thickness?
MR.
8
9
established.
WACHOWIAK:
Has the basemat floor
Yes.
That's
already
Do you remember what it was?
10
MR. SEEMAN:
I believe it was six meters.
11
MR. WACHOWIAK:
Six meters?
And that was
12
already established in the design basically for the
13
structural portion of the building.
14
allowed
15
approximately one and a half meters above the --
to
play
in
was
this
CHAIRMAN CORRADINI:
16
What we were
area
here
that's
And the material for
17
the basemat is specified or flexible relative to the
18
concrete type?
19
MR. WACHOWIAK:
20
CHAIRMAN CORRADINI:
Okay.
21
MEMBER APOSTOLAKIS:
Say that again.
22
MR. WACHOWIAK:
23
CHAIRMAN
24
I believe it's flexible.
In the design --
CORRADINI:
The
type
of
concrete, there's various --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
45
1
MEMBER BROWN:
2
CHAIRMAN
3
Yes, the gray stuff.
CORRADINI:
MR. WACHOWIAK:
4
5
You mean the gray stuff?
The
gray
stuff.
Yes because of the gray
marble out here.
MEMBER
6
BROWN:
Not
the
BiMAC
7
itself, the device?
That's a basemat.
8
stuff is the protector for the basemat, right?
MR. WACHOWIAK:
9
some of the things are here.
11
that
12
containment.
13
distributor.
way
up
here.
The other
So let me go over what
10
are
thing
We have water tanks
They're
way
up
in
the
And we have pipes that go down to a
14
And then the BiMAC itself -- and I think
15
I have this on the next sheet -- is a two-dimensional
16
cone.
17
circular area.
So
it's
like
this
except
And those pipes go out.
18
to
standpipe.
a
And then they
transition
20
course,
21
diameter.
22
near the edge.
23
edge, some of the horizontal do need more than one
24
standpipe in order to completely cover the wall.
much
lower
as
The
covering
19
are
a
it's
you
go
standpipes,
out
along
of
the
And the standpipes are higher as you get
And, in fact, as you get near the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
46
MEMBER BROWN:
1
Why do you want all of the
2
core to pool into one little puddle, as opposed to
3
being more distributed?
4
the old days.
MR. WACHOWIAK:
5
6
Okay.
This is somewhat
exaggerated in terms of the inclination.
CHAIRMAN CORRADINI:
7
8
That was not a good idea in
He can't speak about
it.
9
MR. WACHOWIAK:
Which?
10
MEMBER BROWN:
You can calculate it on
11
the next page if you want to.
MR. WACHOWIAK:
12
It's showing what our
13
concept was.
14
or
15
thermal dynamic behavior is in the range of zero to
16
20 degrees.
17
up our concept based on.
that
the
The experiments show that the optimum
range
of
angle
for
having
the
right
Approximately ten degrees is what we set
Is that on the next page?
18
MEMBER BROWN:
19
MR. WACHOWIAK:
preliminary.
And,
Yes.
Once again, all values
20
are
21
different
22
specifically are.
23
And it's 100 square meter floor area.
now.
We
in
won't
fact,
get
they
into
are
what
all
they
So it's a very low, low angle.
This shows where that the layer basically
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
47
1
hugs the pipes.
In fact, that was in the original
2
concept, we thought that we would do this this way.
3
In fact, we will be filling in most of that area with
4
the concrete.
5
what the right final dimension is.
And we'll let the core itself decide
CHAIRMAN CORRADINI:
6
7
you just said, Rick?
8
sorry.
I didn't understand it.
MR. WACHOWIAK:
9
Can you repeat what
I'm
This layer initially will
10
probably not be set up so that it's in a dish sort of
11
arrangement.
12
floor.
MEMBER BROWN:
13
14
It will probably be just like a flat
will not be ten degrees.
That degree, ten degrees,
It's the pipes.
15
MR. WACHOWIAK:
16
pipes underneath the floor.
The ten degrees is the
17
MEMBER BROWN:
18
MR. WACHOWIAK:
But the floor itself will
MEMBER BROWN:
That's the brown stuff on
19
seem --
20
21
Okay.
top of the pipes.
MR. WACHOWIAK:
22
And the actual sloping of
23
the floor will be based on getting water into the
24
sump, just like in a normal lower drywell.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
48
MEMBER
1
2
APOSTOLAKIS:
So
that
will
be
another -MR. WACHOWIAK:
3
We have channels in the
4
floor to get water into the sumps and things like
5
that.
6
need to have that slope.
But for BiMAC performance itself, we don't
So your question, why do we want to pull
7
8
it to the middle?
We don't.
This was an initial
9
concept of how it might be arranged, but it is likely
10
to be just a flat floor underneath there.
11
pipes need to be sloped because the pipes, we need to
12
establish the flow patterns that -MEMBER BROWN:
13
But the
So you are saying that
14
brown layer is going to be thicker in the middle and
15
thin around the edges?
16
MR. WACHOWIAK:
17
MEMBER BROWN:
18
MR. WACHOWIAK:
19
thin around the edges.
20
of putting it right now.
MR. WACHOWIAK:
walk on it.
Thicker in the middle,
That's probably the best way
grated thing on top of that?
23
24
I put it in simplified --
MEMBER BROWN:
21
22
Yes.
And what is the little
Is that a -It's a floor so we can
We're not expecting to --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
49
MEMBER BROWN:
1
2
Is it an open floor?
I
mean, that's -MR. WACHOWIAK:
3
at
4
It's
5
probably be grating with some kind of a sheet metal
6
on
7
there.
it
least
so
that
to
you're
be
not
grating,
dropping
but
it
stuff
will
through
But, once again, that has nothing to do
8
9
going
We haven't decided yet.
with the performance of the BiMAC.
MEMBER BROWN:
10
11
wanted to know what it was.
12
MR.
I understand that.
That's all.
WACHOWIAK:
know,
And
we
specific
materials
here,
14
zirconium
refractory
15
Actually, it's now based on what we have looked at in
16
the experiments.
17
choice of things that we can use.
And we're working
18
on what that specific material is.
It's not been --
material
had
the
13
in
sorry.
presented
earlier
this
designs.
We find that we have a much greater
CHAIRMAN CORRADINI:
19
20
you
I just
Based on what?
I'm
You said based on what?
MR.
21
WACHOWIAK:
After
we
finished
22
optimizing the dimensions here, what we found is that
23
we
24
dimensions that we'll finally end up with on the
have
greater
choices
of
materials
given
the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
50
1
piping.
2
CHAIRMAN CORRADINI:
3
MEMBER APOSTOLAKIS:
4
that later?
They are going to
MEMBER APOSTOLAKIS:
Oh.
You will get
Okay.
MR. WACHOWIAK:
9
10
CHAIRMAN CORRADINI:
get to that.
7
8
So what kinds of
tests have you performed?
5
6
Thank you.
Yes.
I'll get you there.
Just to illustrate here -MEMBER SHACK:
11
Before you get there, let
12
me just ask the question about -- you know, you're
13
very concerned about ablation of that layer, but the
14
thermal shock itself never seems to be discussed as
15
to
16
testing of that.
whether
it
will
spall
and
crack.
There's
no
And, again, I'm not familiar with the
17
18
literature
on
this.
19
particular
mode
20
discussed anywhere?
of
CHAIRMAN
21
Could
failure,
you
which
CORRADINI:
address
isn't
Can
I
that
really
append
22
something to this?
In chapter 21 and in the BiMAC
23
experimental report, it makes great pains of saying
24
the best estimate is that metallic will come down
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
51
1
first,
2
oxidic.
but
it
was
bounded
100
percent
of
the
Metallic doesn't behave the same way in
3
4
thermal shock as oxidic.
5
the concern here.
So that's what I think is
MR. WACHOWIAK:
6
7
by
Do we have an RAI on that
specific thing or is that --
8
MR. SEEMAN:
That is on the RAI.
9
MR. WACHOWIAK:
We have an RAI on that.
10
And we basically got that about a week ago, and we
11
haven't
12
responding to that in the RAI.
responded
to
that
yet.
CHAIRMAN CORRADINI:
13
But
we'll
be
So if we are not
14
going to discuss anything further here, just to go
15
one step further, this is where I guess I am asking
16
the question about the ABWR calculation being a bound
17
to the BiMAC performance because if this is designed
18
in such a fashion that I spall, I get right to the
19
tubes.
20
concentrated focus and a totally different erosion
21
pattern than if I had it as the ABWR expected it to
22
be nicely spread out.
23
the question.
I burn through the tubes.
MR.
24
Now I've got a
That's what's got me asking
WACHOWIAK:
Okay.
There
are
two
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
52
1
things associated with that.
The first one is that
2
we show these things here.
And they're somewhat
3
misleading in your mind.
4
--
It's more of a flat floor
5
CHAIRMAN CORRADINI:
6
MR.
WACHOWIAK:
Yes.
--
That I got.
than
what
you're
7
thinking.
The calculation that we did that compared
8
to the ABWR took into account the shape of the pipes
9
so that the volume of the melt, if you will, had that
10
characteristic.
The
11
associated with the double-inverted ungulate that we
12
have here.
DR. KRESS:
13
thickness
of
the
melt
was
I have a conflict of interest
14
on this, but I would like to ask a question of fact.
15
On your picture, the line on the left is identified
16
as a deluge line.
17
MR. WACHOWIAK:
18
DR. KRESS:
19
feed
line
20
separate?
the
Is that incorrect?
BiMAC
MR. WACHOWIAK:
21
22
to
Yes.
pipes.
No.
The
That's a
deluge
is
We called them all
deluge.
23
DR. KRESS:
Called them all deluge?
24
MR. WACHOWIAK:
Yes.
And some of them
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
53
1
will go directly into the BiMAC channel.
2
will go into the lower drywell area -DR. KRESS:
3
4
right.
You call them all deluge,
Yes.
MR. WACHOWIAK:
5
And some
-- directly into the sump
6
because we also want the sump filled with water to
7
protect in case something gets spilled over into the
8
sump so we don't have -DR.
9
KRESS:
There
10
lines that feed the pipes?
11
MR. WACHOWIAK:
12
DR. KRESS:
13
MR. WACHOWIAK:
14
because we have four --
15
DR. KRESS:
16
are
three
of
these
Twelve lines.
Twelve lines feed?
There are 12 lines total
Six feed the pipes, and six
go directly on top?
MR.
17
WACHOWIAK:
Essentially.
And
we
18
could change that if we -- that's not cast in stone
19
yet of how many go to which places, but what we
20
initially decided was that we needed about half to go
21
directly to the BiMAC.
22
protecting, you know, just spilling onto other things
23
that we want protected.
24
DR. KRESS:
And the other half can go to
Well, lines that feed the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
54
1
BiMAC tubes, do they go to the center distributor?
2
You have a distributor running in that direction.
3
MR. WACHOWIAK:
4
CHAIRMAN CORRADINI:
5
MEMBER BROWN:
Do you have a slide that
shows that?
CHAIRMAN CORRADINI:
8
9
I think it's the
next slide.
6
7
Do you have that?
question that I had, too.
DR. KRESS:
10
Tom is getting to a
I was -This shows two of them, two
11
downcomers, one on each end.
12
six.
You said there were
I was trying to figure out how they --
13
MR. WACHOWIAK:
14
long-term feed for natural circulation.
15
pool is above the core.
16
these tubes to the distributor and out.
DR. KRESS:
17
18
Long-term, these are the
So the water
Water would be going into
I understood you to say there
are six of those lines.
19
MEMBER APOSTOLAKIS:
20
MR. WACHOWIAK:
Twelve.
And then the concept here
21
on this cartoon, right?
22
here but a cartoon; that these deluge lines would be
23
fed
24
smaller lines would go down there initially.
down
to
this
This isn't a design drawing
distributor.
So
six
of
those
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
55
DR. KRESS:
1
2
the distributor?
MR. WACHOWIAK:
3
4
Oh, the blue is also feeding
The blue is the one that
comes from the pools up on the top.
CHAIRMAN CORRADINI:
5
6
get into really colors.
7
circulation?
MR. WACHOWIAK:
9
CHAIRMAN CORRADINI:
Yes.
The dark blue, there
are 12 or 6 of them?
MR. WACHOWIAK:
11
12
So let's now
So the yellow is for natural
8
10
Okay.
That would go into here?
We're estimating that should be about six of them.
CHAIRMAN
13
CORRADINI:
And
there
are
14
another six that would discharge on top of the melt
15
or on top?
MR. WACHOWIAK:
16
sump
or
things
that
On top of the melt or in
17
the
we
think
need
to
have
18
additional protection, like the sump, is a way that
19
if you get material in there, you could focus the
20
heat onto the pedestal wall and down through the
21
basemat.
22
So what we want to do is we want to make
23
sure that the sump is filled so that the BiMAC pipe
24
is here and then the sump behind the pipe is also
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
56
1
filled with water so that there is more protection
2
from getting material there.
If there are other things down in here
3
4
that
we
5
protection, which we don't know right now since we
6
don't know all of what is down there, we could direct
7
that deluge onto those specific things so that it is
8
a water-covered thing when the melt is coming out
9
versus just dry and good impact, --
to
MR. WACHOWIAK:
have
some
short-term
So I don't know how
-- flexibility that we
have.
CHAIRMAN
14
15
want
you want to --
12
13
we
CHAIRMAN CORRADINI:
10
11
think
CORRADINI:
There's a question over here.
That's
I'm sorry, George.
MEMBER APOSTOLAKIS:
16
fine.
Yes.
I am still
17
trying to understand the answer to Tom's question.
18
So there is this dark blue line from the GDCS.
19
MR. WACHOWIAK:
20
MEMBER APOSTOLAKIS:
21
blue lines do you have?
MR.
22
23
GDCS.
GDCS.
How many dark
Six?
WACHOWIAK:
Associated
with
this
piece here, it will likely be six.
MEMBER APOSTOLAKIS:
24
So where would the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
57
1
second one be?
2
MR. WACHOWIAK:
3
MEMBER APOSTOLAKIS:
4
the second one would be.
It could be seven.
I want to know what
I am missing something.
5
MR. WACHOWIAK:
They would just be --
6
MEMBER APOSTOLAKIS:
7
MR. WACHOWIAK:
8
MEMBER APOSTOLAKIS:
9
MR. WACHOWIAK:
Around.
-- just around.
Okay.
Around.
Like the ABWR, they just
10
come down the lower drywell, and they are open pipes
11
to here.
So it would be a setup similar to that.
MEMBER APOSTOLAKIS:
12
13
next
14
distributor and so on.
then,
each
MEMBER BROWN:
15
16
slide,
said.
one
will
No.
It's an open line he
MEMBER APOSTOLAKIS:
18
MR. WACHOWIAK:
19
MEMBER APOSTOLAKIS:
20
MR.
its
own
Oh, okay.
Into the distributor.
All right.
WACHOWIAK:
Into
the
common
distributor for all of the pipes.
22
MEMBER APOSTOLAKIS:
23
DR. KRESS:
24
have
It just dumps water down into the --
17
21
And if we go to the
thing
progresses,
Okay.
Okay.
Now, in the long term as this
that
khaki-colored
stuff
will
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
58
1
disappear?
MR. WACHOWIAK:
2
3
The khaki-colored stuff
will disappear.
4
CHAIRMAN CORRADINI:
5
DR.
6
It
won't
be
part
of
the
will
come
to
an
will
come
to
an
operation?
MEMBER
7
8
KRESS:
Sacrificial layer.
SHACK:
It
equilibrium.
MR.
9
WACHOWIAK:
It
10
equilibrium.
There will be some layer and some crust
11
and then material above it, core material above it.
12
So this where it may start to be a floor will end up
13
in some shape that is determined by the heat transfer
14
characteristics of the melt.
So
15
the
melt
will
bring
that
into
16
equilibrium.
17
enough so that the equilibrium doesn't fall into the
18
blue pipe.
DR. KRESS:
19
20
The key is that it needs to be thick
And you have determined what
that crust thickness will be?
MR. WACHOWIAK:
21
Yes.
And that was what
22
was on the previous page that as long as we have 20
23
centimeters, we're going to be protected.
24
more of a minimum value here.
That's
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
59
MEMBER BROWN:
1
2
That's
the way it's built.
MEMBER SHACK:
3
4
That's a start.
Yes.
That is not your
equilibrium value.
MR.
5
WACHOWIAK:
That
would
be
the
6
starting value, but, in all likelihood, it's going to
7
be
8
centimeters of using zirconium oxide.
different
than
this
because
CHAIRMAN CORRADINI:
9
this
was
20
So let me get back
10
to my question again, which maybe is another RAI.
11
This
12
ex-vessel.
13
concentrated
14
that's to me the crucial thing.
15
based on what the melt composition is at that turn
16
point.
is
essentially
Your
at
just
heat
that
turn
like
the
AP1000
flux
is
going
point,
17
MR. WACHOWIAK:
18
CHAIRMAN CORRADINI:
right?
except
to
be
And
so
And that's crucial
Yes.
And so I'm still
19
back to the original statement in the BiMAC report
20
and in chapter 21 which says things are bounded by
21
100 percent of the melt oxidic, but the guesstimate,
22
the best estimate is metallic melt coming down.
23
And so if I have metallic melt coming
24
down, even particularly that I've now got a flat
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
60
1
floor,
and
I
direct
that
metallic
2
corner, I'm going to have one interestingly high heat
3
flux there.
And I'm curious.
4
melt
to
that
You've gotten to the
5
curve that I am interested in, which is that red line
6
is based on some theoretical calculation.
7
line is the test results, which says we're hunky-dory
8
because we're underneath the red line.
I'm
9
10
struggling
MR. WACHOWIAK:
Okay.
how
much
CHAIRMAN CORRADINI:
14
MR. WACHOWIAK:
The blue line was
Oh, I'm sorry.
I believe the test was
looking at that.
CHAIRMAN CORRADINI:
16
Right, knuckle or
17
whatever you want to call this thing.
18
should remember this, and I don't.
MR. WACHOWIAK:
19
I'm sorry.
I
I apologize.
Yes.
Well, maybe you
20
should.
21
like to use, all this non-dimensional stuff.
This is one of those things that you guys
22
The blue line is the pipe that is --
23
MEMBER APOSTOLAKIS:
24
I
done based on the initial calculations, not the test.
13
15
with
trust the red line.
11
12
still
The blue
"You guys."
Which
guys are we?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
61
1
(Laughter.)
2
MR. WACHOWIAK:
3
4
Academia.
I know what he's saying.
If we told the construction guys to go
out and build this non-dimensional thing --
5
(Laughter.)
6
MR. WACHOWIAK:
So the black line is a
7
pipe that is essentially all horizontal.
8
not much to the vertical section of it.
9
matter of fact, we think that the melt is probably
10
not even going to get to the vertical section of that
11
piece of pipe because it's out on the diameters.
12
know, the BiMAC itself, the volume can hold, what,
13
three or four cores.
14
go all the way out to the sides.
15
CHAIRMAN CORRADINI:
16
There is
And, as a
You
So we don't think it's going to
That's if it's a
cone.
MR. WACHOWIAK:
17
Right.
So this one here,
18
there's nothing at that knuckle.
19
cones at the ends near where the entry to the channel
20
is, that's what this channel is meant to represent.
21
And
22
vertical
23
there would be this issue that you are talking about.
this
is
the
piece
horizontal
and
the
piece.
knuckle
CHAIRMAN CORRADINI:
24
But this one if it
right
This
is
there
I'll stop.
the
where
But what
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
62
1
I'm worried about is that pinch point and that you're
2
going to get enough boiling to essentially shut the
3
natural circulation down flow.
4
will essentially then choke off, right, just stop
5
flow because if I have any sort of CHF or enough
6
boiling, I'll essentially stop any sort of down flow
7
because now I'm bringing the pressure drop.
And the whole thing
8
The two-phase pressure drop is moving,
9
getting larger and larger and moving back up the pipe
10
so that I don't have enough natural circulation head
11
to drive it.
MR. WACHOWIAK:
12
13
And I think that was what
the experiment was meant to investigate.
14
CHAIRMAN CORRADINI:
15
DR.
KRESS:
MR.
WACHOWIAK:
16
DR. KRESS:
line
Yes,
I
is
a
CFD
think
that's
Do you know what code they
used for that?
21
MR. WACHOWIAK:
22
looked up before I came on.
23
DR.
24
blue
right.
19
20
Okay.
calculation?
17
18
The
Yes.
KRESS:
It wasn't something I
Two-dimensional
or
three-dimensional?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
63
MR. WACHOWIAK:
1
We did two-dimensional
2
and three-dimensional cases.
3
the three-dimensional results.
4
chapter 21 that looks at different things.
There's a
5
batch
with
6
three-dimensional code versus the two-dimensional.
of
those
cases
I think the M cases are
that
There's a table in
were
7
DR. KRESS:
8
you used to calculate the red line?
10
don't know those.
11
21 when we get to a break.
13
I
So obviously two different
correlations.
14
MR. WACHOWIAK:
15
DR. KRESS:
16
I can look those up.
We can maybe look inside chapter
DR. KRESS:
12
the
Do you know what correlations
MR. WACHOWIAK:
9
done
Yes.
One for a flat plate and one
for vertical.
17
MR. WACHOWIAK:
18
inclined plate and one for a vertical.
CHAIRMAN
19
Well, one for a minimally
CORRADINI:
But
where
Tom
I
20
thought was going is there is a fuzziness to the blue
21
line and there is a fuzziness to the redline.
22
worried
23
That's where I'm still troubled at.
24
what.
about
the
intersection
of
the
I'm
fuzziness.
I assume that's
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
64
MR. WACHOWIAK:
1
2
experiments
3
know, clearly this is where we have the least amount
4
of market, is right at that fuzzy intersection.
supposed
DR. KRESS:
5
6
were
And that was what the
to
address
because,
you
This is all nucleate boiling.
None of it is film boiling, of course.
MR. WACHOWIAK:
7
It's essentially nucleate
8
boiling, especially on the vertical parts.
9
some.
There are
We looked at some SLUG flow in --
10
DR. KRESS:
11
MR.
Lot of void --
WACHOWIAK:
--
the
horizontal
12
sections, but, once again, we didn't get to there as
13
long as we had subcooling at the inlet.
14
go
15
regime and still perform this way.
16
DR. KRESS:
17
MEMBER ABDEL-KHALIK:
somewhat
past
how
nucleate
into
the
SLUG
Okay.
18
different,
19
distribution in that header?
would
that
MR. WACHOWIAK:
20
boiling
But it could
If the tubes are
affect
the
flow
That was another one of
21
the objectives of the experiment, to look at an array
22
of tubes.
23
quarter BiMAC, if you will.
24
heat
flux
So one of the scaled experiments had a
on
heat
flux
And we looked at local
differences
across
the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
65
1
different tubes.
2
that.
And we'd have results presented for
3
And I believe we have an RAI also on some
4
of those results, that some of the tubes had reverse
5
flow, I think is what you saw in the bore head.
6
that right?
So
7
8
asymmetric
9
patterns
10
that
issues,
that
because
that
were
there
of
some
were
established,
of
these
different
and
we're
Is
flow
looking
into that right now.
DR. KRESS:
11
flow
Was there any concern about
12
parallel
stability
when
13
parallel pipes with different heat fluxes on them?
14
MR. WACHOWIAK:
15
CHAIRMAN CORRADINI:
you
have
lots
of
Parallel flow stability.
You got some.
What
16
you were saying, that you had set up a situation
17
where certain pipes were going this way.
18
this almost like an oscillatory.
MR. WACHOWIAK:
19
Yes.
You get
And I think that's
20
what some of the questions that we have -- there were
21
some
22
experiments.
23
those.
reverse
flow
And
we
DR. KRESS:
24
situations
have
been
shown
asked
in
to
the
explain
I was more concerned about
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
66
1
some
flows.
You
have
the
same
pressure
drop
2
basically across these two.
3
at pressured off if you're in film boiling. Some of
4
them will give you that pressured off if you're in
5
nucleate boiling but a different flow rate.
Some flows would get it
6
So that sometimes give you -- you never
7
know when you're going to be in one or the other
8
because it's a probabilistic thing that decides.
MR. WACHOWIAK:
9
DR. KRESS:
10
11
It's a good question.
But you didn't see any of
that in the test at all?
12
MR. WACHOWIAK:
13
DR. KRESS:
We didn't see that.
Not the question.
Were each
14
of the tubes instrumented so you could get the flow
15
rate through each of the tubes?
MR. WACHOWIAK:
16
17
I believe that that is
the case.
CHAIRMAN
18
CORRADINI:
That
was
true.
19
There is a picture of this magnetic flow meter where
20
they are tracking flow rate per tube.
21
DR. KRESS:
Per tube?
22
CHAIRMAN CORRADINI:
23
DR. KRESS:
24
CHAIRMAN CORRADINI:
Yes.
And void fraction?
I think so because
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
67
1
there was a void measurement.
2
void measurements.
3
DR. KRESS:
4
CHAIRMAN CORRADINI:
5
DR. KRESS:
6
Oh, magnetic flow meters -Right.
-- are weird when you have
void fractions.
MEMBER ABDEL-KHALIK:
7
8
There were integral
I believe Tom is
asking about leatherneck-type flow and stability.
DR. KRESS:
9
Yes, absolutely.
MEMBER ABDEL-KHALIK:
10
you
know,
11
is,
12
greater than the onset of significant voids, which is
13
about
14
region where we would be concerned about OFI-type
15
problems in parallel pipe systems.
16
looked at?
five
we're
percent
talking
And that question
void
MR. WACHOWIAK:
17
about
fraction.
void
And
fractions
that's
a
Has that been
Where I think we -- I
18
don't know the answer to that specific question.
It
19
hasn't been posed yet.
20
find is that all of our initial calculations that we
21
did assumed that there would be saturated water at
22
the inlet to the downcomer.
23
are probably going to have some subcooling, four or
24
five degrees of subcooling, at the inlet, almost only
One of the things that we did
And it turns out that we
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
68
1
because of the height of the water pool that is going
2
to be there.
3
Many of these things that we were worried
4
about go away when we have the initial subcooling.
5
We maintain almost a nucleate regime all the way
6
through the pipe when we have the initial subcooling.
7
So if that specific question that you
8
have can be found in one of the questions that we
9
already have or somehow if you can get that to the
10
staff
and
send
that
11
question
12
prepared to discuss.
specifically.
to
us,
we
But
it's
MEMBER ABDEL-KHALIK:
13
be
an
answer
beyond
Okay.
problem
what
that
I'm
I think that
14
would
15
Regardless of how much subcooling you have at the
16
inlet, you have subcooled boiling.
17
voiding you have in the piping is probably greater
18
than OSV, which is five percent void fraction.
19
typically OFI happens at the same point that the OSV.
to
look
at.
The amount of
And
So five percent void fraction I would
20
21
start
22
piping systems.
worrying
about
flow
MR. WACHOWIAK:
23
24
interesting
could
instability
Okay.
in
parallel
And we have it on
the heat flux range there, but what Dr. Theophanus
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
69
1
was saying, where he thought he was going to get into
2
a
3
horizontal tubes now, is about a 70 percent of what
4
fraction is where he would get backpressure issues
5
now.
6
question, and I don't know that we answered that
7
question.
problem,
on
the
vertical
or
in
the
This question about stability is a different
If you failed one of the
tubes, would you be injecting water into the melt?
CHAIRMAN CORRADINI:
10
11
is
DR. KRESS:
8
9
this
We would call it the
Con. Ed. approach.
12
DR. KRESS:
13
MR.
That's where it comes from?
WACHOWIAK:
I
don't
know
that
we
14
expect the melt to get all the way down to the tubes.
15
We still expect there to be some layer of material
16
-DR.
17
18
KRESS:
Some
crust
and
material
there?
MR. WACHOWIAK:
19
-- in crust in between
20
that.
So we wouldn't be looking at injecting water
21
directly into the melt.
22
works.
23
one tube, would we be able to get a local ablation
24
right at that one point to drop down in the tube?
Now, that's if everything
But if a tube dries out, you know, if it's
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
70
1
That seems intuitively hard for me to
2
come to the point where if we failed one tube, that
3
we would get instant or a direct flow or a direct
4
melt path to the basemat.
5
could see that.
If we failed the tube, I don't think we
6
7
would be thinking at -DR.
8
9
A band of tubes, though, I
KRESS:
Very
about
an
ex-vessel steam explosion under those conditions.
10
MR. WACHOWIAK:
11
DR. KRESS:
12
little
Okay.
That happened to one of the
German designs is the reason I brought it up.
MR. WACHOWIAK:
13
I guess I would like to
14
see how that came about.
15
less of an issue because it's not as -- even though
16
I'm
saying
that
17
subcooling,
we're
18
initial pours, where you have the big subcooling,
19
where
20
thought about that.
that
21
22
there
at
be
are
or
a
a
near
larger
few
degrees
saturation
concern.
in
I
of
the
hadn't
MEMBER BROWN:
Can you go back to slide
MR. WACHOWIAK:
I'm going the wrong way.
seven.
23
24
might
Long term I think there's
Seven.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
71
MEMBER BROWN:
1
You mentioned earlier the
2
water comes down the drainage pipes, goes down the
3
distributor, flows up, and then it kind of dribbles
4
down into where all the melting stuff is.
5
what I see?
MR.
6
7
10
Dribbles
is
--
okay.
Initially you could -MEMBER
8
9
WACHOWIAK:
Is that
BROWN:
If
you've
got
a
pipe
coming up and you've got it open and the water comes
up and flows -MR. WACHOWIAK:
11
It flows out of the top.
12
Eventually, though, the amount of water is such that
13
that pool will be about 17 meters deep by the time
14
we're done.
15
good dribble.
MEMBER BROWN:
16
17
18
So yes, it dribbles, but it's a pretty
You are not depending on
the tubes, then, for cooling any more at that point?
I mean, obviously if you've got -MR. WACHOWIAK:
19
pipes
basically
If we have the melt on
20
the
establishing
a
flow
pattern
21
through there, so the pipes are somewhat inclined,
22
heat the pipes up, the water comes out, the water has
23
to be replaced.
24
pattern in the --
It's a natural circulation flow
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
72
MEMBER BROWN:
1
2
So it's high enough that
it will refill the --
3
MR. WACHOWIAK:
Oh, yes.
4
MEMBER BROWN:
-- deluge and then come
5
back through there?
MR.
6
7
deluge lines.
not
through
the
There's another set of
lines?
10
MEMBER SHACK:
11
MR. WACHOWIAK:
12
MEMBER BROWN:
13
MR. WACHOWIAK:
14
MEMBER BROWN:
15
It's
That's those --
MEMBER BROWN:
8
9
WACHOWIAK:
The downcomer lines.
Well, where are they?
Yes, I remember.
That's the yellow lines.
Okay.
So, in other words,
they're down within the level of the water.
MR. WACHOWIAK:
16
Yes.
They're up higher
17
than where the core would be, but they're within the
18
level of where the pool is going to be.
19
MEMBER BROWN:
it
is,
that
So the initial problem,
20
whatever
causes
21
material to settle down there, you have blocked that
22
pipe where it starts dribbling out.
23
issue, is that correct?
MR. WACHOWIAK:
24
all
the
core
in
the
That would be an
Yes, that would be an
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
73
1
issue.
And so a couple of things that we're doing
2
with this is the exit pipes and these inlet pipes are
3
all way up against the wall.
4
the melt is is here.
5
probabilistic
6
debris might come out if we failed one of the edge
7
CRD tubes and found that we have significant space
8
between the wall and where the edge of that --
analysis
And we have done essentially a
of
MEMBER BROWN:
9
The core itself where
where
the
pile
of
core
So nothing would blow out
10
towards that and then come down the wall and plug it
11
into -MR. WACHOWIAK:
12
Right.
We looked at what
13
kind of protection we would need for these pipes and
14
also for the sump, which the design of the sump in
15
this isn't just a square sump that sits off to the
16
side.
17
the wall just for that concern that you have raised.
MEMBER APOSTOLAKIS:
18
19
It's actually something that is molded into
this heat flow?
What happens to all
Water comes out.
20
MR. WACHOWIAK:
21
MEMBER APOSTOLAKIS:
22
MR. WACHOWIAK:
23
MEMBER APOSTOLAKIS:
24
MR. WACHOWIAK:
It goes on top?
Yes.
Same thing?
And the heat?
Yes.
The heat is removed by
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
74
1
the PCCS, passive containment cooling system.
2
MEMBER APOSTOLAKIS:
3
MR.
WACHOWIAK:
Okay.
Any
steam
that
is
4
generated here will be condensed in the PCCS, which
5
then feeds back to the GDCS pools, which then will be
6
fed back down by these lines.
DR.
7
8
picture,
it
9
well-covered.
KRESS:
looks
When
like
you
the
look
liner
But those are cylinders.
at
is
that
very
They touch
10
at one point on the wall and one point on each other,
11
which is not as well covered as it looks, it seems to
12
me like.
MR. WACHOWIAK:
13
Right.
That's what I
14
mentioned early on in this, that on some of these,
15
what we call near-edge tubes, there are going to be
16
two or three vertical pipes connected to that so that
17
we do have full coverage on the wall.
CHAIRMAN CORRADINI:
18
19
People have been asking about once it is there.
20
you
21
starts up?
walk
me
through
the
logic
22
MR. WACHOWIAK:
23
CHAIRMAN CORRADINI:
24
So that will branch.
of
how
this
Can
thing
Right.
I'm still way back
at the transient.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
75
MEMBER
1
2
And
maybe
back
up
one
picture as you walk through this.
MR. WACHOWIAK:
3
4
BLEY:
numbers down here.
I should just put random
Okay.
CHAIRMAN CORRADINI:
5
So at time zero,
6
some CRD because of events that we're going to take
7
up after lunch gets us a core melt in the lower
8
plenum and a CRD breaks or leaks or a weldment opens
9
up or something.
MR. WACHOWIAK:
10
11
Yes, or the vessel will
creak rupture at one of those.
12
CHAIRMAN CORRADINI:
13
these dudes starts flowing out.
14
starting point, can you kind of talk me through how
15
this is initiated and the timing of what is supposed
16
to happen when?
17
MR. WACHOWIAK:
18
MEMBER BLEY:
Right.
So one of
So with that as the
Okay.
And I guess from what you
19
have told us, that material is now expected to be a
20
flat surface, not -MR. WACHOWIAK:
Essentially flat.
It
22
won't be perfectly flat but essentially flat.
So
23
somewhere in this part of the cylinder -- you know,
24
this is several, ten meters across.
21
And the vessel
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
76
1
itself is seven meters across.
So the CRD tubes are around this area.
2
3
We would have melt that comes out.
4
would probably drip down the CRD tubes.
5
wouldn't
6
amount
7
candling down the CRD tubes.
8
right term.
actually
came
out.
get
a
But
lot
until
let's
say
a
And you
significant
that
it's
not
That's not really the
So it comes down.
9
In actuality, it
And it would first
10
come into contact with the floor.
Right now we know
11
it's at least a grating, but it's probably going to
12
have some sort of a sheet metal on top of the floor.
The melt will just go through that sheet
13
14
metal like it's nothing there.
It's not going to
15
provide any type of protection.
We don't expect it
16
to.
17
protect is if the CRD itself fell down, we would
18
expect the grating to stop it from acting like a
19
spear and coming down and breaking the concrete.
The only thing that we would expect this to
CHAIRMAN CORRADINI:
20
21
preferentially send it to the wall.
22
MR. WACHOWIAK:
23
CHAIRMAN CORRADINI:
24
Nor will the grating
or.
No.
Or slash it there or
That would be my --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
77
MEMBER BLEY:
1
2
Why do you think that?
But
go ahead.
MR. WACHOWIAK:
3
The estimate is that it's
4
a thin metal layer and that the super heat in the
5
melt material would just act like it's not there.
MEMBER SHACK:
6
7
addenda
says
8
somewhat beefier.
you
could
be
making
that
Melt right through it,
yes.
MEMBER
11
12
think
MR. WACHOWIAK:
9
10
you
But one of your recent
APOSTOLAKIS:
What
kind
of
temperatures are you talking about?
13
MR. WACHOWIAK:
Three thousand, 4,000.
14
MEMBER APOSTOLAKIS:
15
MR. WACHOWIAK:
F?
Yes, I think.
Yes.
We
16
have melted the core.
And then we have had to remelt
17
it in the bottom of the vessel so that we fail the
18
vessel.
19
remelted in the bottom of the vessel.
20
just fuel pellets falling out or anything like that.
And then it's going out after it's been
So it's not
21
MEMBER APOSTOLAKIS:
Yes.
22
MEMBER
thought
SHACK:
I
Okay.
So --
your
recent
23
comment was, though, you were thinking of beefing up
24
that
floor
presumably
to
more
protection
against
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
78
1
falling thingie bobs.
MR. WACHOWIAK:
2
Right.
"Falling thingie
3
bobs" is probably the best term that I've heard for
4
that so far.
5
wouldn't be anything that would protect or that would
6
stop --
But, even if we did that, it still
7
MEMBER SHACK:
8
MR. WACHOWIAK:
9
MEMBER
10
Stop hot core melt.
-- molten core melt.
SHACK:
I
mean,
not
putting
a
ceramic layer there.
MR. WACHOWIAK:
11
The only thing is we just
12
don't want things to fall directly onto the concrete
13
and have some sort of mechanical damage because of
14
the concentrated impact.
Once
15
again,
we
don't
expect
that
to
16
happen either because these CRDs are all connected by
17
wires
18
probably
19
They're probably going to be hanging by their wires
20
and
21
hydraulic
22
scrambler rods.
23
you know, miles of cable connected to it, too.
and
zooms
not
going
everything
everything
to
else
lines,
So
24
and
fall
straight
anyway
water
else.
down
because
lines
that
They're
go
anyway.
there
to
are
their
Plus, each one of the motors has,
while
that's
an
issue,
theoretical
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
79
1
issue, I think, in practicality, what we will see is
2
there is no way those can -MEMBER BROWN:
3
Why wouldn't the heat of
4
the core melt just destroy that stuff, just like it
5
does the grating?
6
MR. WACHOWIAK:
7
MEMBER BROWN:
8
MR. WACHOWIAK:
9
were, "Well, okay.
It will.
Okay.
So if it comes --
But the initial thoughts
We opened this hole.
And this
10
thing that is hanging down will just fall."
Well,
11
the thing that is hanging down there will probably
12
fall into the tree forest of CRD tubes.
And then the wires will melt, and the
13
14
tubes will melt and fall.
15
direct spear coming down, but it will be some sort of
16
a progression of material falling down from the top.
17
But that's kind of off from what we were getting at
18
You know, it won't be a
here.
As the material is coming down, we are
19
20
heating the air space of the lower drywell.
And I'll
21
get to why that is important in a minute.
We would
22
be heating the air space.
23
starts to collect on the floor, there is a net, an
24
array of thermal couples either on top of the floor
And then as the material
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
80
1
or embedded in the floor.
2
CHAIRMAN CORRADINI:
3
MR. WACHOWIAK:
TBD?
It's probably going to be
4
a combination of both.
The array or spacing of those
5
is not TBD.
6
there are and where they are located.
We've got that in the design of how many
7
But let's just say that in every given
8
region, there are two thermal couples that are there
9
and
that
when
had
the
prescribed
indication
that
number
there
of
is
adjacent
10
regions
increased
11
temperature there or lack of continuity because if
12
the melt melts the thermal couple wire, we are going
13
to assume that it was the melt that caused that.
14
So if we get the right array and we have
15
done a heat transfer calculation to detect what that
16
array is, then the system that actuates these valves
17
here will send the signal to those.
18
are squib valves, essentially two-inch squib valves,
19
similar to what we use in standby liquid control
20
systems today.
That
system
send
a
21
22
values
or
23
valves.
24
wire
would
signal
to
Right now they
actuate
those
squib
actuate
those
squib
But just prior to or just outside of the
coming
into
these,
there
is
a
temperature
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
81
1
switch.
The air temperature in the lower drywell
2
needs to be above the saturated temperature given no
3
core melt, which is 575-600 degrees K?
4
CHAIRMAN CORRADINI:
5
MR. WACHOWIAK:
F?
F.
F, F.
So if the air
6
temperature is hot enough to set those temperature
7
switches, then the squib valves will actually fire
8
and will start putting water down into the -CHAIRMAN CORRADINI:
9
So can I just repeat
10
up to this point what you have said?
11
combination of there are two things that have got to
12
occur.
13
something above something like 600 F.
14
be some sort of logic in terms of either measure
15
temperature or lack of continuity of embedded thermal
16
couples.
Those
sensors
are
going
to
So it's a
have
to
see
And there will
17
And you need both of those.
18
both of those to essentially register to allow the
19
squib valves to fire.
20
MR. WACHOWIAK:
21
MEMBER APOSTOLAKIS:
22
MR. WACHOWIAK:
You need
Yes.
Why both?
The reason we have the
23
temperature switches there is because we are really
24
worried.
Okay.
We are worried about them going off
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
82
1
when we don't want them to.
2
MEMBER APOSTOLAKIS:
3
MR. WACHOWIAK:
4
dealing in severe accident space here.
5
much better day at the nuclear plant if we prevent
6
the severe accident than if we have to deal with the
7
severe accident.
8
in the GDCS tanks to be able to go into the core,
9
rather than somewhere for some reason going into the
10
I was just curious.
And, remember, we are
And it's a
So we really want that water to be
lower drywell before we need it.
11
MEMBER APOSTOLAKIS:
12
MR. WACHOWIAK:
Accidentally.
So that's the reason why
13
you have to have the confirmatory high temperature in
14
the gas space.
15
MEMBER APOSTOLAKIS:
16
MR. WACHOWIAK:
the
ABWR.
similar
18
completely actuated by air space temperature.
DR. KRESS:
19
The
And, once again, that is
17
20
to
Okay.
ABWR
deluge
system
is
Why didn't the GDCS water go
into the core in the first place?
MR. WACHOWIAK:
21
Because the valves that
22
lead from the GDCS tanks to the core failed in the
23
closed position or the two digital control systems
24
that control those valves failed.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
83
1
And I was going to get to this point.
2
The control system that actuates this is a third
3
diverse control system from everything else that we
4
had been operating.
5
MEMBER BLEY:
Diverse power?
6
MR. WACHOWIAK:
Diverse power, diverse
7
platform.
It's not a programmable system.
8
system with its own batteries.
CHAIRMAN CORRADINI:
9
It's a
We will come back to
10
this because I can see the PRA types are getting
11
actually excited again.
MEMBER ABDEL-KHALIK:
12
Would a common mode
13
failure of the squib valves also affect this?
The
14
same common mode failure that would prevent the water
15
from getting into the vessel would prevent the water
16
from going into the BiMAC?
17
MR. WACHOWIAK:
If we used a common squib
18
valve or common squib valve charge there, then yes.
19
But we're specifying that that has to be diverse.
20
One of the reasons why we have 12 lines
21
coming down here, rather than just 4 lines coming
22
down, like out of each of the GECS pipes, is we want
23
to use a completely different squib valve here than
24
what we're using for the inject lines.
So while
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
84
1
those are specifically designed squib valves for that
2
purpose of injection, these are more similar to squib
3
valves used in nuclear applications used today.
So
4
we're
concerned
with
that.
We're
5
specifying that it needs to be a diverse valve so
6
that we don't have that common mode failure.
7
would eliminate any benefit from having this in the
8
first place.
9
failure.
That
It would be to get that common mode
10
CHAIRMAN CORRADINI:
11
DR. KRESS:
Keep on going.
There is no problem between
12
the race between the melt getting down to the tubes
13
and the timing of turning on those squib valves, is
14
there?
MR. WACHOWIAK:
15
That all depends on our
16
final configuration of the layer here, but, you know,
17
we've got tens of minutes to deal with.
DR. KRESS:
18
19
Before the melt gets to the
tubes?
MR. WACHOWIAK:
20
Before the melt gets to
21
the tubes.
And that was when we were still looking
22
at a 20-centimeter layer.
23
more time now that we have done our optimization.
DR.
24
KRESS:
So I think we've got even
Have
you
dealt
with
that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
85
1
transient
period
2
actual
3
sacrificial layers?
4
before.
thermal
calculationally
effect
is
on
to
the
see
what
the
tubes
and
the
I guess this has been asked
CHAIRMAN
you
are
6
exactly where I was going to ask the question.
So
7
keep on going.
5
CORRADINI:
DR. KRESS:
8
No,
but
Well, that's the question.
9
Have you dealt with that transient period to be sure
10
there's not a way for that melt to get there faster
11
or for the thermal effect to get there faster?
MR.
12
WACHOWIAK:
I'll
have
to
go
back
13
through the report and see if that was addressed in
14
there.
CHAIRMAN CORRADINI:
15
We didn't see it.
I
16
looked for this in particular.
17
a different way, but it is really Tom's question,
18
which is so you said you have got these temperature
19
things
20
sacrificial layer or above the layer.
21
this double-check on the air temperature.
22
they both say yeah, something is amiss.
23
buggers off.
somewhere
in
the
And so let me say it
concrete
or
the
And you've got
And now
Fire these
Let the water in.
Now the water is coming in.
24
in
I am trying
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
86
1
to figure out how much melt is there and what is the
2
composition of the melt where it fires or does it not
3
matter.
4
all metallic and there is a hell of a lot of it and
5
we're okay or it's all oxidic and there's a hell of a
6
lot of it and we're okay or there's not a whole lot
7
of either and I dump a lot of water in there, now a
8
hell of a lot of it comes in and we're okay.
And you have done calculations to show it's
And I've been looking for all three of
9
10
those possibilities.
And I have to admit I've not
11
seen it anywhere.
12
it kind of goes back to Tom's point.
And that's the source of my -- and
MEMBER ABDEL-KHALIK:
13
Now, before those
14
valves activate, the pipes are presumably filled with
15
nitrogen.
Is that correct?
MR.
16
17
expect, yes.
WACHOWIAK:
That's
I
would
It's -MEMBER ABDEL-KHALIK:
18
what
Is that a better
19
arrangement than if the pipes were to be kept always
20
full of water?
MR. WACHOWIAK:
21
The issue that we have if
22
we keep the pipes full of water by putting the valves
23
down here, you know, if we start with a full pipe
24
with a valve down here, we run into the issue that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
87
1
the
melt
eject
may
impact
the
valves.
2
environment for the valves is adverse.
The
3
So what we have elected to do was we put
4
the valves up and on the deck, up top, so that they
5
are
6
itself.
7
at.
not
affected
whatsoever
by
the
melt
material
So that was a trade-off that we had to look
8
Are we more worried about how fast the
9
water gets there or more worried about the valves
10
being
able
to
survive
11
environment?
12
reliability in a severe accident environment.
We
chose
in
a
making
severe
the
accident
valves'
higher
13
So we are taking the hit on the timing
14
there, but we don't think that the timing is going to
15
be that great.
16
starts going through the pipe to when it gets into
17
the
18
compared to how we detect the actual amount of melt
19
on the floor.
20
to get thermal couples registered at each of the
21
cells that we need to activate that.
BiMAC
is
I think the time from when the water
going
to
be
a
fairly
period
I think that is going to take longer
MEMBER ABDEL-KHALIK:
22
short
I'm just concerned
23
in this case, sort of following up on what Tom's
24
question is, about if the pipes are already very hot
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
88
1
so that the wall temperature is above the minimum
2
film
3
gravity-feed these pipes, the water is not going to
4
get in there.
boiling
temperature
MR. WACHOWIAK:
5
and
you're
trying
to
I think that is one of
6
our RAIs that we're working on right now.
7
initial look at that was that we don't think that
8
they're going to be that hot when the water gets
9
there.
MR.
10
SEEMAN:
There
would
11
material that has no test to go through.
12
MR. WACHOWIAK:
And the
be
enough
But that is one of the
13
staff's questions that they've asked us to provide a
14
response on.
MEMBER BROWN:
15
16
couldn't -MR. WACHOWIAK:
17
18
Are you worried that they
Apparently they couldn't
find it either.
MEMBER BROWN:
19
Are you worried about the
20
valves being damaged by the stuff that comes down or
21
just the environment at the high temperature --
22
MR. WACHOWIAK:
23
MEMBER
24
BROWN:
It would be ---
caused
them
not
to
operate?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
89
1
MR. WACHOWIAK:
It would be both.
2
MEMBER BROWN:
You've already said that
3
you're not going to have this material get out to the
4
wall where it is going to plug the pipes.
5
you're saying, well, we're worried about it getting
6
out and damaging the squib.
7
there?
8
MR. WACHOWIAK:
9
MEMBER
10
And now
Did I miss something
The main --
BROWN:
Am
I
mischaracterizing
that?
11
MR. WACHOWIAK:
12
the thermal environment that is there.
13
MEMBER BROWN:
14
the
valve
would
be
15
couldn't open whatever --
Yes.
Okay.
such,
Right.
Because we're
18
degrees.
So, you know, it's going to be not a very
19
nice
20
they're splashing and things like that, too.
21
get small amounts of material and things, that could
22
also be a problem.
valves
MEMBER BLEY:
23
to
they
the
for
here
that
expecting
environment
space
enough,
17
24
gas
So the internals of
hot
MR. WACHOWIAK:
16
The main part was
down
be
up
over
there.
600
Plus,
If you
Have the squib valves been
tested in a 600-degree environment?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
90
MR. WACHOWIAK:
1
Once again, we're moving
2
them out of the 600-degree environment so that we
3
don't have to do that.
4
5
MEMBER BLEY:
They're far enough that
MR. WACHOWIAK:
But the equipment in the
they --
6
7
drywell, my understanding is that the equipment in
8
the drywell, all needs to be qualified to like a
9
575-degree temperature.
That's what we do for the
10
ICS valves and other valves in the drywell.
11
CHAIRMAN CORRADINI:
So just to repeat,
12
Rick, to let you go, so did we miss it?
13
look for it.
14
it's back to Tom's original question.
15
transient analysis somewhere that we missed?
I didn't
I gave the three possibilities.
But
Was there a
16
Because I didn't see it; that is, the
17
start-up, everything you just walked us through and
18
all the possible branches of how this thing actually
19
gets to the steady state that the tests have been
20
done at.
MR.
21
22
WACHOWIAK:
I
think
that
is
the
genesis of your question, right, Ed?
CHAIRMAN CORRADINI:
23
24
Am I missing something?
My question is, so
the staff can answer it in light of this, is there an
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
91
1
RAI
that
asks
to
describe
the
analysis
of
the
2
start-up of this, the start-up transient, since we
3
get to the steady state?
4
Because it's very clear this has been
5
documented when you get to the steady state, how it
6
performs.
7
actuation to the point of steady state.
I'm still trying to get from the point of
8
MR. FULLER:
9
CHAIRMAN CORRADINI:
10
Okay.
RAI asking how this happens?
11
MR. FULLER:
12
MEMBER APOSTOLAKIS:
Not yet.
13
the microphone.
14
say it again and who you are?
Yes.
This is Ed Fuller from
the staff.
To answer Professor Corradini's question,
17
18
no, we have not asked that question yet.
19
CHAIRMAN CORRADINI:
20
MR. WACHOWIAK:
21
Okay.
So I would expect that
this afternoon.
22
(Laughter.)
23
MR. WACHOWIAK:
24
You have to go to
Ed, can you go to the microphone and
MR. FULLER:
15
16
Has there been an
And I know we have talked
about that before, and I think --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
92
CHAIRMAN
1
CORRADINI:
That's
fine.
I
2
think Tom has characterized it best in that we are
3
trying to put in our minds how this thing starts up.
4
And we might have five different incorrect versions
5
of it.
So we want to get clear how you guys view it
6
and how it is supposed to operate.
MEMBER BLEY:
7
8
me.
9
to it.
Just a simple question for
That picture you had you don't have to go back
Those overflow pipes that dump water, are
10
they essentially level with the floor?
11
it looks like with the sacrificial floor.
12
not much above it if they are above it.
13
MR. WACHOWIAK:
14
MR. SEEMAN:
15
17
They are
Do you know what the --
I thought that was that 1.5.
It would be like two meters.
MEMBER BLEY:
16
That's what
So I believe it is.
So it's right down there
near it.
18
MEMBER APOSTOLAKIS:
Didn't you say that
19
eventually you were going to have 17 meters?
20
MR. WACHOWIAK:
Of water.
21
MEMBER SHACK:
But they only just make
22
this above any potential depth of the melt.
23
CHAIRMAN CORRADINI:
24
is the overflow line.
Yes.
Dennis' point
So once you start natural
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
93
1
circulation, where does it flow back down?
MEMBER BROWN:
2
It is the right-hand pipe
3
that kept going back out, that little cutout that
4
he's got there in the dark blue, this one here.
5
yellow pipe is where it goes back.
6
CHAIRMAN CORRADINI:
7
MR. WACHOWIAK:
8
MEMBER BLEY:
10
yes.
11
come back?
And then how high that
I was going to ask both,
Where does it dump out, and then where does it
MR.
12
WACHOWIAK:
That
MEMBER APOSTOLAKIS:
15
two meters, you said?
16
with water, right?
likely
be
MR. SEEMAN:
18
MEMBER APOSTOLAKIS:
water come out?
Right.
CHAIRMAN
21
guys, it boils out.
22
back down.
CORRADINI:
From
the
little
From the yellow one, it flows
MR. WACHOWIAK:
to scale.
So how does that
From the two-meter pipe?
20
23
So it would be what,
And then it starts filling up
17
24
would
somewhat higher than the water pipes.
14
19
Right.
is, that's what Dennis is --
9
13
The
This is not necessarily
This would be a much bigger pipe.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
94
MEMBER BLEY:
1
2
You said that only two or
three meters.
3
MEMBER APOSTOLAKIS:
4
MEMBER BROWN:
5
the natural circulation gets started.
6
full of air, and then the water fills up.
7
we're
8
pipes?
going
get
MEMBER
9
10
to
it
Yes.
So it gets covered before
started
So it's all
through
APOSTOLAKIS:
It's
And then
the
empty
covered
by
water.
MR. WACHOWIAK:
11
It's all under water,
12
though.
The water itself -- so the BiMAC is in this
13
little flat area down here that you can't even see.
14
And the water level essentially will be here.
15
MEMBER BLEY:
16
MR. WACHOWIAK:
17
20
About even with the top
of the suppression.
MEMBER BLEY:
18
19
Up even with the pool?
And the downcomer wouldn't
be up real high, where it's likely to be the coolest?
It will be just partway up?
MR. WACHOWIAK:
21
It will be partway up
22
because, remember, it can't be -- we don't want it to
23
be in the way of servicing the CRDs and things there.
24
We wanted that all to be out of the way.
We want it
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
95
1
to be under the grading once again so that people
2
aren't dropping stuff in there during the outages.
3
You know, that would not be a good thing.
MEMBER
4
BLEY:
On
this
picture,
the
5
grading is kind of that thing that looks like the top
6
of a capstan there, right under the CRDMs?
MR. WACHOWIAK:
7
8
two white dots, I guess.
9
MEMBER BLEY:
Underneath the thing, the
Okay.
10
MR. WACHOWIAK:
11
MEMBER BLEY:
12
MEMBER BROWN:
Way down at the bottom.
Oh.
Way down there.
Okay.
So a downcomer is just a
13
pipe that water fills up.
14
It's open at the bottom.
15
the place, it then starts coming back down because
16
it's hot water.
17
to get started.
18
bottom.
20
When the water builds up in
But the pipe hasn't been filled up
It's not going to fill from the
MR. WACHOWIAK:
19
It's open at the top.
When we fill the water in
with these things --
21
MEMBER BROWN:
Is it?
22
MR. WACHOWIAK:
Here this is open to that
23
channel so the water will go, will fill in here, and
24
will also fill up here.
So the water level here and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
96
1
here and here will all be the same.
MEMBER
2
BROWN:
Well,
you
show
the
3
downcomer pipe as being above the grading level on
4
figure 8.
5
it?
That's why I was asking the question.
6
MR. WACHOWIAK:
7
MEMBER BROWN:
Okay.
It's I don't know how ever
8
many meters it is above the grading level.
9
it's open up at the top.
10
11
builds up.
And pretty soon the water
And it's got to come back down that pipe.
DR.
KRESS:
That
is
part
MR.
WACHOWIAK:
the
That
is
part
of
the
what
is
the
transient question.
16
MEMBER BROWN:
17
MEMBER
All right.
APOSTOLAKIS:
So
18
natural circulation part at that stage?
19
water has already -MR.
20
21
of
transient.
14
15
And then
Does that get started okay?
12
13
See
WACHOWIAK:
I mean, the
The
right
three-dimensional --
22
MEMBER APOSTOLAKIS:
The transient --
23
MR. WACHOWIAK:
24
pattern will be out through this pipe --
The natural circulation
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
97
1
MEMBER APOSTOLAKIS:
2
MR.
3
WACHOWIAK:
Right.
--
and
then
into
the
channel pool itself.
4
MEMBER APOSTOLAKIS:
5
MR. WACHOWIAK:
6
be coming back in.
7
MEMBER
Right.
And then the pool would
APOSTOLAKIS:
In
the
other
8
picture, the previous picture, it would come down
9
from where, from the deluge?
MEMBER BROWN:
10
11
No.
where the initial water goes down.
12
MEMBER APOSTOLAKIS:
13
MEMBER BROWN:
14
the downcomer pipe.
MR. WACHOWIAK:
16
MEMBER APOSTOLAKIS:
17
MEMBER BROWN:
This is just -Oh, for the right.
But that's a downcomer.
Isn't that representing a downcomer?
MR. WACHOWIAK:
19
20
The initial water.
But the right-hand one is
15
18
The left-hand one is
place.
It's coming from the same
They come from GDCS.
21
get
MEMBER STETKAR:
It's the GDCS pools,
condensate
from
22
which
back
23
exchangers, which flow back down the lines.
MR. WACHOWIAK:
24
the
PCCS
heat
Anything that steams out
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
98
1
will come back that way.
2
pool will also circulate.
MEMBER APOSTOLAKIS:
3
4
But the water from in the
So how does that
happen?
MR. WACHOWIAK:
5
It is hard to show with
6
these two-dimensional pictures because we don't have
7
the right slice to answer your question, I think.
DR. KRESS:
8
On the next picture, when you
9
initially turned on the first deluge lines to go into
10
the pool, the blue line, what keeps the water from
11
going
12
tubes?
back
up
to
yellow,
13
MR. WACHOWIAK:
14
DR. KRESS:
15
MR. WACHOWIAK:
instead
of
through
the
Nothing.
Some would, right?
Some would.
That's why I
16
was saying that the water level in here and the water
17
level in these would be the same.
18
that's going to prevent it from going up through
19
here, which gets back to the question, what is the
20
temperature on this when it starts so that we'll get
21
an
22
pipes?
even
distribution
of
23
DR. KRESS:
24
MR. WACHOWIAK:
water
There's nothing
through
all
those
Including the yellow pipes?
Yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
99
MEMBER BLEY:
1
For Charlie's question, if
2
I understand this right, the downcomer actually feeds
3
into the distributor?
4
it dump into the bottom of the pool?
MR. WACHOWIAK:
5
6
the distributor.
Down and under or through
the pipes.
MEMBER BROWN:
9
10
The downcomer feeds into
That's correct.
MEMBER BLEY:
7
8
Is that what it does or does
Okay.
So it feeds just
like the initial feed?
11
MR. WACHOWIAK:
Yes.
12
MEMBER BROWN:
So that the downcomer has
13
to then go -- in order to do the natural circulation,
14
it has to go back out through the pipes and then out
15
through these little things, where it dribbles, not
16
really dribbles, dribbles outward?
17
MR. WACHOWIAK:
Yes.
18
MEMBER APOSTOLAKIS:
I'll come back to
19
this transient business.
20
It's one meter, two meters, three meters.
21
what happens.
22
point starts coming down.
Tell me
I mean, it keeps going up or at some
CHAIRMAN CORRADINI:
23
24
The water starts rising.
to tell you just yet.
They are not going
They're going to get back to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
100
1
us on that.
2
MEMBER APOSTOLAKIS:
3
CHAIRMAN CORRADINI:
4
to tell us just yet.
5
on that.
What?
They are not going
They're going to get back to us
6
MEMBER APOSTOLAKIS:
Okay.
7
CHAIRMAN CORRADINI:
Is that fair?
8
MR. WACHOWIAK:
9
CHAIRMAN CORRADINI:
DR. KRESS:
10
Good.
Yes.
Okay.
And tell me what you do in
11
natural convection.
Is this hot water coming out?
12
Is it assumed mix with everything?
13
got a colander of water to feed down.
14
misnomer.
15
mixing, that calculation doesn't work.
16
usually the way you calculate it.
That's a
If you've got bad distribution or bad
MEMBER APOSTOLAKIS:
17
And then you've
But that's
But I also want to
18
understand as the level rises, at which point does it
19
go to pipes and come -DR.
20
KRESS:
It
just
gives
you
more
21
driving force to drive stuff through the line in the
22
calculation the way it's calculated.
CHAIRMAN CORRADINI:
23
24
In chapter 21 and in
the BiMAC test report, they give one figure on what
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
101
1
they expect the heat flux distribution.
2
agree with this one part, which is that downcomer or
3
the center pipe is going to see much less heat flux.
4
So the natural point is it is going to want to boil
5
up the small pipes and get fed by the large pipe
6
because of the maldistribution of heat flux to the
7
big pipes, which are the small pipes.
But
8
9
I
think
George's
I guess I do
point,
I
think
George's point, is as you're filling up, how does the
10
thing get to steady state again?
11
the how do we get to steady state question.
MR. WACHOWIAK:
12
So we're back to
And I think some of the
13
things that we will be looking at here is as it first
14
starts to dribble over, if you will, it is putting
15
water on top of the melt.
16
boil away.
And that is all going to
17
Until we establish a good crust on the
18
top, we are not going to be filling up that whole
19
pipe.
20
condensing in PCCS, and coming back down the deluge
21
lines.
going
to
be
boiling,
going
to
PCCS,
So early on the flow path is going to be
22
23
It's
through the whole containment -MEMBER APOSTOLAKIS:
24
Right.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
102
MR. WACHOWIAK:
1
-- until we establish a
2
good crust on top of the core.
3
to start filling up into the --
4
MEMBER APOSTOLAKIS:
5
And then it's going
That's where I'm not
sure I understand.
CHAIRMAN CORRADINI:
6
So let me ask this
7
one question.
There were spreading experiments back
8
in 1988 and '89 at ISPRA about core melt spreading
9
and if it stays where you think it's going to -- if
10
it spreads evenly or it goes where you think it is
11
going to go.
Have
12
that
you
if
assessed
you
to
yourself
14
asymmetrically coming down -- I'm back to transient;
15
I'm sorry, I can't get off of it -- that something
16
comes down, as you said, not at the wall because the
17
wall is way far away from where the first CRD, it
18
starts piling up here, that it's not going to simply
19
stay piled up and focus an attack on the basemat
20
there?
MR. WACHOWIAK:
something
convince
13
21
have
those
that
is
And we'll have to get
22
back to you on that.
I know we looked at that thing
23
early on when we were trying to decide what to do and
24
when we were deciding if we were going to put the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
103
1
BiMAC in.
2
I know that we had a whole bunch of those
3
analyses that we looked at, but I don't know that
4
that
5
document.
earlier
made
it
MEMBER APOSTOLAKIS:
6
7
material
into
the
final
When are they coming
before the full Committee?
8
CHAIRMAN CORRADINI:
They?
9
MEMBER APOSTOLAKIS:
They.
CHAIRMAN
10
11
October.
12
plan.
October.
CORRADINI:
14
weeks.
15
writing a letter?
These
they?
In a month and a few
So you will have the answers then?
CHAIRMAN CORRADINI:
16
They?
Right at this moment, that's the
MEMBER APOSTOLAKIS:
13
These?
You are
Well, we have never
17
held off an -- it's an interim letter.
18
held off an interim letter if we have open questions.
19
They will appear as conclusions and notes to the
20
staff.
MEMBER APOSTOLAKIS:
21
22
We've never
No, no, no.
When we
say, "Welcome back" to you, when will that happen?
23
CHAIRMAN CORRADINI:
For them?
24
MEMBER APOSTOLAKIS:
Yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
104
1
2
CHAIRMAN CORRADINI:
I don't think by
MEMBER APOSTOLAKIS:
But during the full
October.
3
4
Committee presentation, I mean, if you have already
5
done it -CHAIRMAN CORRADINI:
6
Let me just try to
7
help this out.
8
RAIs and staff will come in and present and we feel
9
comfortable,
10
So if they can answer some of the
then
that
will
be
included
in
the
interim letter.
11
If this still is in the path where they
12
are trying to answer the RAIs, we have to decide,
13
will we write an interim letter on chapters 19 and
14
21, listing some things that are still open that
15
we're concerned about.
MEMBER
16
17
APOSTOLAKIS:
Is
that
your
understanding?
18
MR. WACHOWIAK:
19
DR.
KRESS:
Yes.
I
know
this
is
just
a
20
diagram, but if you look at that smallest tube at the
21
top --
22
MR. WACHOWIAK:
23
DR. KRESS:
24
Yes.
-- it looks like there's a
part of the liner that is not protected.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
105
MR. WACHOWIAK:
1
2
cartoon.
It's not shown.
3
DR. KRESS:
4
MR. WACHOWIAK:
5
It's just a -The idea is that when
this pipe comes out, we would have --
6
DR. KRESS:
7
MR. WACHOWIAK:
8
That's because of this
We would actually have --- coverage of tubes all
across the all.
DR. KRESS:
9
And actually having coverage.
MR. WACHOWIAK:
10
Yes.
We have looked at
11
3-D modeling of this thing.
12
that some of these have to be branched into more than
13
one.
DR.
14
KRESS:
And we can arrange it so
Have
you
considered
any
15
issues about thermal warping of those tubes due to
16
the temperature distribution?
17
or something and uncover part of the liner?
MR. WACHOWIAK:
18
Maybe they got offset
We're not expecting those
19
tubes themselves to get that hot.
20
to be the expectation is that all of these tubes are
21
filled with water and are in a -DR. KRESS:
22
It's still going
I was concerned about the
23
temperature distribution on a lot of them, rather
24
than actual temperature.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
106
MR. WACHOWIAK:
1
Yes.
And I think what
2
our experiment showed is that the bulk temperature in
3
those tubes is remaining all right around saturation.
4
So if we're talking about between subcooled -DR. KRESS:
5
6
Okay.
You wouldn't have any
-MR. WACHOWIAK:
7
-- and saturation, you
8
know, that type of temperature, but the extremely
9
high temperatures, I don't think we are expecting to
10
see in the tube themselves because they're built for
11
it.
CHAIRMAN CORRADINI:
12
13
going.
I
14
minutes.
15
to get to?
like
to
take
a
break
in
a
few
Do you have a natural break point you want
MEMBER ABDEL-KHALIK:
16
17
would
We've got to keep on
Could you stop at
that slide?
18
MR. WACHOWIAK:
Which one?
19
MEMBER ABDEL-KHALIK:
20
MR. WACHOWIAK:
21
MEMBER
Twelve.
Got it.
ABDEL-KHALIK:
Now,
the
22
implication here is that CHF is the limiting heat
23
flux, but OFI can happen at a lot lower heat flux
24
than CHF.
And if that happens, some of the tubes
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
107
1
will actually dry out.
2
MR. WACHOWIAK:
3
MEMBER ABDEL-KHALIK:
4
done?
Okay.
So has that been
Has that analysis been made?
MR. WACHOWIAK:
5
I guess I don't know the
6
answer to that question whether in determining that
7
CHF was the right parameter to use here, that that
8
analysis was done and then not included in the report
9
or if it was not done.
So
10
I
don't
know
11
question.
12
think we can have it answered.
13
MEMBER
the
answer
to
your
So if we can get that question, then I
ABDEL-KHALIK:
The
question
is
14
simply, we want to know what the OFI limit is for
15
this set of piping.
16
MR. WACHOWIAK:
Okay.
17
CHAIRMAN CORRADINI:
Can I help or modify
18
the questions?
19
it.
20
this part of the BiMAC report.
21
the color coding and everything, I just couldn't get
22
it.
Maybe I missed it.
I really couldn't understand
I tried.
With all
They probably had instabilities in their
23
24
I would expect -- I think you said
actual
measurements.
And
I'm
curious
how
they
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
108
1
discerned
when
2
because I think Said's point is well-taken that since
3
they're looking at the red line, compared to the blue
4
or the black line, they probably crossed the boundary
5
where
6
behavior.
7
it?
they
they
were
had
it
and
actually
And I'm curious.
getting
they
didn't
counterflow
Were they able to detect
8
MR. WACHOWIAK:
9
CHAIRMAN CORRADINI:
10
when
Right.
That's another way
of asking the same sort of question, yes.
MEMBER ABDEL-KHALIK:
11
you
reached
saturation
I mean, the fact
12
that
13
preclude the possibility of having leatherneck-type
14
instability.
MR. WACHOWIAK:
15
temperature
Okay.
16
good question.
17
Okay.
18
this one might take a little bit.
19
CHAIRMAN
does
not
I think that's a
We should be able to answer that.
Now is a good time for a break because I think
CORRADINI:
So
just
to
make
20
sure, we are not going to come back to the BiMAC at
21
this point?
22
MR. WACHOWIAK:
23
CHAIRMAN CORRADINI:
24
Okay.
No.
We are pushing on.
We are pushing on.
Go ahead, Said.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
109
1
MEMBER ABDEL-KHALIK:
2
to talk about their testing or -CHAIRMAN CORRADINI:
3
4
Aren't they going
I think we might
drag them back to that, but -MEMBER ABDEL-KHALIK:
5
Drag us back to the
6
testing, but in that, we will need to be close on
7
what we go into, the proprietary or not.
8
going to have --
9
CHAIRMAN CORRADINI:
10
MEMBER ABDEL-KHALIK:
11
CHAIRMAN
But I'm
On the testing?
On the testing.
CORRADINI:
Maybe,
then,
we
12
might hold off on that.
So let's take a break until
13
a quarter of.
14
(Whereupon, the foregoing matter went off the record
15
at 10:25 a.m. and went back on the record
16
at 10:44 a.m.)
17
CHAIRMAN CORRADINI:
Let's get started
18
again.
So the question is to the members is that
19
Rick is going to go on to other issues.
Do we have
20
questions
about
21
experiments?
that
are
in
open
session
the
22
And if not, if we are going to get to
23
details about dimensions and angles and such in that,
24
we should hold off and bring that up later.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
110
1
MR. WACHOWIAK:
2
CHAIRMAN CORRADINI:
3
Yes.
So we'll move on at
this point through your presentation.
MR. WACHOWIAK:
4
All right.
A couple of
5
things that came up in the last thing we were able to
6
take a look at on the break.
7
the report on the instability.
MR. SEEMAN:
8
9
Glen found something in
Leatherneck flow instability
is discussed on page 23 of the test report.
And that
10
discusses that it wasn't possible to reach that in
11
flow instability at the experimental levels that were
12
used in the test.
MR. WACHOWIAK:
13
At the heat flux levels
14
that were used, you couldn't get there.
15
for it and couldn't get to it.
16
MEMBER
17
He looked
So that's discussed.
ABDEL-KHALIK:
But
does
that
exclude it from the expected conditions?
18
MEMBER BLEY:
19
levels compare with what we might see?
MR. SEEMAN:
20
How did those heat flux
They were bounding.
The
21
heat flux levels in the test were bounding compared
22
to expected, the heat fluxes in the BiMAC.
DR. KRESS:
23
24
It was calculated by CFD, a
whole lot of decay heat and a lot of melt.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
111
1
MR. SEEMAN:
Right.
2
DR. KRESS:
Decay heat was for six hours
3
after shutdown?
4
MR. SEEMAN:
Right.
5
DR. KRESS:
How did you arrive at that
6
That was the --
particular value?
MR. WACHOWIAK:
7
We looked at the core
8
damage sequences.
9
sequences were all six hours or later before we had
10
melt go to the vessel.
DR.
11
12
And essentially the significant
KRESS:
The
core
melt
sequences?
13
MR. WACHOWIAK:
14
DR. KRESS:
15
MR. WACHOWIAK:
16
dominant
dominant.
The significant ones.
The significant ones.
It's a little more than
It was -MEMBER ABDEL-KHALIK:
17
to
18
like
19
scaled to determine whether or not the statement that
20
you made that the heat flux that you used in the
21
experiment
22
relevant.
is
more
about
bounding
how
for
MR. WACHOWIAK:
23
24
see
But I guess I would
the
the
experiments
actual
Okay.
that was in the scaled experiment.
were
system
Right.
is
Because
That was in the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
112
1
scaled experiment.
2
MEMBER
3
CHAIRMAN
CORRADINI:
right,
The
one-half,
one-quarter, right?
6
MR. WACHOWIAK:
7
CHAIRMAN CORRADINI:
8
of it at one-half scale.
9
MR. SEEMAN:
10
Right,
right.
4
5
ABDEL-KHALIK:
scaling.
Yes, quarter -I'm sorry.
Quarter
Excuse me.
That is an RAI to discuss
We have an RAI on that scaling.
MEMBER ABDEL-KHALIK:
11
Are we going to get
12
to this later on in a closed session where they talk
13
about the experiments and --
14
MR. SEEMAN:
15
MEMBER ABDEL-KHALIK:
16
MR. WACHOWIAK:
that
maybe
I'll
We want to.
We will.
Okay.
Thank you.
All right.
bring
17
thing
18
question about dribbling out of the pipes came from
19
somebody else --
20
MEMBER ABDEL-KHALIK:
21
MR. WACHOWIAK:
up
later
The other
since
the
Just keep on going.
We'll keep going.
All
22
right.
So the next area that we want to look at is
23
the high-pressure melt eject scenarios.
24
here is direct containment heating and then the local
The issue
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
113
1
failures of the liners.
In the direct containment heating, the
2
3
way
we
address
4
bounding parameters for the high-pressure melt eject
5
and
6
pressure suppression containment to see if we could
7
withstand that scenario.
then
DR.
8
9
looked
process?
at
one
the
KRESS:
way,
is
dynamic
This
we
looked
response
is
using
of
the
at
the
ROAAM
You used the ROAAM process to do this?
MR.
10
that,
WACHOWIAK:
Yes,
yes,
which
is
11
basically establishing the theoretical basis and then
12
doing an expert elicitation, I guess, or review on
13
that.
The local liner failures, we looked at
14
15
that.
And, once again, we can't preclude local liner
16
failures in the high-pressure melt eject, but in the
17
way that our containment liner is constructed, which
18
is different than what some of the other plants have
19
done in the past.
We don't have a freestanding shell with
20
21
concrete
outside
of
it.
Each
of
the
plates
are
22
actually anchored into the concrete.
23
flow path out around the liner if we get a small
24
hole.
So there is no
It's got to go through the entire container
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
114
1
itself.
2
that.
And we have sensitivities associated with
3
So this is one place in the report where
4
it gets difficult to understand what it was we were
5
trying
6
presented it.
to
9
at
mainly
because
of
the
way
Regime III is the expected regime.
7
8
get
This is basically what we expect to happen.
it's not highlighted in the report.
we
Okay?
And
What we expect
10
to happen isn't what's shown because we don't think
11
that we're going to have any interesting phenomena
12
there.
CHAIRMAN CORRADINI:
13
I'm sorry.
14
just remind us really briefly?
15
this only if we had a failure of?
16
remind me?
Can you
We would get into
Can you just
I'm sorry.
17
MR. WACHOWIAK:
Depressurization.
18
CHAIRMAN CORRADINI:
And that would be by
19
the squib, again by another set of depressurization
20
valves?
MR. WACHOWIAK:
21
22
We have depressurization
valves --
23
CHAIRMAN CORRADINI:
24
MR. WACHOWIAK:
Right.
-- that could provide it.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
115
1
We have traditional safety relief valves that could
2
also provide depressurization.
And then we also have
3
the isolation condenser system, which if it goes into
4
operation will also provide depressurization.
5
three of those systems would need to fail before we
6
get into high-pressure melt eject.
So all
And I think -- I am trying to remember
7
8
now.
9
percentage of the total core damage frequency falls
10
In our latest results.
into this range.
CHAIRMAN CORRADINI:
11
12
That is a fairly small
Okay.
Sorry.
Thank
you.
MR. WACHOWIAK:
13
So regime II is where we
14
spent most of our time looking at, finding bounding
15
parameters to address the way that we would load the
16
containment from a high-pressure melt eject.
17
What we have shown in the end with these
18
results is that the peak pressures in the containment
19
are well within the ultimate pressure that we talked
20
about earlier.
21
70 pounds in the dynamic peak because basically the
22
pressure
23
handle those sorts of dynamic pressurization.
We get about .6 megapascals, 6 bars,
suppression
containment
is
designed
to
The way that we calculated this -- and
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
116
1
it's
described
2
essentially a closed-form solution that looks at the
3
load
4
containment, which was parked against the IET tests
5
from
6
containments.
or
that
way
in
looks
back
report
at
when
Then
7
the
the
on
is
we
pressurization
the
regime
--
pressure
I,
which
have
in
the
suppression
we
call
8
hypothetical, should be regime H, but it's regime 1,
9
hypothetical.
We push the parameters on the model to
10
see what it would take to fail the containment.
11
in the end, we found that we could get up to the
12
place where containment would fail, but we would have
13
to use input parameters, like the timing of the melt
14
release or the amount of the melt or the rate of
15
ablation of the vessel, those sorts of things that
16
were outside of anything that we had seen in the
17
thermal data.
So we call that the hypothetical regime.
18
19
And
It
was
mainly
could
calculating a failure of the containment, even though
22
this
23
containment.
24
associated with the significant core melt sequences,
calc.
were
we
21
bounding
methods
that
demonstrate
our
our
so
20
is
that
calculated
It's
capable
not
of
failing
And in the actual cases that we ran
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
117
1
we don't even get anywhere near the bounding case.
MEMBER BLEY:
2
Just to tie a couple of
3
things together that you said before, if you actually
4
had
5
depressurization systems, --
this
scenario
where
you
6
MR. WACHOWIAK:
7
MEMBER BLEY:
failed
all
of
your
Yes.
-- to get to this point,
8
what is the earliest core melt you might get out of a
9
scenario like that?
Do you remember it?
10
MR. WACHOWIAK:
11
MR. SEEMAN:
12
I am not sure the earliest,
but I should have a pretty good --
13
MEMBER BLEY:
14
MR. WACHOWIAK:
15
Do you have it?
Rough time.
Yes.
It is a few hours,
but -MEMBER BLEY:
16
Okay.
So a scenario like
17
this wasn't considered a significant scenario because
18
the probability, the frequency of it is very low
19
because your experiments were done, you said, with a
20
six-hour decay heat because the significant scenarios
21
all had greater than six-hour --
22
MR. WACHOWIAK:
23
MEMBER BLEY:
24
Yes.
-- take time to melt.
So
this one would be earlier than that, but it doesn't
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
118
1
fit
in
that
category
of
2
understanding you correctly.
3
MR. SEEMAN:
4
MEMBER BLEY:
5
MR. WACHOWIAK:
6
ones.
significant
if
I
am
That was six hours.
This one was -They were the earliest
That's where we based the six hours on.
7
MR. SEEMAN:
Right.
8
MEMBER BLEY:
Okay.
9
DR. KRESS:
Are these map?
10
MR. WACHOWIAK:
11
MEMBER
Yes.
ABDEL-KHALIK:
What
phenomenon
12
would have such a short time constant that would give
13
you that peak at two seconds?
MR.
14
WACHOWIAK:
If
we
deposited
the
15
entire core into the drywell within just a couple of
16
seconds, we would end up having to start with a CRD
17
tube that came out.
And that four-inch hole would
18
need
a
19
like, I think it was, a couple of seconds was the
20
parameters that we got to iterate those types of
21
pressurization.
to
ablate
DR. KRESS:
22
23
into
one-meter-wide
hole
within
The containment is inerted
during this?
MR. WACHOWIAK:
24
Yes.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
119
DR. KRESS:
1
2
So you don't have to worry
about hydrogen iteration?
3
MR. WACHOWIAK:
Yes.
4
CHAIRMAN CORRADINI:
I think you want to
5
take a step back.
I think Said is asking, let's say
6
you have a hole, let's say you have the melt, let's
7
say you have a temperature.
8
that is pumping up the pressure inside containment to
9
these?
What is the mechanism
10
MEMBER ABDEL-KHALIK:
11
MR. WACHOWIAK:
12
MEMBER ABDEL-KHALIK:
13
MR.
Right and then --
Okay.
WACHOWIAK:
I'm sorry.
Right, right.
The
way
that
the
14
high-pressure melt eject, the DCH scenario, goes is
15
you start with the core material jets.
16
of core material out of the hole.
17
by a high-pressure steam jet.
You get a jet
And it's followed
Okay?
18
And the high-pressure steam jet fragments
19
and mixes with the core material as it's on the floor
20
and it drives that up the sides of the walls of the
21
lower drywell and disperses into the atmosphere on
22
the containment.
23
And as it's dispersing, the surface area
24
of the material in the melt is able to interact with
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
120
1
all of the gas that's in the -MEMBER ABDEL-KHALIK:
2
3
What I am trying to
understand, --
4
MR. WACHOWIAK:
5
MEMBER
Okay.
ABDEL-KHALIK:
--
all
of
the
6
things that you talked about sort of tend to drive
7
the pressure up.
8
this curve.
9
period of time?
So we're on the pressurized leg of
What brought it down for that short
10
MEMBER BLEY:
11
MEMBER ABDEL-KHALIK:
12
MR. WACHOWIAK:
13
MEMBER ABDEL-KHALIK:
14
The dip in the curve.
Right.
This part here.
Right.
Or on the
other case, what you call the nominal thing, --
15
MR. WACHOWIAK:
16
MEMBER ABDEL-KHALIK:
dominates
Here.
17
phenomenon
18
causes the pressure to turn around?
MR.
19
during
WACHOWIAK:
this
Vent
-- what physical
time
period
clearing.
that
The
20
pressure suppression containment is set up so that if
21
the water in the vents, covering the vents, pressure
22
goes up, it pushes that water down when the vent's
23
clear and the steam goes and we start involving the
24
suppression pool air space.
And then they recover
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
121
1
it.
It
becomes
2
participate in the -DR.
3
more
of
KRESS:
you
a
In
need
--
not
order
surface
all
to
make
this
or
drop
size
4
calculation,
5
distribution.
6
or is that pulled out of experiments on --
area
Does that come out of expert opinion
MR. WACHOWIAK:
7
events
I would have to go back
8
and look at where we got that, but it's probably a
9
combination of expert opinion and experiments because
10
I don't know that we have seen very many of these DCH
11
events.
CHAIRMAN CORRADINI:
12
13
I would guess it's
the tuning of the Sandia experiments.
14
DR. KRESS:
Sandia experiments.
15
MR. WACHOWIAK:
But I wouldn't exclude
16
expert elicitation in that as well because this is
17
one of these phenomena where that --
18
MEMBER BLEY:
19
that
is
colored
20
experiments.
by
If you did, you would hope
knowledge
21
MR. WACHOWIAK:
22
DR.
KRESS:
of
the
Sandia
Okay.
So
the
ROAAM
process
23
basically tells you how much melt and what picture it
24
is?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
122
1
MR. WACHOWIAK:
2
DR. KRESS:
3
Yes.
That maybe feeds out of the
primary vessel?
CHAIRMAN CORRADINI:
4
5
worried
6
depressurize, but I do think that this one part I
7
guess
8
"expert opinion" versus tuning, who are the experts
9
you are talking about?
we
about
this
need
to
given
I guess I am not
all
understand.
the
So
ways
when
you
you
say,
Was it a staff insight at GE
10
that got together and set the distribution curse for
11
the ROAAM calculation or was it the Santa Barbara
12
folks?
MR. WACHOWIAK:
13
Santa Barbara folks set
14
that up.
15
reviewed
16
comments are included in the report.
by
basically
nine
experts.
CHAIRMAN CORRADINI:
17
18
And then we took that report and had it
And
their
And that's in 21?
Did I miss that?
19
MR. WACHOWIAK:
In 21.
20
CHAIRMAN CORRADINI:
21
MR.
Okay.
WACHOWIAK:
22
correspondence
is
in
there.
23
reviewed it are in there.
Their
And
CHAIRMAN CORRADINI:
24
the
letters,
people
Thank you.
who
Thank
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
123
1
you.
Move on.
MR. WACHOWIAK:
2
Okay.
This is a simple
3
demonstration of how we decide these things in our
4
process,
5
something in the containment event tree.
6
from the DCH that we calculated is this line here.
7
Okay?
whether
And
8
or
the
not
we're
containment
going
to
include
The load
fragility
that
we
9
presented before, this is an earlier version of it.
10
Those two lines or two curves don't intersect in any
11
significant way.
12
failure due to DCH physically unreasonable for our
13
containment.
So this allows us to call energetic
And we don't treat it explicitly in the
14
15
main calculation for the level 2.
16
sensitivity that we have looked at.
17
are wrong?
18
this is a fairly simple, straightforward example of
19
how
20
unreasonable.
we
We do have a
Well, what if we
What does it do to us in the level 3?
apply
that
to
come
up
with
So
physically
21
Now, the next one is the fuel coolant
22
interaction that would generate a steam explosion.
23
We looked at ex-vessel steam explosions.
24
issue that we have here.
That's the
And the failure modes are
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
124
1
damaging the pedestal wall, which would get us a
2
release to the environment.
3
We
find
that
if
we
had
a
very
deep
4
subcooled pool of water in the lower drywell, that
5
that becomes an issue.
6
damaging the BiMAC pipes themselves because they can
7
be crushed, if you will, by an impulse load.
8
also find that deep subcooled pools will affect BiMAC
9
pipes.
Another issue is physically
And we
I didn't put very deep down because the
10
11
loading
on
the
12
evaluation, pipes are more fragile than the lower
13
drywell or at the wall.
The
14
pipes,
way
we
at
least
address
in
this
the
is
current
that
we
15
minimize the amount of water in the lower drywell
16
prior to the vessel breach.
17
up, do we want to have this BiMAC pre-flooded or
18
almost pre-flooded?
19
looked at that trade-off and decided that we don't
20
want to have that set up that way.
And the question came
And this is the area where we
CHAIRMAN CORRADINI:
21
So just to repeat
22
because I do remember you had covered this before.
23
When you say "deep," it's still below the equipment
24
hatch?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
125
MR. WACHOWIAK:
1
2
the
equipment
3
equipment hatch.
trying to remember.
MR.
6
7
still
is
up
to
the
Which is still?
I'm
Is that about three meters?
WACHOWIAK:
2.2
meters
is
the
equipment hatch.
CHAIRMAN CORRADINI:
8
9
Deep
Very deep is above
CHAIRMAN CORRADINI:
4
5
hatch.
Yes.
2.2.
And in these
calculations, I'm trying to -- I lost my Chapter 21
10
somewhere in here.
11
at asymmetric loading of the pedestal.
I was most
12
concerned
temperature
13
where it occurred relative to the wall.
14
MR. WACHOWIAK:
with
the
In your calculations, you looked
pour
rate
and
Yes.
the
I think in the
15
report, you can see we did two or three different
16
locations with respect to the wall.
17
and then a couple of off-center.
18
when we calculated the effect on the wall, we didn't
19
-- the wall section itself got the whole impulse.
Off-center
20
21
would be.
22
symmetric wall.
developed
We did a center
But, once again,
what
the
impulse
And then the calculation on the wall was a
23
CHAIRMAN CORRADINI:
24
DR. KRESS:
Okay.
Did you use a particular FCI
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
126
1
code for this model?
2
MR. WACHOWIAK:
3
DR. KRESS:
4
CHAIRMAN CORRADINI:
No, please.
5
DR.
thing
6
The
developed
in
I don't remember which
code we used to develop the FCI loads.
DR. KRESS:
There is a model in MAAP.
CHAIRMAN CORRADINI:
10
No.
It's the PM
alpha SPROS.
MR. WACHOWIAK:
12
13
KRESS:
MR. WACHOWIAK:
9
11
Texas?
Wisconsin?
7
8
Of course.
And then LSI for the wall
response.
CHAIRMAN CORRADINI:
14
Was it coupled or
15
was it -- so there was a pressure source term, and
16
then
17
response of the wall?
it
was
fed
to
essentially
18
MR. WACHOWIAK:
19
CHAIRMAN CORRADINI:
the
impulse?
What
a
finite
element
Yes.
the
And cracking
20
was
21
little dashed line in terms of the pedestal wall
22
failure there?
23
was the mechanism?
failure
that
the
Is that cracking of the wall or what
MR. WACHOWIAK:
24
was
Okay.
No.
It was reaching the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
127
1
strain limits on the rebar.
2
CHAIRMAN CORRADINI:
3
MR. WACHOWIAK:
Okay.
So yes.
We cracked the
4
concrete well before we would this front integrity of
5
the rebar.
6
CHAIRMAN CORRADINI:
7
MR. WACHOWIAK:
Okay.
Thank you.
And that's shown in the
8
-- you have to get the color version of the report to
9
really see what is going on there.
And then also be
10
aware that the boundary condition of the pedestal
11
wall was not -- the anchorage at the bottom wasn't
12
realistic.
13
So that first node that you would see
14
there for the strain in the rebar is not necessarily
15
what
16
boundary condition that was used.
we
would
expect.
DR. KRESS:
17
It's
just
because
of
the
In terms of the upper bound
18
load calculation, does that have anything to do with
19
how much metallic melt is assumed to be in with the
20
core melt?
MR. WACHOWIAK:
21
Right.
And there are
22
several cases that are in the report that we looked
23
at different compositions of the melt and different
24
core rates and things like that.
And this box here
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
128
1
is meant to envelope all of those cases.
DR.
2
3
KRESS:
The
specific
cases
that
follow.
MR. WACHOWIAK:
4
There would be different
5
cases along there.
6
got a one to two-meter subcooled pool and we have the
7
upper bound load, we start seeing from the LS DINA
8
calculations
9
kilopascal seconds is where we failed the rebar.
that
So what we find is that if we've
somewhere
between
400
and
600
So for these types of pools, we're not
10
11
expecting to fail the pedestal.
12
that in that same set of pools, this bounds out what
13
the impulse to the floor is.
14
least
15
calculation, that we start to see their incipient
16
failure rate within the loads that are generated by
17
those one to two-meter deep pools.
with
the
BiMAC
pipes
But what we see is
We do see that, at
that
we
use
in
this
So what we say is if we get more than .7
18
19
meters of water, this isn't a PRA now.
20
with
21
drywell.
22
then
23
going to fail by a pedestal failure or fail because
24
the BiMAC is -- the pipes are going to be damaged.
more
we'll
than
.7
meters
of
water
If we start
in
the
lower
When the floor is expected to come out,
assume
that
the
containment
is
either
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
129
DR. KRESS:
1
2
matter there?
3
different depth?
Does the amount of subcooling
If it were saturated, would you get a
MR. WACHOWIAK:
4
Yes.
We wouldn't see
5
very much at all.
In what the calculation shows, if
6
it's saturated, we don't see any impulse, essentially
7
the melt --
8
DR. KRESS:
9
MR.
10
WACHOWIAK:
--
calculates
itself.
And it's -DR. KRESS:
11
12
It just gets --
So if you're in a subcooling,
did you assume --
13
MR. WACHOWIAK:
14
DR. KRESS:
15
MR. WACHOWIAK:
Fifty degrees.
Fifty degrees?
In the calc, so quite a
16
bit of subcooling.
17
to get to that.
18
calculate the subcooling in the pools.
19
that was beyond our state of knowledge as to know
20
specifically what the temperature was in that water
21
in the drywell.
Now, once again, we didn't try to
We thought
So there --
MEMBER BROWN:
22
23
We're not even sure we're going
Is subcooling better or
worse?
MR.
24
WACHOWIAK:
Subcooling
makes
it
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
130
1
worse.
DR. KRESS:
2
3
You get real energetic
explosions the more subcooling you have.
MEMBER
4
5
Yes.
BROWN:
Okay.
for
the
ex-vessel?
6
MR. WACHOWIAK:
Yes.
7
MEMBER BROWN:
Okay.
8
CHAIRMAN CORRADINI:
9
That's
it gets back to your point.
And so I guess then
So every one of this
10
curve, the red curve, which is the load from the high
11
level, what was the subcooling in the core rate?
12
you remember?
14
I'm sorry I'm looking at --
MR. WACHOWIAK:
13
rate.
I don't know the core
The subcooling was 50 degrees K.
15
CHAIRMAN CORRADINI:
16
MR. WACHOWIAK:
17
Do
50 degrees K?
I remember that one.
I
don't remember what -CHAIRMAN CORRADINI:
18
Okay.
All right.
19
And then for the blue, it's more like 10-20 degrees
20
K?
21
MR. WACHOWIAK:
22
CHAIRMAN CORRADINI:
23
MEMBER
24
right?
BROWN:
Okay.
Okay.
Fifty
degrees
Kelvin,
That's what you --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
131
1
MR. WACHOWIAK:
Yes.
2
MEMBER BROWN:
Okay.
3
large number for subcooling.
4
conservative, I guess?
So that's such a
In this case, that's
Is that?
5
DR. KRESS:
6
it's a pretty good subcooling, yes.
MEMBER BROWN:
7
8
Yes.
Tom?
Anybody?
It's a pretty low --
And you wouldn't expect
that much under this scenario?
MR. WACHOWIAK:
9
The water itself is --
10
there are two ways that water can get into the lower
11
drywell essentially in these cases, is that it came
12
out of the reactor.
13
saturated at 1,000 pounds when it came out of the
14
reactor.
15
good subcooling in the lower drywell.
So 50 degrees subcooling would be pretty
The other cases where it's condensed on
16
17
the wall -MEMBER
18
19
So it was already started out at
BROWN:
It
would
be
pretty
conservative is what you --
20
MR. WACHOWIAK:
21
MEMBER BROWN:
22
MR. WACHOWIAK:
Yes.
It wouldn't get to that.
It wouldn't get to that
23
point.
And the other place, though, is condensation
24
on
wall
the
that
would
run
down
into
the
lower
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
132
1
drywell.
And we're thinking that that's going to be
2
fairly close to saturation because of that mechanism
3
for getting the water down there.
4
subcooling we think was bounding.
So 50 degrees
But, once again, once we got to the -- we
5
6
didn't really use that.
7
what
8
calculation
9
going to assume that it's subcooled and that we'll
10
the
temperature
shows
We didn't try to calculate
was.
that
We
the
just
water
is
said
if
there,
our
we're
have the steam explosion.
11
We didn't try to cut it that fine because
12
that gets beyond what we think that we actually can
13
know at this point in time.
CHAIRMAN CORRADINI:
14
So, just to repeat
15
because you have heard of this, where would be the
16
failure of the tube?
17
in
18
intersection
19
versus the deformation.
the
tube
along
of
It would be in the buried tube,
the
the
wall,
red
that
line,
blue
which
line,
the
the
load
is
That's where.
20
MR. WACHOWIAK:
Fairy tube.
21
CHAIRMAN CORRADINI:
22
MR. WACHOWIAK:
Fairy tube.
And I don't believe we
23
took into account any of the floor material on top of
24
the tube.
It was just the strength of the pipes
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
133
1
themselves that was considered here.
CHAIRMAN CORRADINI:
2
Okay.
So let me
3
ask, then, my question.
4
that I would most worry about was the exposed tube on
5
the
6
buried tube underneath the concrete.
wall
that's
It would seem to me the tube
connected
to
the
water,
not
Again, I've got this in my head.
7
the
I've
8
got 12 feeder tubes, 6 of which go to the top of the
9
water, 6 of which go below.
And it's those tubes,
10
those six tubes, that feed the header that then boil
11
back
12
exposed to the water where I have a drop of melt,
13
which I have an explosion which crushes those tubes.
up
worry
me
the
most
because
they're
I thought that was the blue line.
14
15
that
That's
not the blue line?
16
MR. WACHOWIAK:
The pictures that are in
17
the report show the horizontal tubes.
18
we'll have to take that bask as a question to see how
19
we address the -- now, once again, remember that the
20
vertical tubes for the most part are going to be
21
buried in some sort of a material.
CHAIRMAN CORRADINI:
22
So I guess
But at some point
23
they have to pop out so water can pop in.
So that's
24
the point I'm worried about is I've got these feeder
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
134
1
tubes to the downcomer and I'm worried about them
2
being crimped off by some further explosion.
3
MR. WACHOWIAK:
Right.
4
CHAIRMAN CORRADINI:
5
MR. WACHOWIAK:
Yes.
I understand the feeder
6
tubes.
7
there and they're buried most of the way --
8
CHAIRMAN CORRADINI:
9
MR.
10
The sidewall tubes, you know, if they're
WACHOWIAK:
I'm there.
--
the
impulse
isn't
going to be doing anything to those --
11
CHAIRMAN CORRADINI:
12
MR. WACHOWIAK:
Yes.
-- the large downcomer
13
tubes.
That's a good question.
14
that we address that in the report.
15
see if it's in there.
Once
16
again,
I'm with you.
that
And I don't know
can
We will look to
be
addressed,
17
though, too, since it's not everything.
18
localized area, we can do things to minimize the
19
impulse on that pipe in the detailed design.
CHAIRMAN
20
21
experiments
22
They're what they call an MFMI event.
23
have gotten data on that.
24
run
by
CORRADINI:
AECL
for
their
Since it's a
There
are
pressure
tube.
They actually
So you might want to look.
There is actual experimental data literally looking
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
135
1
for this sort of effect on a pressure source outside
2
of an array of tubes in a water chamber.
DR. KRESS:
3
4
The horizontal tubes, do they
fail in tension due to just mashing on them?
5
MR. WACHOWIAK:
6
DR. KRESS:
7
The sides fail in tension
because of the -MR. WACHOWIAK:
8
9
Pressing, yes.
Yes.
I think that's the
-- in the report, it shows where the strains are
10
calculated in the pipes.
11
mechanism.
Yes, that crushing is the
12
CHAIRMAN CORRADINI:
13
MR. WACHOWIAK:
Okay.
Okay.
Thank you.
Now we'll move on
14
to containment overpressurization in the long term.
15
We've gotten through these short, energetic things,
16
what could happen to the containment in the long
17
term.
We have systems that mitigate this.
18
This
19
is getting back to more traditional PRA.
20
passive
21
steam that's carrying the heat, we can condense that
22
steam and move the heat outside the containment.
containment
cooling
system,
that
We have a
if
it's
23
Implicit in that is the vacuum breakers
24
that they provide that seal between the drywell and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
136
1
the wetwell.
2
Committee has had many presentations on the vacuum
3
breaker.
So I don't intend to go over any of that
4
5
stuff
6
breakers.
now.
Just
we
took
credit
CHAIRMAN CORRADINI:
7
8
I know that at least most of this
MR. WACHOWIAK:
the
backup
valves
the
vacuum
We don't have our
skeptic consultant with us anyway.
9
for
So let's move on.
If the vacuum breakers
10
and
fail,
then
we
assume
11
containment will overpressurize.
12
done using the fault tree systems in the PRA.
the
That's how that's
We have an active RHR system that we also
13
14
take credit for in the PRA.
And then we do have
15
venting.
16
considered to be a large release.
Once again, in our PRA, venting is still
17
The only thing that the venting does is
18
it changes the way that the source term is addressed
19
in the level 3.
20
design PRA, considered to be a large release and
21
factored into that containment performance.
So all venting is, at least in the
We treated these using a fault tree/event
22
23
tree method.
And it's linked.
24
linked directly.
Level 1, level 2 are
We talked about that in June.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
137
CHAIRMAN CORRADINI:
1
So at this point, I
2
think unless the members -- I'm going to jump in
3
here.
4
might be the point to close it if there are questions
5
about the heat transfer experiments.
Unless the members are in disagreement, this
6
MEMBER BROWN:
The BiMAC.
7
CHAIRMAN CORRADINI:
The BiMAC.
Going
8
back to the BiMAC heat transfers, which have a lot of
9
design detail, we'll have to close the session.
10
that all right?
11
federal officer to tell me how I do that.
12
do that?
MR.
13
14
recess.
Is
Now we'll look to the designated
VANDER
MOLLEN:
You
How do I
call
a
short
And we ask everyone who is not either --
15
CHAIRMAN CORRADINI:
Do we check IDs?
16
MR. VANDER MOLLEN:
Well, we're not far
17
from
it.
People
18
members, who go to jail if they say something of
19
proprietary stuff; or the applicant; and anyone who
20
has
21
applicant.
22
personnel
23
unauthorized to be here.
executed
a
And
to
who
may
stay
proprietary
I
tell
am
me
going
if
either
agreement
to
there
CHAIRMAN CORRADINI:
24
are
defer
is
with
to
the
anybody
Okay.
staff
who
the
GEH
is
So are there
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
138
1
any open session questions at this point about the
2
severe accident management discussion that Rick has
3
provided?
4
(No response.)
5
CHAIRMAN CORRADINI:
If not, let's take a
6
short, very short -- don't leave the room except for
7
a quick bathroom -- break.
8
All right?
9
(Whereupon, the foregoing matter went off the record
10
at 11:15 a.m., to reconvene in closed
11
session, and reconvened in open session
12
as follows at 11:42 a.m.)
13
CHAIRMAN CORRADINI:
14
11:25 we'll come back.
Go ahead, Ed.
We're
in open session.
EVALUATION OF SEVERE ACCIDENTS
15
MR. FULLER:
16
What I am doing here today
17
is a continuation of what you heard in June from Mark
18
Caruso going through the review of chapter 19.
19
severe
20
Electric prepared is in section 19.3 of the second
21
tier document of the DCD.
accident
evaluations
piece
that
The
General
In our review, we followed the standard
22
23
review plan.
And we used section 19.2 to denote
24
severe accident evaluation.
So I just wanted to make
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
139
1
sure
you
understand
2
later.
3
actually review that and prepare the SER.
The
that
so
of
this
purpose
you're
not
confused
presentation
is
to
4
I am going through three topics just to
5
show you without even going through in any detail the
6
applicable regulations that we use and just show you
7
the SER technical topics, which are nothing more than
8
major section heads in section 19.2, and to just
9
briefly discuss the very few significant open items
10
that we have got.
Next slide, Rocky.
11
We have got a number
12
of regulatory requirements on severe accidents that
13
are in 10 CFR 52.
MR. FOSTER:
14
Ed, if we could possibly
15
move along on this slide because we presented this on
16
June 3rd?
MR.
17
FULLER:
There
was
one
I
added,
18
though, pertaining to the severe accident mitigation
19
design alternatives that relates to the NEPA as well
20
as to our actual FSAR review.
CHAIRMAN CORRADINI:
21
22
This is out of the
10 CFR 51.55, Ed?
23
MR. FULLER:
That's correct.
24
CHAIRMAN CORRADINI:
Okay.
I just wanted
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
140
1
to point that out.
Everything else in this slide is
2
the same as what appeared in the package.
MEMBER APOSTOLAKIS:
3
PRA"
mean,
"Regulatory
What does the "use
4
of
5
statements on severe accidents and use of PRA"?
6
MR. FULLER:
7
MEMBER
8
Where?
APOSTOLAKIS:
MR. FOSTER:
Right
under
The first bullet underneath
the second one.
MR. FULLER:
11
12
Policy
"Regulatory Guidance."
9
10
Guidance.
Oh, the policy statements
that appear back in the early '90s.
MEMBER APOSTOLAKIS:
13
Oh, that means use
14
PRA to the extent supported by the state-of-the-art,
15
--
16
MR. FULLER:
17
MEMBER APOSTOLAKIS:
18
Yes.
Next slide.
-- which is a pretty
general blanket statement.
MR. FULLER:
19
course.
We
are
Okay.
supposed
Now let's get into
20
the
to
evaluate
severe
21
accident prevention and severe accident mitigation
22
features.
23
give our evaluation.
And so the first two sections of an SER
We had no open items on severe accident
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
141
1
prevention.
We
2
mitigation, which I will get to.
3
on
4
accident management and nothing on, no open items on,
5
severe accident mitigation design alternatives.
containment
had
a
couple
performance
on
severe
accident
And we had a couple
capability
and
one
on
6
Note here that I have actually put on the
7
slide that document that I referred to earlier today
8
to aid in your finding it.
9
at the SAMDA submittal.
Okay?
MEMBER APOSTOLAKIS:
10
You want to look
So let me understand
11
that.
Let's go back since you want to show me.
12
19.2.6.
13
of ways for mitigating the severe accident, what this
14
means.
So we expect the applicant to show a number
MR. FULLER:
15
when
you
What he has to do here is --
16
obviously
are
producing
a
design,
an
17
advanced design, to be certified, there are a lot of
18
features for mitigating severe accidents that are not
19
in existing plants.
20
In addition, the NEPA requires you to
21
look at other severe accident mitigation alternatives
22
and do a cost-benefit on whether or not they is a
23
case for including them.
24
go through the process.
And so the applicant has to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
142
And
this
NEDO
2
report, is GEH's document to comply with this.
And,
3
needless to say, they didn't find anything that was
4
in addition to what they have already got that was
5
cost-beneficial.
1
MEMBER
6
7
this
particular
report,
APOSTOLAKIS:
So
there
no
evaluation of an alternative to the BiMAC?
8
MR. FULLER:
9
MEMBER
No.
APOSTOLAKIS:
Was
10
evaluation of an alternative to anything?
11
no I think when -MR. FULLER:
12
13
was
there
an
You said
Not to the BiMAC.
Is the
answer --
14
MEMBER APOSTOLAKIS:
15
MR. FULLER:
Or to anything else.
Yes, there was some, but I
16
can't really specify what they are, you know, right
17
today.
18
with you.
19
report to review, but it's been out for a year.
I would have to get the report and go over it
I am sorry that you guys never got the
MEMBER ABDEL-KHALIK:
20
21
answer
that
22
evaluated?
question?
What
Can the applicant
alternatives
were
23
MR. MILLER:
This is Gary Miller, GEH.
24
I think the scope of that was to evaluate
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
143
1
alternatives above and beyond what was currently in
2
the ESBWR design.
3
were
4
evaluations from other plants.
based
on
an
We
5
So we screened alternatives that
industry
used
that.
search,
We
used
other
SAMDA
ABWR
SAMDA
6
evaluation, anything we could to generate a list of
7
potential alternatives that are not currently in the
8
ESBWR design.
9
their merits.
And then we screened those based on
MR. FULLER:
10
Okay.
Let's go to the next
11
slide.
The first significant open item has to do
12
with the BiMAC performance test report.
13
whole lot of RAIs on the BiMAC during the course of
14
the review.
15
number of them; in particular, the two listed here,
16
resolution awaited the results of the BiMAC tests and
17
the review of the documentation of them.
We asked a
Quite a few of them got resolved, but a
18
We got that report in the May time frame.
19
We did a very quick, intense review and generated 27
20
RAIs.
And,
as
a
result
of
this
discussion
21
morning, maybe a couple of more will be generated.
So this is an ongoing open item.
22
this
The
23
review focused on several major areas:
adequacy of
24
the test facility, scaling and its applicability to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
144
1
the ESBWR, the range of test data as compared to what
2
one might expect in severe accident loadings, the
3
adequacy of the theoretical predictions as compared
4
to
5
operational ESBWR safety.
the
data
and
And
6
implications
we
found
7
generally
8
but
9
multi-channel tests.
we
do
adequately
have
an
Regarding
10
of
that
the
the
test
scaled
prototypic
RAI
scaling
the
on
range
of
design
on
facility
conditions,
basis
of
measured
the
test
11
data, we focused on the perceived lack of relevant
12
tests for near-edge tubes.
13
treatment of the range of heat fluxes chosen for the
14
tests.
15
are written on.
We wanted to see a better
And there were some other issues that RAIs
16
Comparing theoretical predictions against
17
the data, our contractor did a little assessment for
18
the
19
predictions are supporting the measurements.
single-tube
tests.
And
it
seems
like
the
20
And, finally, the implications on ESBWR
21
operational safety, we have some RAIs on thermal load
22
boundary conditions; the use of CFD simulations to
23
obtain boundary conditions; and, taking cue from Dr.
24
Powers, who is not here today, asked questions on the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
145
1
structural integrity of zirconia; and we also asked
2
about the effects of crusts on heat loads and some
3
other issues as well.
4
majority of the 27 RAIs that are in those 4 areas.
So that captures the vast
5
Next slide.
Another significant --
6
CHAIRMAN CORRADINI:
Can I interrupt you
7
just to ask one thing?
8
calculations under the third sub-bullet, "Adequacy of
9
the theoretical," what was being used?
MR.
10
So when you are doing these
FULLER:
It
is
a
model
that
11
Khatib-Rahbar has put in place many years ago.
12
ask Mohsen to address that?
DR. KHATIB-RAHBAR:
13
All right.
Dr.
Can I
This is a
14
very simple one-dimensional pressure drop calculation
15
essentially for low pressures, where you're driving
16
the flow through this, the head you're providing by
17
the liquid column.
18
the tubes.
19
the range of that.
And you're just finding the stability of
CHAIRMAN CORRADINI:
20
21
And it's basically coming out of
So this is given a
heat load, you're doing a natural circulation?
22
DR. KHATIB-RAHBAR:
23
CHAIRMAN CORRADINI:
24
DR.
Precisely.
So did you guys --
KHATIB-RAHBAR:
This
is
the
same
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
146
1
thing which was done by General Electric, by the way,
2
as well.
3
look
CHAIRMAN CORRADINI:
Okay.
at
of
the
prediction
But did you
4
guys
5
distribution in comparison to what GEH was proposing
6
or -DR. KHATIB-RAHBAR:
7
the
heat
Distribution?
flux
Do you
8
mean axially or distribution you mean in terms of
9
what?
CHAIRMAN
10
11
CORRADINI:
Along
the
tube
length and along the wall.
DR. KHATIB-RAHBAR:
12
Yes, we did that.
We
13
actually used a non-uniform heat distribution along
14
the pipes because it's a single tube case.
15
single tube, the heat flux along the pipe, which
16
tried to simulate the same thing, which was done in a
17
GE -CHAIRMAN CORRADINI:
18
but
I
guess
I'm
Right.
asking
That I'm
19
with,
20
different.
21
Did you check their heat flux distribution?
22
words,
23
distribution of some shape.
something
slightly
I'm saying you were using their input.
they're
saying
they
expect
the
In other
heat
flux
Did you recheck that?
DR. KHATIB-RAHBAR:
24
So for a
No, no.
The GE heat
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
147
1
flux distribution is based on the CFD analysis.
CHAIRMAN CORRADINI:
2
3
Okay.
Which they
have done.
4
DR. KHATIB-RAHBAR:
5
We have not done a molten core analysis.
CHAIRMAN CORRADINI:
6
7
DR. KHATIB-RAHBAR:
CHAIRMAN CORRADINI:
11
DR. KHATIB-RAHBAR:
This is just a single
Okay.
It's heat flux-driven
analysis.
13
CHAIRMAN CORRADINI:
14
MEMBER ABDEL-KHALIK:
15
is
16
non-uniformly
17
steady state calculations --
for
a
single
tube,
heated.
Okay.
whether
And
you
18
DR. KHATIB-RAHBAR:
19
MEMBER ABDEL-KHALIK:
20
That was I
tube.
10
12
Okay.
guess my question.
8
9
Which they have done.
Thank you.
But this analysis
it's
do
a
uniformly
or
sequence
of
Precisely.
-- to find the '05
--
21
DR. KHATIB-RAHBAR:
Precisely.
22
MEMBER ABDEL-KHALIK:
Yes, yes.
But there is no
23
calculation whereby you have a group of pipes that
24
are not geometrically identical --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
148
1
DR. KHATIB-RAHBAR:
2
MEMBER
3
Correct.
ABDEL-KHALIK:
--
that
are
in
parallel.
DR. KHATIB-RAHBAR:
4
That's correct.
Had
5
we done a single tube analysis only, the issue of how
6
we challenge stability is not considered if that is
7
what you are concerned about.
In
8
those,
I
think
there
are
some
9
discussions in the GE report that they talk about
10
that I think in passing, but that is something we
11
have not addressed.
It is just a single tube, pressure drop
12
13
analysis.
MEMBER
14
15
You are absolutely correct.
ABDEL-KHALIK:
DR. KHATIB-RAHBAR:
17
a steady state problem, though.
18
MEMBER
ABDEL-KHALIK:
DR. KHATIB-RAHBAR:
Right,
But it is
right,
a
And that's where the
problem is.
22
MEMBER ABDEL-KHALIK:
23
MR. FULLER:
24
Precisely.
sequence of steady states.
20
21
state.
Steady state.
16
19
Steady
please.
Right, right.
Let's go to the next slide,
There was another significant open item that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
149
1
is now cleared up on vacuum breaker performance.
2
were concerned about vacuum breaker leakage.
3
has provided information on isolation valves on the
4
drywell side of the drywell-wetwell interface to show
5
that, in addition the vacuum breakers, you had the
6
isolation valves, which would close in appropriate
7
ways
8
leakage and loss of pressures suppression capability.
to
reduce
MR. FULLER:
vacuum
breaker
Just to be sure,
I'm sorry.
No.
It's on the
drywell side.
13
CHAIRMAN CORRADINI:
14
MR. FULLER:
15
of
And GEH
it's on the drywell side or on the wetwell side?
11
12
probability
CHAIRMAN CORRADINI:
9
10
the
We
Okay.
Maybe GE can elaborate a
little bit.
CHAIRMAN CORRADINI:
16
I remember it being
17
on the wetwell side, but, again, I could be goofy.
18
This is not the forum.
19
MR. WACHOWIAK:
20
CHAIRMAN
21
isolation valve:
22
side of the vacuum?
24
CORRADINI:
Where
is
the
on the drywell side or the wetwell
MR. WACHOWIAK:
23
State the question again.
The isolation valves are
on the wetwell side of the vacuum breaker.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
150
MR.
1
2
I'm
not
sure
about
the
drawing, then.
MR.
3
4
FULLER:
WACHOWIAK:
But
they're
in
the
drywell.
MR. FULLER:
5
Oh, yes.
I'm sorry.
I
6
meant to say they were -- yes, they were in the
7
drywell between the structure between the tube and
8
the vacuum breakers themselves.
MR. WACHOWIAK:
9
10
vacuum breaker.
11
MR. FULLER:
12
CHAIRMAN
13
Except it goes into the
knock you off base.
Right.
CORRADINI:
I
don't
want
to
Keep on going.
MR. FULLER:
14
Okay.
Next slide.
Here is another
15
case of an open item that was in existence when we
16
prepared the SER with open items, which has since
17
been resolved, having to do with the liner strain
18
exceeding
19
metal/water reaction conditions.
Level-C
limits
under
100
percent
And temperature boundary conditions for
20
21
the drywell head was set incorrectly.
And apparently
22
that has been corrected.
23
here, who did this particular part of the review, in
24
case you have any questions.
Jim Xu of the staff is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
151
Okay.
1
Hearing none, let's go to the next
2
slide.
The final open item has to do with accident
3
management,
4
accident
5
low-probability,
6
and the BiMAC but more in terms of how they're in the
7
process
8
management guidelines because we can do all of the
9
review of severe accident evaluations we want, but
10
the real reason for doing this is to make sure that
11
the plants have in place procedures and training to
12
handle these accidents if they have them.
not
in
the
management
of
way
earlier
in
potential
developing
GE
was
talking
terms
of
the
high-consequence
their
about
severe
very
events
accident
13
So looking downstream to the actual COL
14
licensees and holders, they will have to have good
15
accident management procedures in place.
Given
16
before
we
that
we
19
need
20
particularly
21
plants.
relative
to
COL,
is
understand for these new kinds of reactors, that we
the
a
it
18
what
give
that
important
understand
we
believe
17
to
that
that,
technical
that
for
the
basis
is,
existing
22
So we have been asking RAIs and back and
23
forth between COL applicants and GE in this case.
24
And we're trying to get them to the point where they
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
152
1
will
give
us
2
appropriate manner.
So
3
the
information
there
have
we
need
been
a
in
an
number
of
4
supplements to this basic accident management RAI.
5
And
6
basically comes down and asks, what is your technical
7
basis
8
hoping we get a decent reply this time.
the
for
latest
ESBWR
one,
accident
CHAIRMAN
9
which
GE
has
management?
CORRADINI:
I
10
appreciate everything you just said.
11
try a different way?
MEMBER APOSTOLAKIS:
12
13
just
am
gotten,
And
not
we're
sure
I
Can you kind of
I am confused.
Is
this an issue of a design certification or of COL?
MR.
14
FULLER:
It
is
an
issue
of
COL.
15
However, if you go back into history, you find that
16
when
17
management guidelines, it did it through a pretty
18
structured process that began by having EPRI develop
19
the technical basis for dealing with all of these
20
severe accident phenomena and how you start bringing
21
them towards procedures.
22
owners' groups were involved.
the
industry
formulated
severe
accident
And NEI was involved, and
23
So what happened was at a certain point,
24
that technical basis got transferred to the various
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
153
1
owners'
groups.
2
plant-specific guidance or specific to their kinds of
3
plants,
4
plant-specific guidelines.
which,
in
And
turn,
they
were
developed
developed
the
into
the
Okay?
5
And there are no actual regulations on
6
this, but there is an agreement between the NEI and
7
the NRC, which basically led to industry initiative,
8
if you will, that all of the plants agreed to comply
9
with.
And
it's
summarized
in
NEI
91-04,
which
10
includes some correspondence between the NRC and the
11
NEI at the time.
12
And so we're taking this as precedent to
13
go forward with the new plants because we want to
14
make sure that the same kinds of processes are in
15
place for the new plants.
MEMBER APOSTOLAKIS:
16
17
but why now and not at the COL?
MR. FULLER:
18
19
22
23
Because take, for example,
ESBWR.
MR. OESTERLE:
20
21
I understand that,
for a second.
Ed, let me jump in here
This is Eric Oesterle from the staff.
I am going to start at the endpoint with the COL
applicants or the feature licensees.
With respect to the NRC issuing them a
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
154
1
license, they have to demonstrate to us that they are
2
the ultimate responsible authority for implementing
3
the operational programs, the operational procedures,
4
including operating procedures, emergency procedures,
5
severe accident guidelines and procedures.
6
So it is the COL applicant that has the
7
burden of responsibility from the staff point of view
8
to provide that information to us.
9
technical
basis
10
procedures
and
11
design certification applicant.
for
developing
guidelines
However, the
all
really
of
rests
those
with
the
12
So there is going to be a lot of dialogue
13
and interaction between the COL applicants and the
14
applicants for design certification so that the COL
15
applicants
16
procedures for review.
can
It
17
provide
is
all
us,
the
part
of
staff,
what
with
we
these
consider
18
operational programs that we look at under chapter
19
13.
20
of the COL application.
And it has connections to various other portions
Now,
21
the
reason
the
burden
of
22
responsibility is on the COL is because they will
23
ultimately operate the plant, not GEH.
24
designers
of
the
plant.
And,
yet,
They're the
they
have
a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
155
1
tremendous amount of input into these procedures that
2
the COL applicant is responsible for.
MEMBER APOSTOLAKIS:
3
4
We've done all this
at the COL.
MR. FOSTER:
5
option
to
6
the
7
applicants and determine when you want to provide it,
8
but we need it before we go through licensing.
MR.
9
GE.
We could have, but we gave
Okay.
FULLER:
Talk
Let
me
with
the
explain
COL
why.
10
Particularly for a plant like ESBWR, there's a lot of
11
-- what we perceive is a plant is not going to behave
12
the same in many ways as the existing plants because
13
we've got severe accident mitigation features.
Timings are going to be different.
14
And
15
strategies will likely be different in any cases.
16
And if we don't get those identified now and give
17
somebody a COL and then find out later that we didn't
18
understand the technical basis, then it's hard to
19
resolve after the fact.
20
give COLs.
So we want it done before we
21
And it makes sense to have the designer
22
involved because they have done all of the severe
23
accident work already to understand their plant.
MR. OESTERLE:
24
Just one more data point
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
156
1
for this discussion again.
2
from the staff.
This is Eric Oesterle
After we develop all of these technical
3
4
bases, then we have to establish regulations.
5
the regulations with respect to design certification
6
state that once you receive a certification of the
7
design,
8
regulations
9
aspects
10
finality.
11
holder.
that
design
has
the
exception
with
for
that
design.
finality
of
That
Well,
under
the
the
operational
does
not
have
That finality rests with the COL license
Although the technical basis provides the
12
13
foundation
14
regulations, that is one of the other very important
15
reasons why it is up to the COL applicant to do this.
all
of
these
MR. FULLER:
16
17
for
presentation.
Okay.
CHAIRMAN CORRADINI:
19
(No response.)
20
CHAIRMAN CORRADINI:
for lunch.
the
That concludes my
Members?
Okay.
Let's break for lunch.
Let's adjourn
1:00 o'clock.
22
MEMBER APOSTOLAKIS:
No.
23
CHAIRMAN CORRADINI:
Please.
24
and
Any more questions?
18
21
discussions
(Whereupon, the closed session was concluded at 12:08
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
157
p.m.)
1
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
158
A-F-T-E-R-N-O-O-N
1
S-E-S-S-I-O-N
(1:01 p.m.)
2
CHAIRMAN CORRADINI:
3
4
Let's resume for our afternoon session.
INTRODUCTION
5
MR. WACHOWIAK:
6
7
All right.
CHAIRMAN CORRADINI:
9
MR. WACHOWIAK:
All tomorrow.
-- all tomorrow --
CHAIRMAN CORRADINI:
10
ourselves.
But we want to pace
We don't want to get behind.
12
MR. WACHOWIAK:
13
CHAIRMAN CORRADINI:
14
So this next
session, which is expected to go today and then --
8
11
On time at 1:05.
Okay.
And you're not going
to present very much.
MR. WACHOWIAK:
15
My intention is not to
16
have a lot of presentation.
17
front
18
because
19
questions before on what it was you were reviewing
20
and what is the purpose of all of this stuff.
21
just want to make sure that we all understand what we
22
have today, what we are going to have in the future,
23
and where that is going to be.
to
put
I
things
want
to
in
make
So go ahead.
24
I have a few slides up
context.
sure.
Then
There
we'll
were
go
some
And I
Everybody saw that picture
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
159
1
before.
I've had that 100 times.
2
CHAIRMAN CORRADINI:
3
MR. WACHOWIAK:
Sure.
This is our team that we
4
have here today.
You know me, Rick.
5
you've met before.
6
PRA area; Glen, who was up here earlier this morning,
7
one of our PRA engineers.
8
and Justin a PRA engineer.
He's principal engineer in the
Jonathan is a PRA engineer
Lou Lanese is somewhere.
9
Gary Miller
You've seen
10
Lou.
Oh, there he is back in the back.
11
regulatory affairs contact, make sure that we don't
12
have to go to jail like these guys do.
13
(Laughter.)
14
CHAIRMAN
15
CORRADINI:
So
He's our
he's
your
designated federal jail-server?
16
(Laughter.)
17
MR. WACHOWIAK:
the
board
Brandon
And also I didn't have up
18
on
Schaffer.
19
engineer for the ESBWR design cert.
20
to management at this point.
He's
a
project
So he's our link
Going down through the list, down the
21
22
next thing, here's what I think we're doing.
23
What you guys want to do is get an understanding of
24
the technical quality of the PRA.
Okay?
Okay?
And where
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
160
1
we were left before, you had some open questions and
2
you wanted to look at some more detailed things.
3
this is our forum or our opportunity for doing that.
4
And
Okay?
5
Look at the completeness that we have.
6
And then we really want to investigate the details of
7
what is going on in the PRA.
Okay?
PRESENTATION OF SELECTED PRA ACCIDENT SEQUENCES
8
MR. WACHOWIAK:
9
And the way we are going
10
to do this is we have got four sequences that you
11
suggested we amended.
12
context for this.
But those sequences are an entry
13
point into this.
It's not that you were interested
14
in those particular sequences.
And we'll use those for the
Okay?
Go down to the next one.
15
I just have a
16
couple of things here that talk about the quality and
17
scope of the PRA and then put it in the context for
18
the design certification.
The first is a partial quote from reg
19
20
guide 1.200.
21
that you do in the context of how you're going to use
22
it.
MEMBER ABDEL-KHALIK:
23
24
And you really need to look at any PRA
I'm sorry, but we
do not have copies of these slides.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
161
MR. WACHOWIAK:
1
You don't?
This was the
2
last one that I sent to Lynn yesterday morning.
3
she not get copies out?
MEMBER ABDEL-KHALIK:
4
5
She said there's
nothing in the box.
MR.
6
WACHOWIAK:
We'll
7
there are only a few slides here.
8
copies of this.
9
the car on the way here.
to
get
--
We'll have to get
We'll need them for
the record.
MR. WACHOWIAK:
12
13
have
Apparently they didn't make it in
MR. VANDER MOLLEN:
10
11
Did
I have got them on here.
I've got the .pdf file on here.
We had printed hard
14
copies before they left, but apparently they didn't
15
make it in the box.
So I apologize for that.
So we want to look at the PRA in context
16
17
of what it is we're doing with the PRA.
18
the past I think there is some thought that there is
19
a
20
anything.
plant
PRA
that's
general
and
can
In general, that is right.
21
Okay?
be
So in
used
for
There is a
22
framework that is there.
But you always have to
23
tailor the PRA to what it is you're going to use it
24
for.
In our case here with the design, we were using
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
162
1
the
PRA
to
2
application.
support
the
design
certification
Okay?
ISG, which has just come out on the PRAs
3
4
for
combined
construction
operating
licenses,
5
basically says that ASME capability category I is the
6
metric that we're looking for, which is a PRA that
7
can discern, can find vulnerabilities and can discern
8
importances at the system/train level.
9
So many of the -- if we get to a point
10
where we're going into more details than what we
11
have, it's because this application of the PRA wasn't
12
intended to do that.
13
On to the next one.
14
MEMBER APOSTOLAKIS:
15
MR. VANDER MOLLEN:
17
MR. WACHOWIAK:
Oh, that's on the
Web site.
MEMBER
BLEY:
It
has
not
come
to
us
directly.
MEMBER APOSTOLAKIS:
23
24
I got it off your Web
MEMBER APOSTOLAKIS:
21
22
Apparently not.
site.
19
20
Harold, do
we have this ISG?
16
18
Let me.
ISGs in general do
not come for approval, but I would like to have a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
163
1
copy.
MR. FOSTER:
2
3
It should be available on
the public Web site.
4
MR. WACHOWIAK:
5
MEMBER APOSTOLAKIS:
6
That's where I got it.
We never go to the
public Web site.
MEMBER ABDEL-KHALIK:
7
Back to the comment
8
that you just made, and it's at the top.
9
not go through a process by which you identified
10
So you did
single-point vulnerabilities for your design.
MR. WACHOWIAK:
11
Yes, I think we did go
12
through that.
In the process of building the PRA, we
13
were
for
14
along the way.
15
lot of that in the final analysis is because when we
16
found
17
vulnerabilities.
18
design.
looking
them,
single-point
vulnerabilities
And the reason why you don't see a
we
eliminated
the
single-point
That wasn't allowed to stay in the
So you don't see things like that.
19
all
You
20
don't see a specific search for that in the final
21
analysis because we were removing those as we went
22
along.
MEMBER
23
24
ABDEL-KHALIK:
Yes,
I
can
understand that, but, you know --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
164
MR. WACHOWIAK:
1
Our point was to identify
2
vulnerabilities and eliminate them.
3
the uses of the design PRA.
4
now since we don't have a piece of hardware that we
5
have to change.
We have that flexibility
MEMBER APOSTOLAKIS:
6
That's one of
I remember we had
7
the long discussion last time on the evolution of
8
PRA.
9
support this design, that doesn't mean when the COL
10
time comes up later, we cannot revisit the PRA and
11
update it and all of that.
12
MR. WACHOWIAK:
13
If this design is certified, this PRA exists to
would see that is my next slide.
MEMBER
14
15
If you had my slides, you
APOSTOLAKIS:
Okay.
Let
me
continue the thought.
16
MR. WACHOWIAK:
17
MEMBER
Okay.
APOSTOLAKIS:
Regulatory
guide
18
1.200 also makes a big deal out of a peer review.
19
There is no such requirement here, is there?
MR. WACHOWIAK:
20
No, there is not.
And
21
you will also see that in the ISG, that for this
22
particular PRA, there isn't a peer review.
23
are several reasons for that.
24
staff is actually reviewing the PRA versus the ones
And there
One reason is that the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
165
1
specifically addressed in reg guide 1.200.
2
review is somewhat of a surrogate for a staff review.
3
And I think that's even in the words of reg guide
4
1.200 now.
MEMBER
5
6
here, too.
But
it
could
be
It could be a surrogate here as well.
It could be.
We run into
-MEMBER APOSTOLAKIS:
9
10
APOSTOLAKIS:
MR. WACHOWIAK:
7
8
The peer
Anyway, they don't
require it.
MR. WACHOWIAK:
11
It's not required.
We
12
run into some difficulties, especially with a new
13
design in doing a full, thorough peer review, because
14
those peer reviews tend to -- we try to make them
15
happen in a few-week time frame.
16
And if you have a brand new design that
17
no one has looked at before, it's hard to find peers
18
and
19
whereas, I think it's taken the staff a year to get
20
up to speed on everything they need to know about the
21
ESBWR design in order to do a good review.
get
them
up
to
speed
in
a
few-week
period;
So at this stage, too, it's difficult to
22
23
do
that
type
24
additional.
of
peer
view,
but
we
do
have
the
And we have done some limited things
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
166
1
that I think we have talked about a lot before.
2
have gone through the standards, ASME standards, and
3
the NEI peer review process and internally done an
4
assessment of our models against all the requirements
5
of the RAI that we sent our response in that showed
6
that comparison.
MEMBER BLEY:
7
Rick, excuse me.
We
I know
8
this is your turn, but so I don't forget it later, I
9
would like to ask the staff if they agree that the
10
review they are doing now is at least equivalent to
11
the kind of review reg guide 1.200 would have had an
12
outsider do.
MR.
13
resources
CARUSO:
that
we
Well,
have
I
had
don't
to
do
think
the
this
are
14
level
15
equivalent to a peer review team that has a number of
16
people with different expertise, but I would say, you
17
know, I guess we have had more time than a peer
18
review team would have.
19
comes out to be equal.
So maybe in that sense, it
I think the one thing that I wish could
20
21
have
been
better
was
I
think
because
the
PRA
is
22
evolving, you know, the most significant and detailed
23
review we did was of the first one.
Many of the questions, as I mentioned
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
167
1
last time when we were here, there were, you know,
2
250, 300, 400 questions.
3
were addressed by modifying the PRA.
4
additional modifications that weren't based on those
5
questions.
Many of those questions
And there were
So we sort of had a moving target.
6
We
7
have not had the resources to go back every time we
8
have gotten rev. 2 to go back and do the level of
9
detail of review that we did on rev. 1.
10
that
MEMBER BLEY:
Do you foresee a point at
will
to
11
which
be
able
12
actually get that final thorough review?
MR. CARUSO:
13
happen
that
you
will
Well, I think we feel like
14
what we have done so far in terms of the rev. 1 and
15
looking at the responses and doing additional reviews
16
and
17
quality and the work that GE has done, their own
18
self-assessment,
19
procedures, that we feel that we have done enough to
20
be able to judge the PRA in terms of its application
21
in this context.
the
questions
that
and
we
looking
22
MEMBER BLEY:
23
MR. WACHOWIAK:
24
asked
at
about
their
technical
quality
of
Thanks.
And just to tie this up a
little bit, having just within the past few months
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
168
1
participated on a peer review for an operating plant,
2
I think the level of depth that the staff has gotten
3
into is equivalent to what a peer review team would
4
look at and in some cases would go even deeper.
5
MEMBER BLEY:
6
MR. WACHOWIAK:
there
is
a
Okay.
finding
Now, in a peer review
7
when
8
transmitted to the utility.
9
resolves that and it doesn't ever go back to the
10
review team, where in this case, where we had those
11
issues, the resolution went back to the reviewers.
12
And the reviewers reviewed the resolution.
13
that sense, it's more thorough than a peer review.
MEMBER
14
or
a
comment,
that
gets
And then the utility
APOSTOLAKIS:
Let's
So in
explore
a
15
little bit of the words in this context, presumably
16
the design certification.
17
MR. WACHOWIAK:
18
MEMBER
that
Yes.
APOSTOLAKIS:
20
context in the sense that basically what you want to
21
make
22
right, the accident sequences, the event trees, and
23
fault trees are meaningful, but also that the numbers
24
are meaningful because you are arguing that when you
that
your
different
sequences
from
wondering
whether
is
is
am
19
sure
context
I
are
any
other
meaningful,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
169
1
do the focused PRA, your core damage frequency is
2
well below the focus.
So both things matter.
3
The sequences,
4
which presumably are used to identify single-point
5
vulnerabilities or other kinds of vulnerabilities,
6
may affect the design.
Of
7
But the numbers matter, too.
course,
this
PRA
will
not
be
as
8
detailed as a PRA for an operating facility because
9
you don't have certain kinds of information.
10
So is that what you guys understand as
11
well by the words "in this context," that we have to
12
make sure that at least what we have is correct in
13
the terms of the sequences and the cut sets but also
14
that
15
importance?
the
numbers
are
not
really
of
secondary
The numbers also could be reasonable.
16
MR. CARUSO:
17
I
agree
Mark Caruso with the staff.
with
that.
I
think
we
have
18
looked at the numbers a great deal.
19
especially at common cause failures.
20
used operational data where it was appropriate, where
21
we
22
motor-operated valves.
were
at
pumps
and
we
MEMBER APOSTOLAKIS:
23
24
looking
this common understanding.
We have looked
You know, we
were
looking
at
As long as we have
That's all I wanted to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
170
1
make sure of, that we're not missing anything.
MR. WACHOWIAK:
2
Yes.
And/or determining
3
that we don't have vulnerability is one thing and for
4
determining that we meet the goals with a full PRA
5
and with the focused PRA.
6
completely doable with what we have.
Those sorts of things are
7
MEMBER APOSTOLAKIS:
8
MR. WACHOWIAK:
9
10
If we're going to go and
try to determine allowed outage time for tech specs
for certain equipment --
11
MEMBER APOSTOLAKIS:
12
MR. WACHOWIAK:
13
16
-- we don't have that
MEMBER APOSTOLAKIS:
No.
That's
what
I
that
within
the
limitations of what we have, -MR. WACHOWIAK:
18
MEMBER APOSTOLAKIS:
Right.
-- both the vent
analysis and the quantitative analysis --
20
MR. WACHOWIAK:
21
MEMBER APOSTOLAKIS:
22
MR. WACHOWIAK:
23
MEMBER
24
No, absolutely not.
meant,
17
19
No, no.
kind of information.
14
15
Okay.
Right.
-- matter.
Okay.
APOSTOLAKIS:
So
that's
good
enough.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
171
MR. CARUSO:
1
I might just add that I
2
believe that some of the other vendors -- and I won't
3
mention any names -- that are coming in for design
4
certification I believe are going to bundle together
5
risk-informed
6
application.
application
50.59
That's part.
MEMBER APOSTOLAKIS:
We will revisit the
context at that time.
11
MR. CARUSO:
12
MEMBER APOSTOLAKIS:
13
in
in much detail at that time.
9
10
spec
I think that we will get into this issue
7
8
tech
Yes, right.
At this point this
is all we -MR.
14
because
WACHOWIAK:
15
slide
after
16
objectives are --
we
MR. OESTERLE:
17
Let's
go
and
go
to
the
next
that
is
what
our
Well, I think hang on just
18
a second.
Eric Oesterle from the staff.
I just
19
wanted to butt in with an important clarification.
20
And this is my interpretation of what the
21
question also asks is that at this point I am not
22
aware of any staff position that would indicate that
23
the staff's review of the applicant's PRA can be
24
considered the peer review as described in the reg.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
172
1
guide.
MEMBER APOSTOLAKIS:
2
3
4
There isn't such a statement there.
MR. OESTERLE:
6
MEMBER APOSTOLAKIS:
Right.
I personally would
not say it's -MR. WACHOWIAK:
8
9
But we are just
asking to figure out where the people stand.
5
7
Yes, you are right.
Go back to the previous
one?
DR. KRESS:
10
Can I see it just again?
I
11
am interested in just what your definition might be
12
of a vulnerability.
13
plant vulnerabilities for the operating plants, they
14
kind of thought of it as a CDF greater than 10-4.
15
it got greater than that, they thought of it as a
16
vulnerability.
You know, when they did the
If
I don't think that would apply to your
17
18
plant.
So I was wondering what you considered may be
19
a vulnerability if your CDF was greater than your
20
target by a certain amount or -MR. WACHOWIAK:
21
The main thing that we
22
were looking for for vulnerabilities was things like
23
single point failures, where if you had an initiating
24
event, like a transient initiating event and some
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
173
1
single failure that would lead you to core damage,
2
lead
3
things that are associated with the normal operating
4
state.
you
to
core
damage,
would
be
vulnerability,
Now, we have a few things in shutdown.
5
I
6
think everybody is aware that if you have a pipe
7
break in shutdown, if we don't get the lower drywell
8
hatch closed within a certain time frame before the
9
water starts coming through the door, then you are
10
significantly on the way to a core damage event.
11
It's
12
situation.
very
difficult
That
13
to
is
we
probably
as
close
to
to
16
that
17
address that particular thing.
procedural
put
a
constraints on the applicants, the COL applicants, so
certain
had
that
15
to
We
to
vulnerability
commit
have.
perfectly
14
they
as
respond
some
things
to
But those are the kinds of things that
18
19
we're
20
there's an initiating event and some small number of
21
failures will take you to core damage, we don't want
22
that to happen.
for.
Now,
23
24
looking
If
there's
things that get there.
we
the
have
something
common
cause
where
failure
But we have to look at what
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
174
1
the specific common cause failures are.
2
common cause failure of eight things, that's not a
3
single point failure.
MEMBER
4
If it's a
Nobody needs to --
APOSTOLAKIS:
So
what
you
are
5
saying is the event sequences are the first thing you
6
look at and then the probability.
7
MR. WACHOWIAK:
8
MEMBER APOSTOLAKIS:
9
10
Because if you say
it's eight things that must fail, I don't care.
MR. WACHOWIAK:
It's not a vulnerability,
yes.
MEMBER
13
14
sense.
15
vulnerability --
APOSTOLAKIS:
Well,
that
made
There is no specific definition of what a
16
MR. WACHOWIAK:
17
MEMBER APOSTOLAKIS:
days,
every
You remember in the
IPEEE
19
identified
20
"These are the changes we made to the project."
no
single
Yes.
18
licensee
vulnerabilities,"
21
(Laughter.)
22
MR. WACHOWIAK:
23
I
mean, it's not a --
11
12
Right.
said,
next
"We
have
paragraph,
No vulnerabilities after
the changes.
So let's go to the next one here.
24
I
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
175
1
think this is my last point I want to make.
Part 52
2
was recently revised.
3
Part 50, which requires the owners of the plant to do
4
a fully standard, compliant PRA prior to fuel load.
This statement was added to
5
So, no matter what anybody does with any
6
of their design cert PRAs, this is a requirement by
7
law that they have to do that.
8
on the other statements that go --
And then, as you see
MEMBER APOSTOLAKIS:
9
MR. WACHOWIAK:
10
Right.
-- farther on down, it is
11
now required by law that they update that every four
12
years.
13
will have to be updated.
And the second piece is
14
that
be
15
standards that are endorsed at the time when they do
16
the update.
So two pieces to this.
it
is
required
to
One, the design PRA
compliant
with
the
17
So if we have some piece of this PRA that
18
we couldn't do because we didn't have the information
19
at this stage of the design and that's something
20
that's required by the standard, well, when this PRA
21
is done, that is going to need to be included in
22
there
23
standard.
or
they
wouldn't
be
compliant
On to the next one, I think.
24
with
the
I just said
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
176
1
that.
That was the update requirement that is in
2
there.
And I think that's it.
3
quite it.
4
allow to do these conclusions.
5
already.
7
Oh, no.
I had to put in a conclusion.
So
6
Right?
do
we
want
Not
I'm not
We discussed this
to
start
with
one
particular sequence?
8
MEMBER APOSTOLAKIS:
I don't know.
9
CHAIRMAN CORRADINI:
I think that was the
10
plan.
MEMBER STETKAR:
11
12
Sequence
descriptions.
CHAIRMAN CORRADINI:
13
14
Let me start.
I guess Mr. Stetkar
has a suggestion.
MEMBER STETKAR:
15
Only because we talked
16
about the severe accident situation, a lot about the
17
BiMAC, and things like that, what I would like to do
18
is start a little bit from the back end.
19
put it in the context of the specific sequence if we
20
want to, but some of the general questions apply more
21
globally.
So now I warn you this is going to get
22
23
We could
real detailed real fast.
So --
CHAIRMAN CORRADINI:
24
Do you want to have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
177
1
them at least -MEMBER STETKAR:
2
3
Please
4
picture
5
individual sequences.
6
in case we want to get them into a specific path
7
through the event model.
issues
You
know,
here
there
than
are
nitpicky
some
bigger
details
deluge valves.
You have 12 valves.
GDCS
There's a top
10
event in the level 2 event tree called BI_SP.
11
it up.
12
if any six deluge valves open," period.
Look
The success criteria for BI_SP says, "I win
Now,
13
of
The sequences are good context
Let me ask you a specific question.
8
9
don't.
This is not a sequence.
what
is
the
basis,
A,
for
that
14
general success criteria?
And, B, how does the PRA
15
differentiate between the deluge valves that supply
16
the BiMAC cooling tubes versus the deluge valves that
17
dump into the lower drywell, two questions?
18
please.
19
MR. WACHOWIAK:
20
MEMBER BROWN:
21
MEMBER STETKAR:
22
Okay.
That was crisp.
I said we are going to
get real detailed real fast -MEMBER BROWN:
23
24
Answers,
Not hard to understand at
all.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
178
MEMBER STETKAR:
1
-- because what I am
2
trying to do is follow up on definitions of success
3
criteria.
4
understand it?
5
PRA?
Are they consistent with the design as we
And how are they implemented in the
6
MEMBER BROWN:
7
MEMBER
8
how
were
those
So let me understand
the question.
MEMBER STETKAR:
Because I haven't seen a
question about this.
MEMBER APOSTOLAKIS:
13
14
And
MEMBER APOSTOLAKIS:
11
12
STETKAR:
reviewed during this review?
9
10
Right.
So you say there
were eight?
MEMBER
15
16
valves.
17
six open."
STETKAR:
There
are
12
The PRA success criteria says, "I win if any
18
MEMBER APOSTOLAKIS:
19
MEMBER STETKAR:
Okay.
What's the basis for
20
six, not seven, not nine, not three, six?
21
technical basis.
22
go
23
valves only go to the BiMAC tubes.
24
deluge valves only go to the lower drywell.
to
deluge
different
That's a
And because certain deluge valves
places,
I
can
win
if
six
deluge
I can win if six
I can
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
179
1
win if three and three, five and one, two and two,
2
and two and four.
3
I want to know what the basis for the six
4
was and how does the PRA differentiate where I am
5
actually getting the water because right now, A, it
6
doesn't differentiate.
7
basis for the six in the beginning.
8
really important because if we don't have successful
9
deluge, it is a bad day in the containment.
And, B, I don't know the
10
MR. WACHOWIAK:
11
(Laughter.)
12
MEMBER STETKAR:
13
a worse day outside.
MR.
14
15
It was already a bad day.
It's a worse day.
It's
There you go.
WACHOWIAK:
I
think
you
get
it,
right?
MEMBER STETKAR:
16
17
And this is
the questions.
So those are
What's the basis for six?
MR. WACHOWIAK:
18
I get it.
Okay.
The basis for six
19
I think came from Theo.
And what he said was that we
20
need to have about a certain I guess flow area of
21
pipe available from the tanks down into the BiMAC.
22
And originally when we had just four valves, our
23
success criteria would have been any two.
24
have decided that we want to expand it out to more
When we
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
180
1
than just four valves, we retained the value of half
2
of the valves.
Now,
3
here
behind
in
is
where
the
I
PRA
think
versus
we
got
what
a
4
little
the
5
implementation is in the design is that the PRA does
6
not take into account that half are not going into
7
the common header in the tubes.
8
other places.
9
to need to go back and look at.
They're going to
So that's something that we're going
10
But, given that, even if we make it that
11
it's got to be whatever the success criteria turn out
12
to
13
differences in the numbers because of the limitations
14
--
be,
I
don't
think
that
MEMBER STETKAR:
15
we'll
end
up
with
That's speculation about
16
where we're headed.
I'm just looking at, do the
17
model and the success criteria support what we know
18
about the design today?
19
design today is that.
20
that.
And what I know about the
And the PRA does not support
21
MR. WACHOWIAK:
22
MEMBER
That's right.
STETKAR:
23
statement.
24
success criteria may be.
That's
a
simple
I don't care speculation about what the
I don't care speculation
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
181
1
about what the numbers may be.
2
accurately model the plant as we know it today?
MR. WACHOWIAK:
3
4
And where we are.
And
that particular one -MEMBER
5
6
It's does the PRA
STETKAR:
Doesn't.
Okay.
So
that's one point.
7
MR. WACHOWIAK:
We're out of sync.
8
MEMBER STETKAR:
Okay.
9
MR. WACHOWIAK:
And because one of the
That's a point.
10
things that happens in doing the design PRAs, we get
11
the information.
12
in the PRA.
We put it into our success criteria
The designers also get the information.
MEMBER STETKAR:
13
but
we're
14
thing,
15
ultimate tail end of that entire process.
16
you see it before anybody else does.
17
it after you do.
18
we finding it?
19
staff and not you?
MR.
on
the
That's an interesting
ACRS
and
22
MEMBER APOSTOLAKIS:
Why?
23
MR. WACHOWIAK:
24
the
The staff sees
Why are
Why are we finding it and not the
know
21
at
You know,
We see it after the staff.
I
20
we're
WACHOWIAK:
why
the
staff
isn't finding it.
Because that model wasn't
in the previous revs. of the PRA.
This rev. 3 is the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
182
1
first time we've actually had this.
MEMBER
2
3
STETKAR:
MR. WACHOWIAK:
5
MEMBER
I
made
that
it
our review.
MEMBER STETKAR:
This is where we started
on rev. 2.
MR. WACHOWIAK:
12
MEMBER
that
Okay.
I thought --
APOSTOLAKIS:
13
established
14
sequences would be affected by this?
what
you
said
15
MEMBER STETKAR:
16
MEMBER APOSTOLAKIS:
17
MEMBER STETKAR:
18
core damage.
19
a specific sequence.
is
Okay.
We
true.
Which
All sequences.
All sequences.
All sequences that go to
That's why I didn't want to talk about
MEMBER APOSTOLAKIS:
20
Yes.
I understand
that.
MEMBER STETKAR:
22
23
sure
Which is where we started
11
21
I
We had the --
STETKAR:
MEMBER BLEY:
9
10
true.
didn't change in rev. 3.
7
8
not
found it in rev. 2.
4
6
That's
It's functional success
criteria.
MEMBER APOSTOLAKIS:
24
Okay.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
183
MEMBER STETKAR:
1
2
That's to get water in
for debris coolability.
MR. WACHOWIAK:
3
In the initial PRA for
4
the BiMAC system itself since we didn't have even
5
close to the design, it was conceptual design, we
6
initially
7
itself that said that that system needed to perform
8
independent of the other systems in the plant that
9
were associated with any sequences where we used,
set
reliability
criteria
for
the
10
core damage sequences.
11
reliability or unreliability of 10-3 or lower.
BiMAC
And it would need to have a
And I don't know that we don't have that
12
13
requirement anymore.
14
things where when we loop back through there, we
15
would make sure that however we set up this system,
16
whether it's 12 valves or whether we need to have
17
those arranged differently, it still needs to meet
18
that reliability criteria in order for us to meet our
19
goals for the containment.
So you found one of probably many things
20
21
where it is not done.
MEMBER STETKAR:
22
23
24
And so this is one of these
This
is
another
sequence-specific.
generic
Let me ask.
one.
Vacuum breakers.
So
Let me ask.
it's
not
In top event,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
184
1
there's a -- this again is in the level 2 containment
2
models.
3
for the operation of the vacuum breakers.
There is a top event called DS-TOPBV.
It's
4
The success criteria for that top event,
5
if I understand the fault tree correctly, is that at
6
least one of the three vacuum breakers must open to
7
equalize pressures and that two of the three vacuum
8
breakers
9
fault tree logic.
must
successfully
re-close.
That's
the
10
My question is, since, as I understand
11
it, if you have a leak rate more than about the
12
square equivalent area of 14 centimeters2, you may
13
have a problem.
Let's
14
say
all
three
vacuum
breakers
15
successfully open so it meets at least one of three
16
opened and one of them stays completely open and the
17
isolation valve does not close.
18
the containment?
CHAIRMAN CORRADINI:
19
20
John?
Can you repeat that,
I'm sorry.
MEMBER
21
Doesn't that fail
STETKAR:
If
all
three
vacuum
22
breakers open successfully and one of them remains
23
open, does not re-close, and its isolation valve does
24
not close because it's got an isolation valve on it
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
185
1
-- so I have an open vacuum breaker path.
2
CHAIRMAN CORRADINI:
3
MEMBER STETKAR:
Manhole cover is --
Manhole cover and its
4
isolation valve are still open.
5
containment according to the success criteria that I
6
understand for bypass scenarios?
7
MR. WACHOWIAK:
8
MEMBER STETKAR:
Does that not fail
Which top event were -It's called DS, dog Sam
9
-- I don't understand the military stuff -- -TOPBV.
10
And the success criteria require any one of three to
11
open.
12
So you fail if all three fail to open.
13
You also fail if any two fail to close, which means
14
one could have remained fully open.
15
if
16
failure?
one
remains
fully
MR. HOWE:
17
another
open.
You do not fail
Isn't
that
really
In the long term, there is
18
actually
tab,
19
requires all three vacuum breakers to be leak-tight.
20
This function right here is primarily just for steam
that
which
is
suppressions,
22
transient.
23
integrity, we require all three to be leak-tight.
then
for
MEMBER STETKAR:
24
kind
long-term
Okay.
of
which
21
And
initial
DL-TOPBV,
pressure
containment
I may have missed
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
186
1
that.
MR. WACHOWIAK:
2
Yes.
The function of
3
this one is like the vacuum breakers in the existing
4
plants.
5
work.
It's to make the pressure suppression system
MEMBER STETKAR:
6
Now, is there no way
7
that you go to -- I don't have the event tree up here
8
in front of me because I've got too many files.
9
you give me a chance, I can bring up the event tree
10
and look at it.
11
going to have to go before --
I mean, this is the way this is
MR. HOWE:
12
If
Right.
To get to long-term
13
success, we have to go through the other top, which
14
requires all -MEMBER STETKAR:
15
16
That's why I -- what
header is that under on the event tree?
17
MR. HOWE:
It would be under the W2.
18
MEMBER STETKAR:
19
MR. HOWE:
Yes, one of those that is
20
going to be DL-TOPBV.
There is an example up here.
21
It's a part of that W2 node.
22
MEMBER STETKAR:
23
stuff is under W2.
24
short-term stuff is under W1?
W2?
Okay.
The long-term
So it's only questioned.
And the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
187
1
MR. HOWE:
That's right.
2
MEMBER STETKAR:
3
MR.
Okay.
WACHOWIAK:
And
then
the
steam
4
suppression is under VB in this containment event
5
tree.
6
associated with that.
7
containment is going to fail early.
8
believe that's the active RHR system.
9
the passive decay heat removal system.
So we've got three different things that are
MEMBER
10
Okay.
The VB node is whether the
STETKAR:
The W1 node, I
And
And the W2 is
W2
has
the
11
long-term.
I didn't get as far as W2 because
12
I was interpreting BV as all functional failures that
13
disabled containment cooling.
MR. HOWE:
14
15
is also under W1.
16
support to the PCCS.
18
Right.
Actually, that DL top
It's really kind of used as a
MEMBER STETKAR:
17
So it's not.
What's the top called?
DL something?
19
MR. HOWE:
DL-TOPBV.
20
MEMBER STETKAR:
21
MR. HOWE:
22
MEMBER STETKAR:
23
MR. HOWE:
24
MEMBER STETKAR:
DL-TOPBV?
Yes.
Thank you.
That's it.
Now, since we're talking
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
188
1
about kind of level 1/level 2 interface, now we kind
2
of
3
sequence, but I was concerned about -- the general
4
topic is GDCS deluge again now.
have
to
walk
through.
5
MR. WACHOWIAK:
6
MEMBER
It's
not
a
specific
Okay.
STETKAR:
Suppose
you
have
7
successful GDCS injection so that the DPVs all open,
8
all the GDCS pools dump into the vessel, and then you
9
go to core melt.
10
MR. WACHOWIAK:
11
MEMBER STETKAR:
You won't.
Yes, you will.
FDW-0033
12
has indeed cut sets.
13
go to melt because you have late makeup failure.
14
This is a late low-pressure melt scenario.
15
possible.
CHAIRMAN CORRADINI:
16
17
18
And there are ways that you can
It is
So somehow the water
doesn't make it back from the PCCS back to the GDCS?
Is that -MEMBER
19
STETKAR:
The
only
thing
I
20
understand is what I look at in the risk model.
So
21
there are requirements that for late makeup, I either
22
need active makeup from some of the active systems or
23
I need to have equalizing or I need to have dumped at
24
least two of the GDCS pools in there.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
189
And there are combinations that indeed
1
2
fail that.
3
I am just talking about functions here.
4
sequences.
5
you want to call them, where I can in this plant have
6
successful low-pressure injection via the GDCS pools.
7
And I can dump them all in there and, yet, still go
8
I don't want to talk about frequencies.
There are cut sets, sequences, whatever
to late low-pressure damage, core damage.
CHAIRMAN CORRADINI:
9
So, just to help me
10
along, can GEH explain?
11
the
12
That's the only way to get that physically.
PCCS
to
deliver
It's got to be a failure of
the
13
MR. WACHOWIAK:
14
(Laughter.)
15
MR. WACHOWIAK:
16
There are
water
back
to
the
GDCS.
Not really.
There are ways to get
there.
17
MEMBER STETKAR:
Okay.
18
MR. WACHOWIAK:
They end up being low.
19
What we could talk about -MEMBER STETKAR:
20
Fine.
They are low-frequency,
21
but I want to get back to right now I am talking
22
about functions and success criteria because I may be
23
confused.
The situation I was thinking about is
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
190
1
suppose you are in this situation.
So all the GDCS
2
pools have dumped.
3
a most likely -- in the real world, that's a most
4
likely situation.
5
I get high temperature down in the lower drywell.
6
And all of my GDCS dump valves dutifully open.
7
does the water go?
The DPVs are open because that's
And then you go to core melt.
And
Where
8
I mean, how do I know that now the entire
9
inventory of water that is coming back into the GDCS
10
pools is going to know that it needs to go down in
11
the lower drywell and feed the BiMAC, rather than
12
going into the vessel, which is also depressurized
13
and
14
That is where the DPVs dump.
just
circulating
in
the
MR. WACHOWIAK:
15
upper
drywell
space?
So there are a few things
16
with this sequence set.
The lines that come off of
17
the -- that go from the tanks to the BiMAC are the
18
same lines as the GDCS.
19
MEMBER STETKAR:
20
MR. WACHOWIAK:
Yes.
So some of it would still
21
go, could still go to the vessel.
22
you guys talk about the -MEMBER STETKAR:
23
24
pressures are, though.
But I think when
It depends where the
I mean, you know, if the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
191
1
pressure is higher in the lower drywell, the water is
2
-CHAIRMAN
3
CORRADINI:
But
how
can
the
4
pressure be higher in the lower drywell?
5
drywell and the upper drywell are connected by about
6
ten
7
different.
square
meters
of
area.
MEMBER STETKAR:
8
So
they
The lower
can't
be
That's why I wanted to
9
find out whether the pressures would be the same or
10
whether I can get a lower drywell bypass condition
11
going on.
MR. WACHOWIAK:
12
I think that that's an
13
interesting question, but I believe that we handled
14
it in a different way, though.
15
sequences where we have already dumped the GDCS pools
16
and those -- I believe we have those marked as high
17
water level in the lower drywell cases.
18
MEMBER STETKAR:
19
goes to a CD-1 low level.
20
specific sequence.
21
we want to talk about --
No?
Not this sequence.
It
Now I'll look at the
It is this FDW-0033 sequence that
22
MR. WACHOWIAK:
23
MEMBER STETKAR:
24
Those particular
That is a --- mapped to a late,
low-pressure, low-level in the drywell sequence.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
192
1
MR. WACHOWIAK:
2
MEMBER
3
that.
Okay.
STETKAR:
So
it's
run
through
It's run through that part of the model.
MR.
4
HOWE:
Yes.
For
that
specific
5
sequence, we have successful early GDCS failure of
6
late GDCS.
7
MEMBER STETKAR:
8
MR. HOWE:
9
Right.
Based on the success criteria
that are defined, that means you have only had at
10
most one of the three pools inject to that core.
11
even if you do get that core melt-through, you still
12
have two other GDCS pools.
MEMBER
13
14
STETKAR:
success
criteria
requires successful injection of one.
15
MR. HOWE:
16
MEMBER STETKAR:
17
The
So
Right.
Success could also occur
if you had all three of them.
MR. HOWE:
18
But the success criteria for
19
the long term GDCS is two of three pools.
20
MEMBER STETKAR:
That's right.
That's
21
right.
The success criteria say that I can win if I
22
have one inject initially.
23
short term if I have all three inject.
24
in the short term if I had two of the three inject in
In fact, I can win in the
And I can win
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
193
1
the beginning.
2
MR. HOWE:
Right.
3
MEMBER STETKAR:
So I can win in both the
4
short term and the long term if all three GDCS pools
5
inject, right?
6
MR. HOWE:
7
MEMBER
That's correct.
STETKAR:
And
under
those
8
conditions, I've won in the short term.
9
in the long term, but there are other things that can
10
I could win
still fail me in the long term.
11
MR. HOWE:
12
MEMBER
Right.
STETKAR:
So,
even
though
all
13
three injected in the short term, at least two of
14
those are necessary but not sufficient.
15
success in the long term.
16
long term having had all three inject in the short
17
term.
You achieve
So I can still fail in the
Is that correct?
MR. HOWE:
18
Yes.
And then that would be
19
due to containment heat removal failure, which we
20
have been as the class 2A and class 2B sequences.
21
And
22
sequences in the level 2.
23
relief.
those,
we
do
MEMBER
24
not
model
STETKAR:
mitigation
of
those
Those are soon to go to
I
didn't
see
that,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
194
1
though, because in the sequence mapping, at least in
2
the tables that I read in the report, this particular
3
sequence, the one --
4
MR. HOWE:
5
MEMBER
Right.
The feedwater --
STETKAR:
--
FDW-0033,
in
6
particular, so now if we're going to talk about --
7
for
8
generic issue because there are other sequences that
9
look
this
one,
like
I
this.
need
the
But
if
context.
you
want
It's
to
a
talk
more
about
10
specific ones, this one indeed can satisfy all those
11
conditions.
12
understand the documentation correctly, it's mapped
13
to a CD-1 --
And according to the documentation, if I
14
MR. HOWE:
Right.
15
MEMBER STETKAR:
-- and, in particular, a
16
low drywell level CD-1 because you differentiate the
17
level in the drywell for the CD-1's.
18
MEMBER BLEY:
19
MEMBER STETKAR:
20
Which is not guaranteed --
failure --
21
MR. HOWE:
22
MEMBER STETKAR:
23
Which is not guaranteed
Right.
-- because then you run
that through that level 2 event tree.
MR. HOWE:
24
Right.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
195
MEMBER STETKAR:
1
that
level
2
event
You know, that's linked
2
to
tree,
where
you
have
the
3
success criteria that says you win if any six of the
4
GDCS valves open up.
So my question first is it's kind of a
5
6
two-part question.
I wanted to understand whether my
7
original concern about where does the water go is
8
valid.
9
can go into the lower drywell, that there isn't any
And the second concern is if indeed the water
10
pressure
difference
or
some
other
11
would preclude that due to any six GDCS deluge valves
12
opening.
13
this condition or maybe do I need more deluge valves
14
to open?
that
Does that success criterion apply under
CHAIRMAN CORRADINI:
15
phenomenon
I don't think -- you
16
guys are in another space and dimension than I am,
17
but if I dumped enough of that water by the way they
18
have the design, you're going to have a very deep
19
pool of many meters already in the lower cavity.
MEMBER
20
21
lower cavity.
No,
yet.
it's
It's not down there.
CHAIRMAN CORRADINI:
22
23
STETKAR:
not
in
the
It's up there.
It's not down there
It's not down there yet.
MR.
24
WACHOWIAK:
If
you
look
at
the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
196
1
long-term success criteria, there are two things that
2
can get you to success in the long term:
3
more
4
equalizing lines open.
pools
inject
So
5
in
or
if
your
one
pool
three-pool
if two or
injects
case,
plus
if
three
6
pools go in early, by definition, you have already
7
won late because three pools win late.
8
go in early, by definition, you have already won
9
because two pools are sufficient to carry you through
10
If two pools
late.
11
If only one pool then goes in, you have
12
to open an equalizing line to stay for the long term.
13
14
15
And
in
particular
case,
there's
not
enough
water to get it all filled down in the lower drywell.
So that's why it's low.
CHAIRMAN CORRADINI:
16
17
that
Where does it end up
then?
18
MR.
WACHOWIAK:
19
suppression pool because -MEMBER
20
It
STETKAR:
You
ends
up
know,
in
the
walk
me
21
through that.
I'm sorry to stop you, but we'll get
22
back to that.
Walk me.
23
or two minutes.
Back up about three minutes
MR. WACHOWIAK:
24
Right.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
197
MEMBER
1
2
STETKAR:
walk
me
back
through.
3
MR. WACHOWIAK:
4
MEMBER STETKAR:
5
MR. WACHOWIAK:
6
And
So we have the -I think you're right.
We have the short-term
and a long-term.
7
MEMBER STETKAR:
Yes.
8
MR. WACHOWIAK:
The short-term requires
10
MEMBER STETKAR:
Right.
11
MR.
9
one pool.
12
requires
either
13
equalizing line.
14
talk to it.
WACHOWIAK:
two
And
pools
or
the
one
long-term
pool
and
an
So it makes the suppression pool
So if three pools work like you say --
15
MEMBER STETKAR:
Oh, okay.
16
MR. WACHOWIAK:
By definition, it makes
17
you win on both short and long-term.
18
inject, you win on short and long-term.
19
pool injects, then you will only win on short-term
20
and you need additional equipment in the long-term.
21
That's why we modeled it that way, was to --
22
CHAIRMAN CORRADINI:
23
If two pools
If only one
When you say, "win,"
you mean to avoid core damage?
MR. WACHOWIAK:
24
To avoid core damage,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
198
1
yes.
And so what we have to ask is, does PCCS work
2
to keep everything going?
CHAIRMAN CORRADINI:
3
All right.
So just
4
to answer my question, then, John, just to finish it
5
off then, if you dumped early with one pool but the
6
equalizing line would work, you would go to core melt
7
and all the water ends up in the suppression pool,
8
instead of a lot of the water.
MR.
9
WACHOWIAK:
To
get
back
to
your
10
point, we have six PCCS heat exchangers.
11
directed into the various GDCS pools.
And the GDCS
12
pools
the
13
connected up at the top with -- when they fill the
14
pools, you fill one.
15
So they're kind of all interconnected up at the upper
16
water-level range.
are,
at
least
currently
in
And they're
design,
And it cascades to the other.
So water that comes back from the PCCS
17
18
heat
exchangers
goes
into
those
pools.
And
if
19
there's a pool that's open to the reactor, that will
20
continue to go to the reactor.
21
the pool, the cascading lines that we talk about are
22
-- you know, we don't know how those are going to be
23
arranged right now.
If, for some reason,
You know, they could have high points.
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
199
1
They might not.
2
safety function for it.
3
4
We just don't know.
There is no
It's just a fill the pools.
But let's say that those don't equalize so that all
--
5
CHAIRMAN CORRADINI:
6
MR.
WACHOWIAK:
I understand.
--
the
pools
don't
7
cascade back to the one that's feeding the reactor.
8
So we don't have that modeled that way yet.
9
pools were to overflow, there's an overflow line on
10
If those
those pools that goes to the suppression pool.
11
So if there are six PCC heat exchangers
12
distributed to three pools, two-thirds of the PCC
13
condensate, the water that's coming up on steam out
14
of
to
the
15
suppression pool through those overflow lines.
So
16
over
is
17
available for the closed circuit in the one pool.
the
reactor,
time
you
ends
will
up
flowing
deplete
the
down
water
that
18
Once again, now that we know how these
19
lines are going to cascade those pools for normal
20
refill, when we see the detailed design on that,
21
we'll revisit that one to see if now in the long run
22
one pool is enough for a success.
23
enough detail on those lines to make that decision at
24
this point.
But we don't have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
200
CHAIRMAN
1
2
MEMBER STETKAR:
the heat here.
I'm
So bear with me.
CHAIRMAN CORRADINI:
6
MEMBER STETKAR:
You have four guys.
Yes.
I really hate to
do this.
MR.
8
9
you.
I am trying to think on
5
7
Thank
satisfied.
3
4
CORRADINI:
WACHOWIAK:
You've
got
one
back
there.
10
CHAIRMAN CORRADINI:
11
MEMBER
BLEY:
But you have Dennis.
Yes.
12
completely tracked this one.
13
MEMBER STETKAR:
But
I
haven't
He's got other ones.
14
But I'm thinking in failure space now.
15
if I have one pool injected successfully in the short
16
term, one and only one, and the equalizing valves
17
didn't open and all the other makeup supplies failed.
18
MR. WACHOWIAK:
19
MEMBER STETKAR:
So I can fail
Right.
So that I have one pool
20
now that is not -- one pool went in and the other two
21
are
22
available because they did not go in initially, their
23
injection valves didn't open or whatever.
still
available.
Aren't
24
the
The
other
success
two
criteria
are
for
still
the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
201
1
number of GDCS -- now, under those conditions, under
2
those conditions, so that I have injected one and
3
let's say that the reason I didn't inject the other
4
two is because none of the -- well, it can't be none
5
of the injection valves opened, but their injection
6
valves did not open.
I'll come back.
7
The DPV valves did all
8
open because they have to open to have any chance.
9
So I've blown down the vessel to the upper drywell.
10
Two pools did not go in.
11
me two pools full of water.
12
didn't open and none of my other makeup supplies.
That leaves
The equalizing valves
So this sequence goes to core damage.
13
14
One pool did.
It's another way of getting to the same sequence.
15
MR. WACHOWIAK:
16
MEMBER STETKAR:
Right.
It goes to what's called
17
CD-1 low level in the drywell.
The level 2 event
18
tree that is linked to that sequence now dutifully
19
asks, do at least 6 of the 12, 6 of the 12, GDCS dump
20
valves open?
21
containment heat removal because, by definition, I
22
have enough water wherever it needs to go.
23
the other part.
24
Now
And if they do, then I can win for
my
question
is,
That's
under
these
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
202
1
conditions, first of all, physics, does the water
2
really know, the water from those two pools now --
3
and I'll grant you you've got two pools -- does the
4
water from those two pools really know that it needs
5
to go into the lower drywell and not a good chunk
6
into the vessel?
CHAIRMAN CORRADINI:
7
8
the
vessel,
9
discharge.
though?
You
How could it go in
said
that
it
failed
And that means the --
10
MEMBER STETKAR:
It's been blowdown.
11
MR. WACHOWIAK:
Right.
12
CHAIRMAN CORRADINI:
But they failed to
open.
MR. WACHOWIAK:
15
16
But they failed
--
13
14
to
The valves, the lines
that --
17
MEMBER STETKAR:
Okay.
18
MR. WACHOWIAK:
-- go from the pool to
19
the vessel have failed.
MEMBER STETKAR:
20
21
Good point.
23
Good point.
Yes, yes.
MR.
22
Good point.
WACHOWIAK:
So
branch
line,
when
those valves open, then -MEMBER STETKAR:
24
I will eventually get to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
203
1
a case where none of this can happen.
MR. WACHOWIAK:
2
3
And then we're back into
the same issue with -MEMBER STETKAR:
4
But the success criteria
5
-- well, there it's a little different on the six
6
because I have effectively disabled three.
7
no -- I am not sure where the water is going in that
8
pool.
9
case, it might be m of 9 with some distribution.
So it might be instead of 6 of 12 in this
10
MR. WACHOWIAK:
11
MEMBER STETKAR:
12
There is
Right.
But that gets back into
the general success criteria.
MR. WACHOWIAK:
13
We're going to need to
14
sync that up that up with the arrangements for those
15
valves.
MEMBER STETKAR:
16
feel
starting
18
getting
19
injection essentially of all the GDCS pools --
concerned
comfortable
that
we
with
I think I am
17
could
CHAIRMAN CORRADINI:
20
21
to
Thanks.
this.
have
I
was
successful
The only way you're
-MEMBER STETKAR:
22
-- and the water coming
23
back to the pools not knowing whether it should go
24
into the vessel and stay in the upper drywell and
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
204
1
essentially in a circulatory mode up there because
2
heat is still coming out.
3
DR. KRESS:
The only way --
4
MEMBER STETKAR:
Of course, it is getting
5
down, you know, how did it know it needed to go down
6
in the bottom?
CHAIRMAN CORRADINI:
7
So just can I get
8
this right?
The only way John's worry would occur is
9
if the first valve on those two pools didn't open,
10
not the downstream valve, because you've got a valve
11
upstream of the check valve that --
12
MR. WACHOWIAK:
13
CHAIRMAN CORRADINI:
14
MEMBER STETKAR:
I'll eventually get to
that.
17
CHAIRMAN CORRADINI:
18
MR. WACHOWIAK:
19
Is that a manual
valve?
15
16
That's a manual valve.
Oh, okay.
That's the maintenance
valves that I wanted to -CHAIRMAN
20
CORRADINI:
So
the
only
DPV
21
valve is the one at the bottom of the loop seal.
22
We'll get to that.
Is that correct?
23
MR. WACHOWIAK:
Yes.
24
CHAIRMAN CORRADINI:
Okay.
Thank you.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
205
MEMBER STETKAR:
1
We'll get to that valve,
2
but that's a systems.
I wanted to try to close some
3
questions that I had in terms of linking success
4
criteria across the level 1/level 2 models, --
5
MR. WACHOWIAK:
6
MEMBER STETKAR:
7
-- which is why I am
starting on the deluge.
MR. WACHOWIAK:
8
9
Right.
In all those systems,
where there are multiple modes of a system that had
10
different
success
11
they're the hard ones to link up between those types
12
of models.
13
this long-term and short-term.
14
an anomaly there.
for
the
same
system,
You end up with something that looks like
MEMBER
15
criteria
STETKAR:
It looks like there's
Well,
yes.
I
mean,
16
those are always difficult to link up.
Well, that's
17
the whole key is you have to walk through them.
18
I think part of the reasons that we wanted to have
19
this type of discussion are sort of two-fold, number
20
one, to answer our own questions because, you know,
21
we have had limited exposure to all of this and very
22
little time to try to examine some things, to do
23
things like this, to understand that, yes, it sounds
24
like it has been thought through and in cases where
And
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
206
1
maybe it hasn't been thought through, to understand
2
why those issues hadn't been raised previous times in
3
the review process, somewhat your concern but also
4
somewhat concern for the staff.
5
MR. CARUSO:
This is Mark Caruso.
6
Rick and I had a conversation on this
7
topic similar to yours not too long ago.
8
this short-term/long-term thing very confusing, too.
9
If you go look at the design basis LOCAs, you will
10
And I found
not find any that ever trigger the long-term.
In addition, there is water that the PRA
11
12
is really taking credit for.
13
the SLICC system that will go in.
14
from the ICCS that will go.
15
But
16
confusing area.
I
MEMBER
17
agree
And there is water from
with
STETKAR:
There is water
you.
It's
This
is
confusing.
a
It
18
takes a little bit of work to walk your way through
19
it.
20
time is not the appropriate forum to do that.
21
year-long review of the PRA, understanding the design
22
information and being able to actually look at the
23
fault trees and walk your way through and the event
24
trees
And a public meeting trying to do things in real
and
walk
your
way
through
the
A
sequences
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
207
1
certainly does seem a way to kind of look at these
2
sort of issues.
3
So that's what I'm trying to understand,
4
is if we, the ACRS, at the tail end of the process,
5
having a couple of days to look at these things in
6
real time can find things that may be deficiencies.
7
8
understand
9
process
You
know,
why
the
hasn't
I
think
longer,
done
that,
we
more
either
really
want
in-depth
your
to
review
internal
10
reviews and certainly why the staff hasn't raised
11
them, as questions.
MEMBER APOSTOLAKIS:
12
I would say, though,
13
that because of the limited time, I think necessarily
14
we are doing a spot-check.
MEMBER
15
16
STETKAR:
Oh,
yes,
obviously.
Sure.
MEMBER APOSTOLAKIS:
17
So the only question
18
in my mind is if I find problems with the few things
19
I am looking at, what can I conclude about the whole
20
-MEMBER
21
22
STETKAR:
That
exactly
is
the
whole issue --
23
MEMBER APOSTOLAKIS:
24
MEMBER STETKAR:
Yes.
-- because if you do a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
208
1
few spot-checks, you would find no problems or very,
2
very minor issues.
3
a warm feeling that essentially the processes work,
4
that
5
reviews, if there were initial errors, that they have
6
been found and corrected or maybe the models were
7
perfect when you started.
the
internal
But
8
9
You develop a sense of, you know,
reviews
if
and
doing
whatever
focused
identifies problems, you're right.
external
spot-checks
Maybe you have
10
been lucky in your spot-checks.
11
just very, very fortunate to find, you know, the only
12
issues.
13
lucky to do that.
But that's where you have to be exceedingly
MEMBER APOSTOLAKIS:
14
15
present?
16
review here?
By the way, who is
Is it your staff that actually did the
MR. CARUSO:
17
Maybe you have been
Originally Nick Saltos did
18
the Level I review.
19
group.
20
to his RAIs and new things in the Rev. 2 model.
I came on board and looked at the responses
Ed has been I think on board and our
21
22
And then he moved to the PRA
contractors in the Level 2 stuff from day one.
23
MR. FULLER:
Not from me from day one.
24
MR. CARUSO:
I guess Bob Paulo started it
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
209
1
out.
MR. WACHOWIAK:
2
3
one except for Marie.
4
MR. CARUSO:
5
MR. WACHOWIAK:
6
MEMBER
7
Nobody was here from day
Marie?
Marie.
APOSTOLAKIS:
MR. DUBE:
9
MR.
WACHOWIAK:
of
how
11
together we had in the initial -MEMBER
12
those
This
conversation
event
APOSTOLAKIS:
particular
trees
But
were
Nick
put
is
not
it's
the
involved anymore?
14
MR. CARUSO:
15
MEMBER APOSTOLAKIS:
16
MEMBER
No.
STETKAR:
Okay.
Okay.
But
17
standard bayonet the guy who isn't here.
18
-MEMBER APOSTOLAKIS:
19
That's not
No, no, no.
I think
we don't get into -MEMBER STETKAR:
21
22
you
No.
10
20
were
involved?
8
13
Don,
care who did it.
That's right.
We don't
It's --
23
MEMBER
24
performance-based.
APOSTOLAKIS:
It's
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
210
1
MEMBER STETKAR:
It's the review.
2
MEMBER APOSTOLAKIS:
3
MEMBER BLEY:
I guess two quick things
4
from me.
5
have one more general issue.
6
MEMBER STETKAR:
7
Do you have any more general issues?
10
MEMBER BLEY:
because
the
I wanted
specifics
I think that will be good
I
had
will
come
up
in
any
walk-through.
MEMBER
11
12
Not general.
I
to walk through that one sequence, but --
8
9
Performance-based.
STETKAR:
And
I
have
more
specifics on -MEMBER BLEY:
13
But I have the one general,
14
and I mentioned some of this last time.
I just want
15
to mention it again and see if you or if you could
16
the staff at the same time have any thoughts on it.
In chapter 6 on human reliability, the
17
18
good
things
19
"Adequate treatment of human actions in the PRA is
20
one
21
accident sequences and their relative importance to
22
overall risk."
23
that.
of
I
the
see
keys
are
to
a
statement
realistic
that
says,
understanding
of
I guess I would certainly agree with
And you also point out that due to the
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
211
1
current
status
2
analysis
3
design phase is preliminary.
of
of
human
the
ESBWR
actions
documentation,
carried
out
the
during
the
And that has to be true, but you then go
4
5
on to say -- and maybe this isn't inconsistent.
So I
6
would like you to comment on the first part.
And
7
then I have a second question after that.
8
You go on to say for type A human actions
9
"We review procedures related to tests" and all these
10
things, but I don't think there are any procedures.
11
And you go on to say for the type C's "A review of
12
normal
13
procedures."
special
failure
emergency
operating
I don't think you have any of those.
So have you done some of that or is that
14
15
all being saved for later?
And if you're doing only
16
the left-hand column kind of PRA, do you do these
17
human reliability analyses or do you plan to?
If you can address that?
18
Then I have
19
some specific things that bothered me a lot on human
20
reliability.
MR.
21
22
good.
23
there.
WACHOWIAK:
good.
That's
There should be things that bother you in
MEMBER BLEY:
24
That's
Fair enough.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
212
MR. WACHOWIAK:
1
I think the wording that
2
we used, "review of procedures," what we are talking
3
about there is typical maintenance-type procedures.
4
There certainly are not any ESBWR-specific procedures
5
at this point in time.
6
MEMBER BLEY:
Right.
7
MR. WACHOWIAK:
So we're looking at type
8
A for restoring things to service.
Typically in a
9
nuclear plant the procedures would say that you have
10
a full flow test or you have a checklist with a
11
secondary check to make sure that you have things
12
lined up properly and valve lineups and things like
13
that.
14
Those are the kinds of things we looked
15
at, were typical maintenance procedures for type A
16
and probably could have worded that better to say
17
that
18
procedures.
that
was
typical
practices,
rather
than
Now, on the post-accident, once again I
19
20
think
it
is
more
21
because
22
emergency operating procedures or abnormal operating
23
procedures, anything like that.
24
that we're in with the human factors engineering.
certainly,
boilerplate-type
as
Ed
knows,
language
we
there
don't
have
That is in this loop
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
213
We
1
make
our
best
guess
at
what
human
2
recoveries we would like to put into the model.
And
3
we using our tools calculate the time frame necessary
4
to perform those actions and then assign screening
5
values based on that.
6
actions goes to the human factors engineering group,
7
which then uses that as input to deciding what types
8
of displays, controls, ergonomics that they need to
9
put in to facilitate those types of actions.
And then that list of human
10
And ultimately they will talk about what
11
kind of procedures that they are going to generate
12
for those.
13
of the system functional requirements of the system.
But certainly those actions become part
As we go through not in the design space
14
15
because
all
this
procedure
development
isn't
16
happening in the design space, it's happening later,
17
but in the later phases of the PRA, we will then take
18
their information of how they have laid out those
19
actions and try to attach some performance-shaping
20
factors to what we already have in the PRA and go
21
back through another loop and calculate and see if
22
they still remain important actions if they indeed
23
were important actions.
24
human factors will go and do additional things in
And if they are, then the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
214
1
their procedure development that they may not do for
2
lower-significance
3
feeds back again for more performance shaping.
items
like
that.
And
then
it
4
And in the end, when we have the training
5
simulator developed because obviously we're going to
6
have to train operators to operate this plant before
7
it operates, we're going to have to train those.
8
then
9
observations and operator interviews and things that
10
we
can
actually
do
MEMBER BLEY:
12
MR. WACHOWIAK:
13
MEMBER BLEY:
the
simulator
Okay.
We don't -So some of this would come
after the design cert?
15
MR. WACHOWIAK:
16
MEMBER
17
of
you would be used to in a full-blown HRA.
11
14
some
And
BLEY:
It has to.
But
before
the
COL
is
complete?
18
MR. WACHOWIAK:
19
MEMBER BLEY:
20
MR. WACHOWIAK:
Yes.
Okay.
Well, no.
The timing of
21
all of this is that the COL doesn't necessarily need
22
to be complete.
23
out there on the emergency operating procedures and
24
the COL, but in general the human factors engineering
Now, I know there are some questions
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
215
1
is a DAC item.
And that would be completed and
2
submitted to the NRC for review after the licenses
3
are issued.
DAC are a special category of things yet
4
5
to be done that have follow-up commitments.
6
human factors engineering falls into DAC.
7
would be after the COL.
MR. OESTERLE:
8
9
the staff.
Yes.
But
So it
Eric Oesterle from
That is exactly where I was going to
10
discuss also because this dovetails right into DAC
11
ITAAC.
And this PRA, at least for me, is confusing
12
enough.
I was hoping that we wouldn't get there, but
13
--
14
MEMBER APOSTOLAKIS:
15
MR.
16
What's DAC?
OESTERLE:
Design
acceptance
criteria.
MR. WACHOWIAK:
17
Design certification and
18
COL applications contain the acceptance criteria for
19
the design, rather than the design.
20
a follow-up item to go in and verify that the design
21
does meet all of the acceptance criteria for review.
MEMBER BLEY:
22
And then there's
I am still trying to get my
23
handle.
And all of the DAC and ITAAC things will be
24
set
before
up
the
design
is
certified,
will
be
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
216
1
written out, established?
2
MR. OESTERLE:
fact,
we're
still
Yes, that's correct.
going
through
that
And,
3
in
review
4
process now to finalize the ITAAC for the ESBWR.
5
we have a meeting with GEH next week to continue the
6
discussions about that.
And
And we'll be here October 21st to talk
7
8
about
chapter
14,
a
9
discussion on the selection criteria and methodology
10
for identifying structure, systems, and components to
11
put into ITAAC and then also to discuss the staff's
12
review of the entire tier I document for ESBWR.
MEMBER BLEY:
13
14
at the same time?
15
going to be separate?
which
includes
a
Do they come together or are they
I like to talk about them
as ITAAC, DAC as one thing.
18
MEMBER BLEY:
19
MR. OESTERLE:
Okay.
DAC are a special subset
20
of ITAAC.
21
ITAAC.
22
verification-type activities.
And I call them design completion-related
The
And
vast
once
majority
we
get
to
of
the
ITAAC
--
are
they
get
incorporated by reference by a COL application.
And
23
24
of
And we'll see ITAAC and DAC
MR. OESTERLE:
16
17
section
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
217
1
following
issuance
of
2
requirements
3
complete all of those ITAAC prior to being allowed to
4
load fuel and to start up.
for
that
MEMBER
5
the
COL
license,
licensee
BLEY:
to
Back
there
are
successfully
to
the
human
6
reliability section in the PRA, there are a couple of
7
things that I just don't agree with.
8
put them on the table.
9
if you want to talk about some of them, fine.
10
And I want to
And after I put them all out,
If
not, we'll worry about them some other time.
The first one is talking about type C
11
12
post-initiating event, human actions.
13
the
14
operator
15
contributors
16
current LWRs.
passive
ESBWR
is
actions
to
such
should
the
risk
The nature of
that
not
post-initiator
be
profile
such
as
they
I don't know why that's here.
17
strong
are
in
It seems
18
maybe wishful thinking.
19
for the errors of omission, but things like errors of
20
commission
21
initiators I'm not fully convinced of that.
are
maybe
I think it is probably true
common
cause
human-induced
But it goes on, and then it talks about
22
23
errors of commission a little bit.
And it says --
24
let me see where to start this -- a commission error
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
218
1
-- and then some words that aren't really necessary
2
-- is considered insignificant when the plant has
3
emergency operating procedures.
And that kind of says we don't even have
4
5
to
think
about
it.
6
inconsistent
7
experiments that have been done, both at Halden and
8
at Westinghouse once upon a time.
with
And
that
experience
is
and
just
plain
with
some
There is an NRC NUREG by Emily Roth and
9
10
Lumau
that
ran
a
bunch
of
operators
11
simulators with fully vetted procedures.
12
difficult scenarios.
13
procedures pretty well.
through
They were
And people wandered from the
14
And the last thing along this same line
15
is no dependencies are considered for human-related.
16
And you haven't done all of this yet.
I think this
17
is in the future for human-related basic events due
18
to type A and type C actions in the same minimal cut
19
set due to highly differentiated time frames and the
20
low combined probability.
21
In general, that is probably true, but in
22
the experience base, you find some separated by even
23
weeks, where the setup from one event, the effect on
24
the person who is now involved in the second event,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
219
1
links over.
So I think while it's often true, there
2
3
are cases where it is not true.
4
you get to that stage of the analysis, instead of
5
just making a blanket statement, not thinking about
6
them, you give it some real thought.
MEMBER APOSTOLAKIS:
7
8
9
And I sure hope when
I agree with Dennis.
I think this section really needs a good editing
job.
Would you go back to a page that you had
10
11
earlier on?
That was a good example, too.
I don't
12
know.
13
can you start from page 1 and keep going down?
14
Look at this.
Where you raise human reliability analysis,
By
15
Stop.
Stop.
virtue
reliability
of
Stop.
its
capacity
17
reliability, the probabilistic safe analysis provides
18
an unsurpassable way of studying the --
20
BLEY:
Then
and
combine
human
MEMBER
systems
to
16
19
with
No.
you're
component
better
not
dismiss them.
21
(Laughter.)
22
MEMBER
23
enthusiast.
24
finished?
APOSTOLAKIS:
Unsurpassable.
That
You
guy
know,
was
are
an
you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
220
MEMBER BLEY:
1
2
general area, yes.
MEMBER APOSTOLAKIS:
3
4
the
5
Should I make them now?
detailed
I
Before you run into
have
two
minor
points.
Already said, what George
said and what -MEMBER APOSTOLAKIS:
8
9
sequence,
MEMBER BLEY:
6
7
I am finished with this
Yes.
Sorry, sorry, sorry.
Go.
MR. WACHOWIAK:
10
Thought maybe you thought
11
his questions were unsurpassable.
12
that we did not address errors of commission and the
13
words that we have in there saying that they aren't
14
significant contributors, those I remember looking
15
at.
16
what it is.
We got that from a reference.
While it is true
And I don't know
Understand where you are now.
Errors of commission are in a couple of
17
18
different areas.
In the things that we looked at for
19
the precursor types of type A errors, while they're
20
all listed as errors of omission, those would also be
21
the same types of things that an error of commission
22
would cause.
23
those things that would unset our acid traps for core
24
melts, if you will, that take away the setup for the
So the aspect is there looking for
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
221
1
plant.
As
2
I
said,
that
entire
list
of
3
pre-initiator operator actions have been given to the
4
human factors engineering group.
5
their
6
setting up alarms and indications and things in the
7
control room.
lists
of
things
that
And that's one of
they
start
with
for
So we try to address it that way.
8
Numerically right now I really have no
9
idea how we would address errors of commission at
10
this point without, you know -MEMBER BLEY:
11
12
can look at.
13
variety of them.
14
When
There are some things you
I won't mention them here.
of
you
10-4,
are
then
calculating
a
There is a
core
damage
human-induced
common
15
frequencies
16
failure of 10-6 doesn't matter too much.
17
calculating 10-8, the way you are going to break this
18
plant somehow means something unusual has to happen.
When you are
19
The likelihood that it is a whole bunch
20
of random things lining up seems pretty small to me
21
when you have the chance of some activities by a
22
human in the plant maybe defeating some of those.
23
think you've got to look really hard.
MR.
24
WACHOWIAK:
And
we
do
have
I
some
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
222
1
protection
against
2
safety-related side for the automatically initiated
3
actions, we don't allow, the digital I&C does not
4
allow,
5
sequences.
6
done.
the
that
operators
They
go
in
to
to
the
design.
interrupt
completion
any
On
of
before
the
those
they're
7
So that's one of the things where I think
8
we have seen in the past, where the automatic systems
9
have initiated something and then the operator said,
10
"Oh, I didn't want that to happen.
MEMBER
11
12
BLEY:
We
Turn it off" --
have
certainly
seen
that.
MR.
13
WACHOWIAK:
--
or
it's
about
to
14
happen and they bypass the thing that is going to
15
make it actuate.
16
actions.
17
actually captured in words or in numbers in the PRA,
18
that is one of the design philosophies that we have.
19
So that helps move us towards the words that we used
20
And we don't allow those types of
So that is one place where while it's not
in the document.
Wasn't
21
there
one
other?
Oh,
the
22
difference in time as a screening value for saying
23
that
we
24
Yes.
We'll have to look back at how we write that so
don't
have
a
dependence
between
actions.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
223
1
that in the future when operator actions are looked
2
at, that we don't use that as a blanket statement
3
that just because of time, it should be screened.
4
But
5
particular model, we looked at those kinds of actions
6
to say, you know, it is separated in time, but it's
7
running the fire pump here versus turning on the,
8
backing up the depressurization valves over there and
9
just --
our
limited
use
of
operator
MEMBER BLEY:
10
of
in
this
I think that is a much
11
better
12
because that is different mindsets getting involved.
13
kind
actions
argument
than
just
the
time
lime
So I think --
14
MR. WACHOWIAK:
I understand.
15
MEMBER BLEY:
16
MR. WACHOWIAK:
Yes.
And here remember with
17
ESBWR, the difference in time is days in many of
18
these cases.
MEMBER BLEY:
19
20
to be back on shift, he can be linked.
MEMBER STETKAR:
21
22
But if the same guy happens
Or each one successively
asks him, "Hey, what's wrong?"
23
MEMBER BLEY:
24
MEMBER
Yes, exactly.
STETKAR:
And
then
I
tell
you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
224
1
exactly what is wrong.
2
wrong.
3
already because nothing has changed.
4
MEMBER BLEY:
5
MEMBER
6
And you tell him what is
And he comes back.
And I know what is wrong
And those have happened.
STETKAR:
That's
the
kind
of
thing.
MR. WACHOWIAK:
7
It's not just operators
8
that are involved at that point because we have the
9
emergency response organization and --
10
MEMBER BLEY:
11
MEMBER APOSTOLAKIS:
12
As long as we don't tell -There is something
here that confuses me a little bit.
13
MR. WACHOWIAK:
Okay.
14
MEMBER APOSTOLAKIS:
And I looked at the
15
analysis of the isolation condenser.
16
--
17
MR. WACHOWIAK:
18
MEMBER APOSTOLAKIS:
19
MEMBER BLEY:
20
Where are you, George?
Table 4.2-6.
In the isolation condenser
section, --
21
MEMBER APOSTOLAKIS:
22
MEMBER BLEY:
23
MEMBER
24
In table 4.2-6
event.
Yes.
-- which is 4.2?
APOSTOLAKIS:
There
Yes.
is
How is it easier for you to find it?
a
top
Do you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
225
1
want a top event, that kind of thing?
2
you want?
MEMBER BLEY:
3
That's what
I think we're going to get
4
through the event -- go ahead.
The fault trees,
5
we've got a lot of questions on the fault trees.
MEMBER APOSTOLAKIS:
B-32-2LOOPSFAIL.
8
MEMBER BLEY:
On page 4.2-147.
9
MEMBER
6
7
Do
you find it?
Okay.
APOSTOLAKIS:
The
description.
10
The description says, "Three heat exchangers loops,
11
remove heat."
Right?
12
MR. HOWE:
13
MEMBER
14
Okay.
Yes.
APOSTOLAKIS:
this
is
a
success, right?
15
MR. HOWE:
16
MEMBER APOSTOLAKIS:
Four.
Right.
That's a success.
17
total?
18
Now, if we go to the fault tree --
And how many have
So if two fail, you are in trouble.
MR. CARUSO:
19
20
So
Excuse me.
Wasn't there an
assumption that there is always one not operable?
21
MR. WACHOWIAK:
No, not in the PRA.
22
MEMBER APOSTOLAKIS:
There is a fault
23
tree for all loops failing.
Isn't there a fault tree
24
for two loops failing somewhere in the --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
226
1
MEMBER BLEY:
2
MEMBER APOSTOLAKIS:
3
MEMBER STETKAR:
4
on the table.
Yes, page 213.
Okay.
Sheet 50, like it says
Sheet 50, .pdf file, page 579.
5
MEMBER APOSTOLAKIS:
6
other things, but they can wait.
MR. WACHOWIAK:
7
Okay.
Fine.
I have some
It's always difficult to
8
try to follow through the fault trees on the paper
9
copy.
MEMBER STETKAR:
10
11
you can search.
MR. WACHOWIAK:
12
13
MEMBER STETKAR:
CHAIRMAN CORRADINI:
well.
No.
You're doing so
Keep on going.
MEMBER STETKAR:
18
19
Do you want to take a
break now?
16
17
Search into those, yes,
but it's a pain in the neck to try to follow up.
14
15
The good thing is .pdf
I heard somebody say,
"Break."
20
DR. KRESS:
We took a break.
21
MEMBER STETKAR:
Rick, I don't know.
I
22
know there were some communications in the last week
23
back and forth regarding these sequences.
24
what you brought.
And I see
And I thought -- and that's good.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
227
1
2
I mean, that's fine, but I'm not sure that we're
actually going to discuss that.
3
MR. WACHOWIAK:
4
MEMBER STETKAR:
Okay.
One of the questions
5
that we had in particular was -- and I was using it
6
as context -- is this sequence FDW-0033, which the
7
feedback, at least as it was finally filtered to us,
8
was, gee, you know, there are no cut sets for that
9
sequence.
So it's not worthwhile discussing it.
10
Well, I hope we communicated back that we
11
wanted to understand why there are no cut sets for
12
that sequence because that's more important to me
13
than the sequence.
14
sequence, but I want to understand why there are no
15
cut sets for that particular sequence.
16
MR. WACHOWIAK:
the
way
it
The sequence is an interesting
got
back
And the communication I
17
guess
18
exactly that way.
19
I think what we said back was, were they aware that
20
there were no cut sets for these sequences --
to
you
wasn't
The sequences were presented.
21
MEMBER STETKAR:
22
MR. WACHOWIAK:
23
around
And
We were not.
-- and do you want to
dust them or propose different ones?
MEMBER STETKAR:
24
And the answer to that
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
228
1
was no.
MR. WACHOWIAK:
2
Well, no.
The answer was
3
go ahead and propose the other ones, but they still
4
want to know why there are no cut sets.
5
MEMBER STETKAR:
6
MR. WACHOWIAK:
7
MR.
HOWE:
description right here.
12
--
MR.
short
kind
of
What it really comes down to
Can you tell me where you
HOWE:
No.
This
is
just
a
stand-alone page.
MR. WACHOWIAK:
17
the question.
This is just an answer to
In response to the e-mail --
MEMBER
19
STETKAR:
We
don't
have
this,
right?
MR.
21
22
a
are in the report?
15
20
There's
MEMBER BLEY:
13
18
Why are there no cut
sets for that sequence?
11
16
So we have an explanation
MEMBER STETKAR:
10
14
Yes.
of why they --
8
9
Okay.
WACHOWIAK:
No.
Because
we
were
coming here.
23
(Laughter.)
24
MR.
HOWE:
What
it
boils
down
to
is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
229
1
really the common cause modeling that's used combined
2
with the success criteria of both that the VIGDCS top
3
and the VE top there.
So you have the --
4
MEMBER STETKAR:
5
CHAIRMAN CORRADINI:
6
got to stop you.
7
said.
Wait a minute.
I've
Maybe he understood what you just
I have no clue.
8
(Laughter.)
9
MR. WACHOWIAK:
MR.
10
Let me ask you first.
HOWE:
Okay.
This
Take a step back.
sequence
that
we're
11
talking about -- and there are a few of them in the
12
model like this -- is examples T feedwater, 033,
13
where we have successful early injection of GDCS, and
14
then failure of what we call the long-term GDCS,
15
which could either be achieved with two of the three
16
pools or one pool, one equalizing line.
17
have failure of that second function.
And so you
And then that goes to core damage.
18
And
19
then the question is, how come those sequences have
20
zero cut sets in the quantified model results?
21
the answer is theoretically you can get cut sets in
22
those sequences.
23
have an extremely low truncation value.
24
of explain why that is.
And
It's just that you would have to
I will kind
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
230
1
The common cause modeling that we use
2
really models combinations of like any two of the
3
eight valves, any three, and beyond that, it just
4
says, "common cause of all valves failed to open."
5
Based on the generic multi Greek letter
6
factors that we have -- and that's for the common
7
cause
8
document also says, really, to use multi Greek letter
9
factors of 1.0 after you've gotten to four valves
10
modeling
--
the
EPRI
utilities
requirement
anyway or four components.
MEMBER APOSTOLAKIS:
11
Now you have a big
12
problem with the actual numbers you use.
13
some guideline delta that are much, much lower than
14
what you just said.
But
15
also
let
me
come
back
You have
to
this
16
discussion.
So you are saying that in a system where
17
there is a common cause failure, there are no cut
18
sets?
MR. HOWE:
19
No.
See, what happens is the
20
primary failure mode for both early GDCS --
21
MEMBER APOSTOLAKIS:
22
MR. HOWE:
23
MEMBER APOSTOLAKIS:
24
MR. HOWE:
Right.
-- and late GDCS, -Right.
-- those failure modes, the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
231
1
dominant ones are all the same.
2
MEMBER BLEY:
The common cause failure.
3
MR.
Right.
HOWE:
The
common
cause
4
failure because early GDCS, to fail that, you have to
5
fail
6
requires either that you fail six or that you fail
7
four and that you fail all of the equalizing lines.
8
So still the dominant failure mode there is CCF of
9
all injection valves.
seven
of
eight
valves.
10
CHAIRMAN CORRADINI:
11
MR. HOWE:
12
CHAIRMAN CORRADINI:
And
the
later
GDCS
So let me translate.
Sure.
So you are saying
13
that you didn't do it because it is included in stuff
14
that was more important than it?
MEMBER APOSTOLAKIS:
15
16
What does it do?
I
mean, that's where I'm lost.
17
MR. WACHOWIAK:
No, no.
18
CHAIRMAN CORRADINI:
The cut sets don't
19
appear.
I mean, I'm just trying to get back to your
20
original answer.
21
properly.
22
are essentially subsumed into something that was more
23
probable.
I'm sorry that I can't explain it
But the cut sets don't appear because they
MR. WACHOWIAK:
24
No.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
232
MR. HOWE:
1
Well, no.
It's just saying
2
that the dominant failures are the same for both
3
early and late.
4
one, you're just going to fail early.
5
see those cut sets that had success early, failed
6
late.
So if you're going to fail either
7
MEMBER APOSTOLAKIS:
8
CHAIRMAN
9
10
That makes sense.
CORRADINI:
Just
to
finish,
because they are so low-probability they got screened
away?
11
MR. HOWE:
12
MR. WACHOWIAK:
13
You'll never
Right.
Let me try one more way.
There would be several cut sets in early and the
14
late.
Okay?
One is a common cause failure of all
15
valves.
16
you have failure of one valve, common cause failure
17
of seven.
18
failure
19
failure of six.
That's common to both of those.
of
I don't have the numbers quite right but
two
individual
Okay.
20
And then
valves,
common
cause
You could get some combination of
21
those things that might show up, but they're all
22
below the 10-15 truncation level because the ones that
23
we
24
included in both the success path and the failure
do
have
that
would
show
up
in
the
model
are
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
233
1
path.
So the common cause failure of all is
2
3
already handled by the success of the early.
4
we're left with are the individual valve failures.
5
And since so many valves have to fail, you end up
6
with cut sets that are below our 10-15 truncation
7
level.
MEMBER STETKAR:
8
So all
I think the answer to
9
the question is the fact that your truncation level
10
is set at 10-15 and the cut sets that -- there are
11
valid logical cut sets that contribute.
12
them are greater than 10-15.
Okay.
13
That's
understanding
important
that
my
sequence
was
original
15
logically -- it was a logical null set.
16
happen because of the way common cause is modeled.
And that's not true.
17
18
null set.
19
sequence.
20
that out Sunday.
this
because
14
It can't
It is not a logical
There are logical contributors to this
So I kind of got that.
I sort of figured
So I would like to look at some of the
21
22
was
Just none of
logical contributors.
And if I look at --
23
MEMBER BLEY:
John?
24
MEMBER STETKAR:
Yes?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
234
MEMBER BLEY:
1
2
Would it
be helpful for the rest of the Subcommittee -MEMBER STETKAR:
3
4
Can I interrupt?
Yes,it would.
What do
you suggest?
MEMBER BLEY:
5
-- if you took us along the
6
event tree and then told us where you are going to
7
focus
8
about?
os
they
can
MEMBER
9
understand
STETKAR:
what
Yes.
you're
talking
Okay.
The
10
particular place that I'm focusing on for the rest of
11
the Committee -CHAIRMAN CORRADINI:
12
13
like to stay with you.
MEMBER STETKAR:
14
15
I didn't ask about BiMAC
crap this morning.
16
(Laughter.)
17
CHAIRMAN CORRADINI:
18
MEMBER STETKAR:
19
Yes, you did.
If I look at the loss of
feedwater -MEMBER APOSTOLAKIS:
20
21
For those that would
What are you looking
at?
MEMBER
22
STETKAR:
I'm
looking
at
the
23
second event tree, second event tree, 22.319, loss of
24
feedwater event tree.
And if you can read that, the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
235
1
two top events that I am particularly focusing on
2
right at the moment are called VI and VE.
MEMBER APOSTOLAKIS:
3
4
Oh.
I see the VE.
Yes.
MEMBER STETKAR:
5
And, in particular, if
6
you -- okay.
Directly under NEDO and directly under
7
33.201 appear VI and VE.
8
that I'm interested in because I kind of want to
9
follow
dependencies
and
The particular condition
how
the
models
are
put
10
together is a sequence that if you trace through the
11
tree, you come in from the left.
You go up at the first branch, which is a
12
13
reactor scram.
14
the up branches.
15
the piece of paper right now.
16
VI.
You see up branch on
CORRADINI:
So
you
are
VI-TOPINJ.
MEMBER STETKAR:
19
20
that
21
injection,
22
injected.
branch,
which
means
means
the
Right.
I
GDCS
have
And go up on
successful
pools
CHAIRMAN CORRADINI:
23
24
You are sort of in the middle of
CHAIRMAN
17
18
You go down and just keep following
enough
of
early
them
Two out of eight
lines, one out of --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
236
MEMBER STETKAR:
1
Enough went
2
in.
3
you come to now a question on VE.
4
down
5
FDW-0033, the sequence that I'm using as the context
6
for this, you then need -- so I'm down on VE, top
7
EQU.
8
you have to go down on VL-TOPINJ and you have to go
9
down on VM-TOPINJ.
10
And then I'm headed down.
Whatever.
on
branch.
And
to
get
And
And I am going
eventually
Then you have to go down on UD-TOPINJ.
to
Then
And you eventually pop out on
FDW-0033.
MEMBER BLEY:
11
12
that
So I'm up on VI.
Which is the first one of
that group that is not a success.
MEMBER STETKAR:
13
Which is the first one
14
of that group that is not a success.
And it's one
15
that goes to CDI, which is a low-pressure late melt
16
with no water in the bottom, the lower part of the
17
drywell.
18
and level 2 also.
So it is sort of interesting for a level 1
19
Right at the moment I am interested in
20
the interactions between VI-TOPINJ up and VE-TOPEQU
21
down, which is sort of what we are talking about,
22
which is why I was trying to find out when you said
23
there were no cut sets whether it was a logical null
24
because of the way common cause is modeled or whether
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
237
1
it is just numerically so small because of all of the
2
combinations.
And your answer is that it is numerically
3
4
too small to exist.
So there are logically valid --
MR. WACHOWIAK:
5
It's both, though, too,
6
because we don't have all of the different common
7
cause splits.
8
because we have all is the only way to get the common
9
ones, then the only cut sets that would be left would
10
We only have one, two, three, all.
So
be the individual valves.
11
MEMBER STETKAR:
12
MEMBER BLEY:
13
MEMBER STETKAR:
14
MEMBER BLEY:
I understand.
Can I?
Sure.
And if you'd had all of
15
them fail, you would have gone down at the first
16
branch there, VI.
17
MEMBER STETKAR:
18
MEMBER BLEY:
19
That's right.
And you wouldn't be up from
the stop area.
MEMBER STETKAR:
20
Right.
So know that if
21
I look at now the fault tree for TOP event VE-TOPEQU
22
-- you will have to excuse me because I have to pull
23
it back up here because I got -- I guess you can tell
24
me what page we're on.
This has to search for a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
238
1
while.
Yes.
You might want to for everybody else.
2
We can look at it on here.
3
The fault tree for top event VE-TOP --
4
MEMBER BLEY:
5
Look at the fault tree you
were looking at, John.
MEMBER STETKAR:
6
That is not the same
7
picture of a fault tree that I was looking at.
I
8
believe it is probably logically equivalent, but it's
9
not clear to me that it's logically equivalent.
10
MR. LI:
11
I think I want to make a clarification.
12
This is Jonathan Li.
This is revision 3, which is -MEMBER STETKAR:
13
Yes.
And I was looking
14
at it.
I've indeed printed out revision 3 and all of
15
the revision 3 documents that we had.
16
different from the revision 3 --
17
MR. HOWE:
18
MEMBER STETKAR:
19
printed this out.
20
changed in section 22.
Right.
I printed this out.
I
It's in section 4, and it was not
21
MR. WACHOWIAK:
22
MEMBER STETKAR:
23
And this is
It wasn't changed in 22.
That's indeed correct.
I made sure that that was absolutely true.
MR. LI:
24
This is 22.3, actually.
No.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
239
1
This
is
actually
we
discovered
2
something with systems, but we didn't show every -MEMBER STETKAR:
3
change
in
22.4
Oh, you didn't show?
So
4
you said something in words, but you didn't show me
5
how you actually changed the model?
MR. LI:
6
7
We evaluate every single
change to why it's -MEMBER STETKAR:
8
9
No.
No, no, no.
you made a change in words, but you didn't show me
10
the actual logic change.
11
PRA?
How can I then review the
12
MR. LI:
13
MEMBER APOSTOLAKIS:
14
that.
Let me understand
There is a -They just have to give
you the results.
MEMBER APOSTOLAKIS:
17
18
That is why I am --
MEMBER STETKAR:
15
16
You said
And then there is a
-MEMBER STETKAR:
19
There was a rule that
20
they set up.
They kept chapter 4 the same from rev.
21
2 to rev. 3 because the vast majority of things --
22
MEMBER APOSTOLAKIS:
23
MEMBER STETKAR:
24
Yes, were the same.
-- were the same.
And
then they created this chapter 22 that documents all
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
240
1
changes from rev. 2 to rev. 3, all changes.
So that
2
if any change was made from rev. 2 to rev. 3, it
3
actually is supposed to be in chapter 22.
4
So now I seem to see a change that was
5
made to a fault tree that in chapter 4 is identical
6
to the way it was in rev. 2, as it should be.
7
chapter 22, I don't see a new picture of the new
8
fault tree, which starts to bother me a lot.
MR. WACHOWIAK:
9
And in
I know we didn't include
10
all of the pictures of all of the fault trees because
11
chapter 4 is -MEMBER STETKAR:
12
But if I am the staff
13
and I am supposed to be reviewing rev. 3 of the PRA
14
and you have made a change to a fault tree in the
15
real PRA in the real computer and not shown me that
16
changed fault tree, how can I review the rev. 3 PRA?
17
MEMBER BLEY:
18
MEMBER STETKAR:
19
To test your conclusion.
To test your conclusion.
I don't have the computer model in front of me.
I
20
can't look at the real computer model.
21
pictures of the fault trees, which I have, are now
22
not consistent with what is really in the computer.
23
So I can't.
How can I do my review?
MR. CARUSO:
24
The only
This is Mark Caruso of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
241
1
staff.
I guess you're asking that question of me.
MEMBER STETKAR:
2
No.
They didn't give it
3
to you to review.
4
you have not seen.
5
change to any event tree or any fault tree between
6
rev.
7
chapter 22, show me the picture of the fault tree.
8
Then it cannot be reviewed.
9
picture and say, "Yes.
10
2
and
rev.
You can't review something that
And if they made any logical
3
and
did
not
document
that
in
I cannot look at the
You said in words that you
made this change."
In fact, this comes back to in words in
11
12
chapter 22, you said, "Oh, yes.
13
the
14
failures of the injection lines with the GDCS pools
15
because we need to do that.
16
doesn't make any difference to the results.
GDCS
models,"
for
We made a change to
example,
that
correlated
It's a small change.
It
17
When I was reviewing rev. 2, I identified
18
that as a fundamental logic problem in the GDCS fault
19
tree.
So I read those words.
20
gee.
I wonder how they did that.
21
actually correlated those failures."
I
22
must
have
a
And I said, "Well,
fault
I wonder how they
tree,
then,
It's not here.
that
23
shows me how they did that.
I can't
24
see how you did that, whether you did it correctly,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
242
1
incorrectly, in an ad hoc manner.
2
here.
3
or not that's done properly.
It's not documented.
It's just not
I can't review whether
4
In particular -- I mean, I as going to
5
bring that up in about ten minutes, but we fell into
6
it.
7
up, which apparently has disappeared in VE-TOPEQU, is
8
that that model, at least everything that I have seen
9
of it for the last year, had a nebulous single basic
In particular, the thing I was going to bring
10
event
hanging
out
there
that
11
failure of all injection valves.
said
common
It was only a single basic event.
12
cause
It
13
wasn't linked to the front model.
And it had a
14
number assigned to it.
15
10-5, which is indeed a factor of 10 lower than common
16
cause failure of all injection valves in the real
17
injection model.
That number was 1.5 times
So there is a disconnect there.
18
There is
19
indeed a cut set that pops up that has common cause
20
failure of all of the equalizing valves with this
21
single basis event hanging out there by itself, which
22
indeed is logically correct.
23
inconsistent.
But numerically it is
And now it seems to have just completely
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
243
1
disappeared form the model.
2
current
3
failures
4
incorrectly.
5
removed them, but I can't see the new model.
model
because
The
new
old
for
model
model
all
used
seems
to
of
the
to
but
just
have
other changes.
8
9
the
accounts
So I am not kind of troubled about what
6
7
actually
So I'm not sure if the
MR. WACHOWIAK:
I understand your point
MEMBER STETKAR:
I mean, now I can't have
that --
10
11
a problem that I used to have because what I used to
12
have a problem with is no longer in the model, but I
13
can't tell whether the new model has problems or not
14
because I can't see the new model.
CHAIRMAN
15
16
CORRADINI:
So
is
this
an
isolated incident?
MR. WACHOWIAK:
17
Probably not because --
18
well, and the reason is that the vast majority of the
19
pages of the PRA are these fault tree pictures.
20
if we had gone through and printed out all of the
21
pages again, it's just as difficult as going back
22
through and revising the model to look at all of the
23
pages again.
And
So we tried to end up with the right
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
244
1
balance of what we should describe in words.
2
there are other places where we have small sections
3
of logic to explain the change.
One
4
thing
that
I
think
we
And
haven't
5
stressed yet on this is that just submitting rev. 3
6
was not the total of the agreement that we had on how
7
we were going to perform this review of the changes
8
that happened to the plant, which are now reflected
9
in the PRA over the process of doing the DCD.
After the staff had a chance to look at
10
11
rev.
3
and
we
had
12
documentation
13
everything that is in 22 and actually putting it into
14
the
15
schedule to come out and take a look at that.
--
sections
1
by
a
chance
"clean
through
21
to
up,"
--
16
CHAIRMAN CORRADINI:
17
MR. WACHOWIAK:
clean
I
they
up
mean
have
an
our
taking
audit
"They," the staff?
The staff, come out to
18
GEH and spend whatever time they need looking at
19
those things to, in fact, ensure that what we said in
20
22 was implemented properly.
So
21
that
hasn't
22
understand
that
you
don't
23
information
and
how
you're
24
something like that.
happened
have
going
yet.
access
to
I
to
that
deal
with
I understand where you are.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
245
1
MEMBER STETKAR:
2
MR. WACHOWIAK:
3
MEMBER STETKAR:
4
Okay.
I've got that.
Okay.
I mean, I understand.
You know, that's annoying, but you get annoyed.
Let me back up to -- and I hate to do
5
6
this, but let me back up to rev. 2 if I can.
7
hate
8
understand.
9
mentioned
to
do
this
to
you
guys,
but
I
And I
need
to
Two things I'm trying to get at, as I
earlier,
are,
does
the
current
PRA
10
accurately model the current design as we understand
11
the design?
12
the risk from the current version of the design,
13
recognizing that both the PRA and the design have
14
been evolving over time?
So is it a reasonable representation of
The second part is indeed to have some
15
16
confidence
that
the
staff's
review
of
the
PRA,
17
recognizing that it's been evolving, has indeed been
18
reasonably thorough, that indeed the review process
19
has worked because that ultimately I think is one of
20
the functions that we provide here.
21
ACRS's job to review the ESBWR PRA.
It's not the
I hope not.
22
So if I back up to rev. 2, if I can do
23
that, in the rev. 2 fault tree for top event EQU, if
24
I solve the combined fault trees, if I did this right
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
246
1
in my head, if I solve the combined fault trees for
2
success of INJ and failure of EQU, I have essentially
3
a cut set that has one contributor as common cause
4
failure of all four equalizing valves.
5
basic event, you know.
That's a
And I have this other basic event that is
6
7
called E50, SQV-CF4 open.
And I think that that --
8
and it is a single basic event.
9
than that.
It's nothing more
And I think that that basic event was
10
supposed to be some sort of surrogate for common
11
cause failures of all the injection valves.
12
true?
MR.
13
HOWE:
That
one
was
Is that
actually
a
14
vestige from the rev. 1 model that we -- that was one
15
of the things that we picked up in rev. 3.
16
removed that.
17
MEMBER STETKAR:
18
MR. HOWE:
Why did you?
common cause modeling that we do now.
20
a leftover from rev. 1.
MR.
22
Okay.
I'm just saying we have the
19
21
And we
WACHOWIAK:
We
used
This was just
a
different
common cause method.
23
MR. HOWE:
Correct.
24
MEMBER STETKAR:
Okay.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
247
CHAIRMAN CORRADINI:
1
2
Is that what you're saying?
3
MEMBER STETKAR:
4
It came in rev. 1.
It
stayed in rev. 2.
5
MEMBER BLEY:
6
MEMBER STETKAR:
7
Was it duplicative?
It got put back.
It's not right in rev.
2.
8
MEMBER BLEY:
9
MEMBER STETKAR:
And
it
got
It's not right in rev.
10
2.
11
something I just learned today because up until ten
12
minutes ago, I was convinced it was still in rev. 3
13
because it was the only thing I could see in rev. 3.
The
14
pulled
It's not right in rev. 2.
staff,
rev.
however,
problem
which
didn't
16
modeling
across
17
couldn't understand what the basic event was.
thought
two
it
the
seem
this
those
in
3,
identify
I
a
in
15
18
as
out
common
functions
was
for
to
cause
because
common
is
I
cause
19
failures of the injection valves, but if it was,
20
recognizing that it's -- then its numerical value was
21
a factor ten times lower than the real common cause
22
failures
23
George's concerns might be too low anyway.
of
the
injection
valves,
which
based
on
But there was that numerical disconnect.
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
248
1
And I'm trying to understand.
I understand that you
2
have taken it out.
I don't know whether the current
3
solution picks up all of the -- it probably does, the
4
current solution.
CHAIRMAN CORRADINI:
5
So can I take it one
6
level up before we do one more specific thing?
7
said there's an audit.
9
When is the audit?
MR. WACHOWIAK:
8
yet.
You
It hasn't been scheduled
We sent a note or it was packaged in with a
10
bunch of other things, other NEDO commitments, that
11
said, "We will be ready after."
12
staff hasn't responded back for which date they want
13
to come after that.
CHAIRMAN CORRADINI:
14
15
We gave a date, and
So I guess I have a
question for the staff about this.
I am sorry you --
16
MEMBER STETKAR:
No, no.
17
CHAIRMAN CORRADINI:
18
this so at least my brain starts reengaging.
I need to talk like
So I guess I hear when an audit happens
19
20
like this, can you explain?
21
analysis
22
documentation and one of the parts of the team looks
23
at the PRA when you come on site?
of
the
sequences
Is it primarily level 1
or
is
it
all
forms
of
The reason I'm asking, that leads me to
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
249
1
another question.
2
just the PRA that they are going to be coming in to
3
look at?
4
5
I just want to understand.
Is it
MR. CARUSO:
Well, this is Mark Caruso
I
admit
again.
6
have
to
I
haven't
had
a
7
conversation with Hossein about what he had in mind.
8
We have done staff audits of PRA or parts of PRAs
9
for operating plant issues, where we have an audit
10
procedure, we go out and we look at the files, we do
11
spot-checks, and then there is an audit report.
12
would be probably a model that would be used if that
13
is the intent.
14
because this wasn't discussed with me.
That
I can't tell you what the intent is
So I am presuming if we are going to do
15
16
an audit, that's what we would do.
17
visited
18
discussed
19
quality
20
number of other things.
21
not an audit.
once
a
before,
lot
of
assurance
is
things.
more
We
procedures.
CHAIRMAN
22
which
We went and
of
did
We
a
look
did
Okay.
at
look
But it was a visit.
CORRADINI:
--
we
the
at
a
It was
But
this
23
would be -- the reason I'm asking the question is --
24
and so here is where I am going -- hearing all of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
250
1
this, I am getting hesitant to write a letter in
2
October until the audit occurs so I can turn back to
3
the staff and say, "Okay.
4
had an X day meeting with GEH."
You have had a visit.
You
5
And now you guys are on the same page.
6
You understand what was put in, what was taken out.
7
What were the details that were there, all the stuff
8
that we were talking about?
So I want to make sure that that's a
9
10
sanity check for me.
11
that needs to get done before we hear from the staff
12
that they are satisfied with the PRA rev. 3 results?
MR. CARUSO:
13
14
15
So is that the sort of thing
Yes.
This is something we
are going to do before we sign off on our final SER.
Right?
We're at the point now where -CHAIRMAN CORRADINI:
16
but
17
part,
18
saying, is this something you're going to do before
19
we
20
hunky-dory to a certain point with open items or is
21
this something that is going to be out there much
22
longer in time?
hear
I'm
from
taking
you
in
a
step
I understand that
further
October
that
back.
everything
I'm
is
23
I'm asking this in front of the members
24
so I get a feeling for what the expectation is in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
251
1
October.
MR. OESTERLE:
2
3
Mike, I'm not aware of any
audits scheduled between now and October.
4
CHAIRMAN CORRADINI:
5
MR. CARUSO:
Okay.
But perhaps I might suggest
6
that you might want to consider asking -- we're still
7
considering this and working out the details with GEH
8
--
9
performing this audit of the PRA as part of your
10
asking
the
staff
what
their
plans
are
for
interim letter.
11
CHAIRMAN CORRADINI:
12
MR. WACHOWIAK:
Thank you.
And so to get back onto
13
this, the date that we have given so far is about
14
Thanksgiving is what we expect to have this done
15
because one -CHAIRMAN CORRADINI:
16
17
serve turkey?
Sorry.
MR. WACHOWIAK:
18
Are you going to
-- because one of the
19
things that we have to do is we have to do is we have
20
to put the document through our change process.
21
to your point, when they would go and look at section
22
4 as updated, what they would have is 22 from before,
23
which has in text what the changes are.
24
document itself will be marked with rev bars.
So
And then the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
252
1
And since they're coming for the audit,
2
they'll probably have the redlined, strikeout version
3
of the file itself and will be able to say, "Okay.
4
You said you were going to make this change.
5
this change is in here.
6
has been made into the document."
I can see how this change
So that works great with the things that
7
8
are generated in Word.
9
of revision control on it.
10
Yes,
CAFTA doesn't have that kind
So we're going to have to
figure out how to -MEMBER STETKAR:
11
That's what bothers me
12
because in many cases, when I checked in chapter 22,
13
there are indeed pictures of logic that have been
14
changed.
15
want to ask a second question before we -- what time
16
are we supposed to take a break, 3:00?
I'd flagged the words about -- in fact, I
17
CHAIRMAN CORRADINI:
18
MEMBER STETKAR:
Very soon.
Okay.
Well, after we
19
come back from the break, then.
But I flagged a
20
couple of words.
21
highlighted.
22
might have changed the fault tree logic here, but I
23
don't see any changes.
24
picture in chapter 22."
I just printed out the pages and
And I said, "Gee, that sounds like they
And I don't see any revised
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
253
1
So
when
you
2
reconciliation
3
between 22 and 4, certainly wherever the pictures are
4
changed is a flag to make sure that the appropriate
5
new pictures get added or revised in chapter 4.
process
go
or
through
this
document
you
call
whatever
it
6
But how does that process where only, you
7
know, maybe one sentence gets picked up that somebody
8
realizes
that
9
document
need
pages
to
be
47
and
changed
56
of
unless
a
4,400-page
you
reproduce
10
every single sheet from the CAFTA output from the new
11
-MR.
12
13
That's
the
difficulty
that we have with the CAFTA thing because -MEMBER
14
15
WACHOWIAK:
software.
STETKAR:
blame
the
This is not a software problem.
MR. WACHOWIAK:
16
Don't
No, no.
What I'm saying
17
is we know how to do this with Word when we change
18
the document.
19
it will show us where on the page we change it.
One of our things that we have to do with
20
21
It keeps those revision controls, and
this is we have to do that piece manually and -MEMBER
22
STETKAR:
Well,
it's
basically
23
your analysts should be making sure that the -- the
24
documentation is updated.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
254
MR. WACHOWIAK:
1
And that's part of our
2
design process that we use for the PRA -- when the
3
model is changed, someone had to change the model.
4
It doesn't change itself, much as we would like it
5
to.
MEMBER STETKAR:
6
7
That's right.
Yes.
MR.
8
9
Sure.
model.
WACHOWIAK:
Someone
changed
the
And then an independent person goes back
10
through and verifies that that change was indeed done
11
correctly.
12
record.
And then that's all signed off in our QA
We have that piece.
And so I have the fault tree picture, the
13
14
old one.
We have that.
The new one, I have that.
15
But if it's one page that's changed in the middle,
16
right now we have to back and manually say it's page
17
45 of -MEMBER STETKAR:
18
That's right except that
19
what we were led to believe -- and we were corrected
20
-- several times in the last meeting that chapter 22
21
had all of the changes that were made, that that was
22
the documentation so that the process is look at 22.
23
If it's not changed in 22, then 4 is 4 in the rev. 3
24
PRA report, the operative 4, 8, or whatever other
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
255
1
2
things that show logic models is indeed still valid.
And that doesn't seem to be entirely consistent.
I
3
am not quite sure how any staff audit would indeed
4
discover that either.
MR. LI:
5
6
Rick, should I say something on
it?
7
MR. WACHOWIAK:
Sure.
Go ahead.
8
MR. LI:
9
When the staff comes in, we have more
This is Jonathan Li from GEH.
10
prosperity.
We have all the activity to show what
11
you can get.
You know, you want printed .pdf version
12
and to tell them what has changed.
13
inhumane
14
readily what's changed because the first you print
15
out will change dramatically, you know.
16
will
17
different.
to
be
the
reviewer
different.
because
The
That's almost
you
cannot
show
The pages
descriptions
will
be
18
But when you come in, we can show them
19
the real fault tree changes between the old one and
20
new one and show them one by one, "This is where we
21
changed."
22
changes one by one, make sure --
if
they
are
MEMBER STETKAR:
23
24
And
really
interested
That's the key.
in
If they
are really interested and they know what to ask about
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
256
1
and are applied that they should go ask about it,
2
that's fine.
MEMBER APOSTOLAKIS:
3
This is a record
4
here.
5
stuff because we explained it to the staff auditor."
6
Isn't there a record that has to stay somewhere for
7
You can't just say, you know, "We have good
future reference?
MR. LI:
8
9
The 22 sections are intended to
describe our changes.
MEMBER APOSTOLAKIS:
10
Yes.
But you are
11
saying that when they come in, it is easy to explain,
12
but when you go to 22, it is hard.
CHAIRMAN CORRADINI:
13
Rick
said?
Because
14
what
15
Twenty-two marked this change in words but did not
16
reflect it by repeating the modified fault tree.
MEMBER STETKAR:
17
he
Can I just repeat
said
it
precisely.
Twenty-two actually did
18
not say anything about the one that I just -- it did
19
not say, "I didn't find a word."
20
in there, but I am not sure that I read every word of
21
chapter 22.
I mean, it might be
22
But because I was using GDCS as one of my
23
spot-check systems, I pretty well thought that I read
24
everything in 22 about like GDCS.
And I don't recall
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
257
1
reading that, "Oh, by the way, in top event EQU, we
2
removed this basic event."
3
that.
I don't think I found
I was kind of surprised.
4
I found some other words in there that
5
we'll talk about after the break that I want to check
6
on that may affect top event INJ also.
7
know whether or how.
So
8
9
10
it's
not
at
all
But I don't
clear
that
--
I
certainly didn't know about that change, even any
words.
CHAIRMAN CORRADINI:
11
Okay.
Let's have a
12
break, 3:20.
13
(Whereupon, the foregoing matter went off the record
14
at 3:00 p.m. and went back on the record
15
at 3:20 p.m.)
CONTINUED PRESENTATION/DISCUSSION
16
CHAIRMAN CORRADINI:
17
18
want to bring something up?
19
MEMBER STETKAR:
20
CHAIRMAN CORRADINI:
21
MEMBER STETKAR:
Mr. Stetkar, did you
I would like to, yes.
Okay.
Good.
I'm probably going to
22
regret this, but could you show me the top logic for
23
top event VI-TOPINJ, the current?
24
today?
That's what it is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
258
1
MR. LI:
2
MEMBER STETKAR:
3
know this picture.
No.
That's fine.
I
That's the good news.
How easy is it for you to chain down?
4
5
Yes.
Easy?
6
MR. LI:
7
MEMBER STETKAR:
8
3 pool tail branch.
9
go down.
Yes.
Okay.
Okay.
Good.
Go down the 3 of
Now go back up and
Move it over and just go down one of the
10
like line A injection failure.
11
did the pool empty, the positive.
12
Stop.
13
there, please.
Thank
you.
Those
14
The
two
at
Okay.
first,
the
That's how you
There you go.
don't
top
do
where
anything
it
says,
15
"Mechanical Failures on Line A, Pool A" and "Pool A
16
Empty Due to Other Causes," are new in rev. 3.
17
that correct?
MR. LI:
18
19
Is
This is the one you have for
this system.
20
MEMBER STETKAR:
21
MR. LI:
Okay.
That probably is true because
22
what we did is we investigate how to gauge costs,
23
tried to add a basis, which could be new.
MR. WACHOWIAK:
24
I'll find it.
I've got a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
259
1
rev. 2 model.
MEMBER STETKAR:
2
3
model
right
here.
4
question.
5
systems analyst.
It
No.
was
I've got the rev. 2
kind
of
a
rhetorical
I was hoping that you were actually the
Indeed, those are new.
6
I am hoping that
7
they are logically correct.
8
that I stumbled over in chapter 22, where there is a
9
sentence that says, "In rev. 3, we have correlated
of
the
injection
This indeed is the thing
10
failures
with
11
pools," which we didn't do before.
failures
of
the
12
I couldn't see this fault tree anywhere
13
because this fault tree is not documented in the rev.
14
3 PRA report anywhere.
15
how you have really done that because I was really
16
concerned about that because my first comment on the
17
first page of the first GDCS fault tree that I looked
18
at back in rev. 2 was that that fault tree did not
19
logically
20
failures that indeed would disable GDCS.
21
it was not complete.
for
all
of
the
MEMBER APOSTOLAKIS:
22
23
account
So I couldn't go check to see
combinations
of
Logically
This fault tree you
have not seen before?
MEMBER
24
STETKAR:
That
statement
is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
260
1
correct.
I have never seen the fault tree that is up
2
on the board.
MEMBER APOSTOLAKIS:
3
4
it from?
MEMBER STETKAR:
5
6
He may have gotten it
from the computer.
7
MEMBER APOSTOLAKIS:
8
CHAIRMAN
9
I mean, is there --
CORRADINI:
Where
is
it
documented?
10
MEMBER STETKAR:
11
MEMBER
12
So where did you get
It's not.
APOSTOLAKIS:
Are
we
keeping
separate documents or what?
MR. LI:
13
No.
Let me explain the process
14
here.
So I think Rick earlier talked to the staff.
15
What we decided is we will update our model.
16
system model for rev. 3 was updated.
17
MEMBER STETKAR:
18
MR.
of
LI:
showing
This change was made?
These
changes
20
thousands
21
feasible for the NRC to do the DCD revision 5 review.
22
So what we did, we do extra.
23
read all 22.4 something, we describe the change in
24
detail.
trees,
we
change
made.
Instead
fault
single
were
19
of
every
So the
don't
to
think
the
it's
So if you
And we develop every single change, say, "Is
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
261
1
that change notable?"
2
CHAIRMAN CORRADINI:
3
MR. LI:
we
did
So after we completed
4
that
5
quantification.
6
in 22.7 and also 22.8 for level 2.
7
check any notable changes.
8
describe it.
have
another
round
of
So the quantification is documented
So, again, we
If it's notable, we would
If it's not notable -MEMBER APOSTOLAKIS:
9
10
process,
Notable.
Is that change what?
"Would describe it"
means you will also show the tree?
11
MEMBER BLEY:
No, not always.
12
MEMBER STETKAR:
Let me ask you something
13
about notable.
Notable, the difficult thing about
14
performing a review is not to look at the things that
15
are important, notable.
16
that are not important, not notable, and understand
17
why they are not notable.
It's to look at the things
18
So to do a review, most of the review
19
process is not to look at the things that are there.
20
It's to look at the things that are not there.
21
So, for example, when I picked up the
22
first page of the GDCS fault tree back in rev. 2, the
23
top event INJ, I looked at that.
24
this as a fault tree logic does not account for all
And I said, gee,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
262
1
of the failures that contribute to the system, to the
2
system failure.
I can't find that by looking at only the
3
4
failures that I can see in cut sets.
I can't find
5
that by looking at the dominant contributors to core
6
damage.
7
that I can look at.
8
tree logic and the system and say the fault tree is
9
not a correct logical representation of the system.
I can't find that by looking at anything
I can only look at the fault
10
It is something that is wrong.
11
complete.
It is logically not
So I made that comment.
12
13
the
fault
tree
is
not
a
14
representation of the system.
15
why that is.
I said, "Gee,
logically
correct
We need to find out
How did that happen?"
16
You have now made a change to that fault
17
tree, which I can't see because you have determined
18
that
19
determined that the original error was not important.
20
If
the
it
change
had
been
is
not
important.
important,
if
it
Well,
had
been
you
the
21
dominant contributor to core damage, you would have
22
fixed it.
It was not the dominant contributor to
23
24
core
damage
because
the
fault
tree
was
logically
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
263
1
wrong because it did not include things.
2
not look at the results.
3
notable contributors and find this because it wasn't
4
there at all.
6
fault
7
error.
9
tree,
12
Its contribution was precisely zero.
you
fixed
that
That's really good.
error,
you
found
the
You found it internally.
You have requantified the model, and the change is
not important.
Okay.
10
11
You could not look at the
Now you're telling me that you fixed the
5
8
You could
That might be correct or it might
not be correct because there might still be an error.
But I can't see the error.
I have to take your word
13
for the fact that numerically this change didn't make
14
any
15
difference to the overall result.
16
a bit uneasy because you have fixed an omission by
17
putting in something that I cannot review because I
18
can't see it.
difference
because
it
19
CHAIRMAN CORRADINI:
20
MEMBER STETKAR:
21
MEMBER BLEY:
22
was
not
a
notable
And that leaves me
I think they get it.
Okay.
Can I add just one short
thing?
23
CHAIRMAN CORRADINI:
24
MR. WACHOWIAK:
You get it, right?
I get it.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
264
1
MEMBER ABDEL-KHALIK:
2
but how is it going to be fixed?
MEMBER BLEY:
3
4
to talk to, Said.
I mean, you got it,
Well, that's what I wanted
Let me say something.
5
MEMBER STETKAR:
6
MEMBER BLEY:
Go ahead.
It's short because I have
7
been struggling with this since I reviewed the stuff
8
you guys wrote.
9
hit me through the discussion today that if you had,
10
in
11
chapter 22, then it would be very possible to say,
12
"Gee, I see something here.
13
and see how it worked out."
14
I'd
15
fact,
And I can't speak for staff, but it
rev.
sure
3
like
completely
and
included
Let me go back and look
that
better
on
the
receiving end, but it's up to you guys what to send.
MEMBER
16
17
updated
APOSTOLAKIS:
So
what
is
the
resolution for this?
MR. WACHOWIAK:
18
I don't know that there
19
is a resolution.
20
analogy for other things that we are doing with the
21
certification.
22
the
23
non-limiting accidents, non-limiting.
24
of the limiting accidents are the ones that we chose
limiting
Let me try to put this into an
In the accident analysis, we submit
accidents.
We
don't
submit
the
And the choice
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
265
1
that we examined as limiting.
2
So it's a very similar case.
3
be an error in our model that we're not showing an
4
accident as limiting because there is an error in our
5
model and the reviewers don't get to see it.
So I am not sure how.
6
There could
You know, that's a
7
similar situation, and it's the way that we have been
8
organized for passing this information on.
9
analysis.
We
determine
what
are
10
things.
11
that contribute to our conclusions.
the
We do the
significant
And we show the staff the significant things
12
So the review that you are looking for --
13
and maybe this gets back to that whole thing about
14
the peer review concept and what they're doing is not
15
equivalent to a peer review because the things that
16
you're
17
specific things that are not sent in for review.
looking
for
specifically
MEMBER STETKAR:
18
are
some
of
the
But this is a little bit
19
different, I think, Rick, because this is the PRA of
20
the design as it exists today.
So it's not the PRA
21
of
equipment
22
thought would probably be important that go through
23
the particular systems that you thought were probably
24
important that would go through the particular event
the
particular
pieces
of
that
you
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
266
1
trees that you thought were probably important, the
2
analogy to your transient analysis stuff.
This is the PRA of the plant design as we
3
4
understand
the
5
therefore, if a piece of equipment is in the plant
6
design, as we understand it, it should be in the PRA
7
model.
8
appropriate
9
that
And
piece
design
the
PRA
and
as
model
and/or
of
it
exists
logic
should
branching
equipment
with
today.
So,
have
logic
to
other
pieces
the
combine
of
10
equipment in the plant.
11
the whole reason that we do the PRA, to look for
12
these
13
important they are to overall risk.
combinations
of
I mean, that's, after all,
failures
and
how
relatively
So I think this is a little bit different
14
15
because
16
particular, are not filtered or I would hope that
17
they're not filtered by your value judgments about
18
what
19
important that reviewers should see it?
20
insignificant that reviewers don't need to look at
21
it?
Because in many cases, the review process looks
22
for
things
23
consistency and and/or type logic.
is
the
PRA
important
like
and
the
that
PRA
logic
reviewers
completeness,
--
looks
models,
what
is
in
so
And what is
for
logical
So I'm not quite sure that that --
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
267
CHAIRMAN CORRADINI:
1
I am listening to
2
you guys discuss this.
And I am not a practitioner
3
of this.
4
that gets you over the concern that Dennis and John
5
have raised.
I am trying to understand a path forward
MEMBER BLEY:
6
7
What
8
correlation and it's all okay but they didn't show
9
you the new correlation?
10
if
they
said
I want to see it.
they
To me it's not the same.
used
a
new
heat
flux
Where's that correlation?
Where did it come from?
Well, if we got a new fault tree model,
11
12
you say, "Yes.
They fixed something.
13
needed to be fixed."
14
is the basis of it?
I saw it.
But how did they fix it?
It
What
What does it look like?
15
To me a simple solution would be the next
16
time around give them the whole thing and tell them
17
what changed.
MEMBER SHACK:
18
I think Dennis' idea, you
19
submit the complete PRA and chapter 22 and you can
20
find -MEMBER BLEY:
21
22
23
You can see what happened.
This one you couldn't go see.
You haven't gone
through it yet to see if you liked what was done.
MEMBER SHACK:
24
You know, it would take a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
268
1
while.
2
to --
The problem is I don't know whether we want
MEMBER
3
4
MEMBER SHACK:
8
MEMBER APOSTOLAKIS:
With everybody here, yes.
you how it differs from rev. 2.
MEMBER STETKAR:
12
MEMBER APOSTOLAKIS:
16
Yes.
Well, my -Right now they show
you rev. 2.
MEMBER SHACK:
14
15
Rev. 3 should have
the updated fault trees and still a chapter to tell
11
13
you're
time -MEMBER BLEY:
10
what
-- go through this in real
7
9
So
saying is --
5
6
APOSTOLAKIS:
work for them.
It's not that much more
They've got rev. 3 on their computer.
It's a .pdf file.
17
MEMBER APOSTOLAKIS:
18
MR. WACHOWIAK:
Yes.
It would have been more
19
work.
Number one, there is a lot of overhead in
20
producing the whole document in our processes, the
21
things that have to be done to make that happen.
22
It's
23
insignificant to produce the whole thing with the
24
change.
an
8,000-page
document.
So
it's
not
So there's that piece.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
269
1
The other piece that we described in 22
2
is that the level 1 and level 2 models' internal
3
events
4
models, the fire and external events and things like
5
that, were evaluated to see if any of these changes
6
that we did to the fault tree models or the event
7
tree models or the data, any of those things, would
8
have impact on those models.
9
updated at the point where we submitted this, but a
10
evaluation was performed on whether there were going
11
to be any changes to those.
were
completely
updated.
But
the
other
And those had not been
12
So one of the reasons that we went this
13
way was so that when DCD rev. 5 went in, that the PRA
14
that reflects, to the best of our ability at that
15
time, reflects DCD rev. 5 could get to the staff at
16
the same time without having to wait six months after
17
the DCD goes in for the PRA and all of the niceties
18
of fixing the document up get reflected because our
19
best estimate, our best of our ability right now is
20
probably about six months from when the DCD gets
21
changed to when the full PRA can be updated.
So
22
this
was
a
mechanism
to
get
those
23
changes in front of the staff and start reviewing
24
them immediately upon getting the difference in rev.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
270
1
5 because if we didn't do that, then they would be
2
there with DCD rev. 5 looking at a design that they
3
don't
4
changed in the PRA because there is nothing there in
5
front of them.
even
10
road
map
of
how
it
should
be
that part of the -MEMBER STETKAR:
8
9
a
So this was our best option from getting
6
7
have
It's okay for the design
changes, but it still doesn't address changes like
this.
CHAIRMAN
11
CORRADINI:
Can
I
just
12
intervene?
13
So, if I understood what you just said, the fact that
14
you issued rev. 5, DCD rev. 5, and the version 3 of
15
the PRA almost simultaneously was facilitated because
16
you didn't choose to, need to, want to enumerate all
17
the details that have been changed and document them.
18
I want to make sure I get this right.
Did I get that right?
MR. WACHOWIAK:
19
Not exactly.
We didn't
20
need to and we didn't necessarily want to change all
21
of the sections in the document to do that.
22
CHAIRMAN CORRADINI:
23
That I get.
So let
me just -MR.
24
WACHOWIAK:
So
we
were
able
to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
271
1
address all of the changes that happened to the plant
2
between the plant design, which was like rev. 3 and a
3
half of the DCD, which is where rev. 2 PRA was.
4
we were able to address all of those changes because
5
the time difference in doing it this way, we were
6
able to catch and address things that changed up to
7
approximately
8
submitted.
a
month
CHAIRMAN
9
CORRADINI:
Again,
you were to have produced a chapter 22 which had
13
essentially what was revision 2 -- I'm sorry -- what
14
were the changes to revision 2, almost like a compare
15
document, that would have been much more onerous and
16
it would have created a time delay, even though the
17
design that you were looking at and the analysis that
18
you
19
That's what I hear you telling me.
MR. WACHOWIAK:
20
21
so
that
my
12
back
just
from
understand how they fit together.
were
of
were
11
done,
matter
they
standpoint,
have
a
when
10
would
it's
before
And
I
The converse is if
with
DCD
rev.
5?
Most of the analysis was
done.
22
CHAIRMAN CORRADINI:
Okay.
23
MR. WACHOWIAK:
The level 1, all of the
24
systems models were done.
Level 2 internal events
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
272
1
were done.
The fire and flood models were not done.
2
What we did was we looked at what the changes were
3
to the results in the level 1 and compared those to
4
the same event tree logic that we had in the external
5
events to see if there were any changes that we made
6
in the level 1 that would have affected those and
7
described what we thought that the differences would
8
be for those models.
9
MEMBER APOSTOLAKIS:
10
CHAIRMAN CORRADINI:
But there is -I'm sorry.
Can I
11
just end up?
12
So
13
facilitated your review such that you are happy with
14
this?
the
So let me turn to the staff and ask.
approach
that
GEH
has
has
that
What is the staff's view on this?
15
MR. CARUSO:
16
CHAIRMAN CORRADINI:
17
taken,
This is Mark Caruso.
Some of the members
are a bit confused.
18
MR. CARUSO:
19
I
think
Thank you.
This is Mark Caruso.
from
our
perspective,
it
was
20
"Okay.
We agree that this could be very
21
much for convenience.
22
there wasn't a lot of stuff being changed, you know,
23
a lot of significant stuff or stuff being changed,
24
that this was probably an acceptable way to go.
I think in the sense where
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
273
1
I guess the other thing I want to say is
2
our sort of criteria or our objectives in looking at
3
this were to see whether or not we thought these
4
changes would change our conclusions.
5
pretty far along.
I mean, we are
And when they presented it to us, it was
6
7
"We haven't really changed much of significance.
8
know, there are no big design changes.
9
big
modeling
10
changes,
11
incorporated.
changes.
small
modeling
There
are
changes.
There are no
some
This
modeling
has
been
We'll describe those to you."
And we said, "Fine."
12
You
And we looked at
13
that as an opportunity to look at what they said and
14
see if we agreed.
15
wasn't enough information to agree or we felt like
16
"Wait a minute.
17
haven't told us enough," we had the opportunity to go
18
back and issue RAIs to get more information.
If we agreed, if we felt there
This could be significant, and they
19
But I think, for example, if this had
20
been the difference between rev. 1 and rev. 2, this
21
would have never flown.
22
process
23
talking about changing from beta models to multiple
24
Greek letter models, from I "don't have this part of
would
not
have
That would have been this
worked.
I
mean,
we
were
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
274
1
the system in there anymore.
2
took that out."
I took this out.
I
That would not have worked.
3
But in this case, I think the changes
4
were such that they were not that significant in
5
terms
6
conclusions.
of
affecting
MEMBER
7
the
overall
APOSTOLAKIS:
But
results
there
and
is
a
8
question here.
9
discussing -- and I don't think it is the right issue
10
-- is whether to replace rev. 2 by rev. 3 or actually
11
have rev. 3 be the complete new PRA and report that.
12
And that would be a lot of work and overhead and all
13
The issue which it appears we will be
that.
But
14
could
at
least
chapter
22
be
15
complete?
16
the sense that it doesn't have some diagrams, some
17
key
18
understand why it can't be.
fault
It appears that even 22 is not complete in
trees,
and
so
on.
And
that
I
don't
19
I mean, why do you have this fault tree
20
on your computer and you can't add it to chapter 22
21
so somebody like Mr. Stetkar can look at it and draw
22
some conclusions?
23
And the next point is that, having not
24
done that, what if we need to go back to the PRA,
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
275
1
say, three years down the line for some issue?
2
MR. CARUSO:
3
MEMBER APOSTOLAKIS:
4
We say, "Well, gee.
No.
I think -Then what do we do?
We have to go back and find Mr.
5
Caruso and Mr. Li and look at that.
6
tell us what is supposed to be there."
7
the way we do business.
That's not
And you said, I think, that the changes
8
9
And they will
were not of great magnitude.
So that would mean that
10
chapter 22, you know, making it complete is not such
11
a big deal, especially since you seem to have all
12
that information.
13
start drawing trees again.
It's not that you're going to
CHAIRMAN CORRADINI:
14
So, just to add to
15
that, I guess my question is, another way of asking
16
this is the fact that -- well, first of all, did you
17
guys realize that 22 was not complete?
18
MR. CARUSO:
19
CHAIRMAN CORRADINI:
20
Yes
That doesn't
give you pause?
21
MR. CARUSO:
22
MEMBER APOSTOLAKIS:
23
MR. CARUSO:
24
Okay.
that.
Well -It's the result.
I think we went on knowing
And I think in some cases, we had asked some
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
276
1
additional questions.
2
CHAIRMAN CORRADINI:
3
questions had been asked for clarification.
MR. CARUSO:
4
Okay.
So additional
We knew it was going to be
5
mostly descriptions of changes and arguments as to
6
why they weren't important.
7
would -- you know, we would have preferred to have
8
all the stuff so that we could look at the stuff, but
9
I think we agreed to take a shot at looking at what
10
they
presented
11
judgment,
12
sleep over, that it was okay.
a
and
seeing
satisfactory
CHAIRMAN
13
And we agreed that we
if
we
could
judgment,
that
CORRADINI:
make
we
From
a
could
a
QA
14
standpoint, to get back to George's question, -- I
15
kind
16
standpoint, how does one reconstruct what was there,
17
then, if it's not there later on?
of
did
a
detour
MR.
18
19
reconstruct.
20
complete
21
complete.
question
CARUSO:
--
They
from
are
a
going
QA
to
I mean, they eventually will have the
rev.
3.
And
the
22
CHAIRMAN CORRADINI:
23
MR. CARUSO:
24
MEMBER BLEY:
record
will
be
made
This is not the end.
It's not the end.
It might be rev. 4 perhaps
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
277
1
but not rev. 3.
MEMBER APOSTOLAKIS:
2
So there will be a
3
rev. 4 with all these fault trees without a chapter
4
22?
5
MR. WACHOWIAK:
6
MEMBER APOSTOLAKIS:
7
And it will be that
PRA?
MR.
8
9
Yes.
WACHOWIAK:
working on right now.
That's
what
we
are
And we have let the staff know
10
when we expect to complete that.
11
are
12
deciding when they are going to come down to GE and
13
look at it.
at
a
stage
where
the
project
MEMBER APOSTOLAKIS:
14
And right now we
managers
are
I just don't see how
15
this Committee can write a letter on the PRA without
16
seeing this.
MEMBER STETKAR:
17
You know, we were led to
18
believe that the rev. 3 PRA was it and it was fully
19
consistent with DCD, rev. 5 and that was this is it.
MEMBER APOSTOLAKIS:
20
21
I don't think we can
write a letter.
MEMBER STETKAR:
22
In some sense, the rev.
23
3 PRA model that's in the computer probably does
24
satisfy in CAFTA -- well, all right -- to a greater
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
278
1
or lesser extent satisfies that.
2
report as it's currently -MR. WACHOWIAK:
3
But the rev. 3 PRA
And this is where I think
4
there is a little bit of a disconnect on this in that
5
the requirements for certification are a description
6
of
7
insights, not the PRA.
the
PRA
MR. OESTERLE:
results
and
This is asking us to
This is Eric Oesterle from
for all of this discussion -MEMBER APOSTOLAKIS:
15
that Rick has a point.
16
different objectives here.
18
the
Just to provide some historical context
14
17
of
the staff.
12
13
description
write a letter.
10
11
a
MEMBER APOSTOLAKIS:
8
9
and
You
are
looking
at
I do think, though,
It seems to me there are two
it
Mr. Caruso mentioned two.
from
the
perspective
of
certifying the design.
19
MR. WACHOWIAK:
Right.
20
MEMBER APOSTOLAKIS:
The staff is looking
21
at it from that perspective when you said, "Are these
22
changes going to change our main conclusions?" and
23
you concluded probably not.
And in the discussion today, I think we
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
279
1
have
2
actually not a good PRA?
3
that has been given to us?
4
step for a good --
5
MEMBER
6
been
focusing
more
on,
is
a
good
PRA,
Does it reflect the design
It is a very necessary
STETKAR:
That
is
a
necessary
step.
7
MEMBER APOSTOLAKIS:
8
MEMBER BLEY:
9
this
A necessary step.
Even a category class 1 in
the left-hand column.
MEMBER APOSTOLAKIS:
10
Exactly.
Right, the category
11
1.
12
disconnect there.
13
your opinion than is necessary for certifying the
14
design?
I mean, are we demanding more in
And that question is to the staff.
MR. OESTERLE:
15
16
So I'm wondering now whether there is a
This is Eric Oesterle from
the staff.
Rick is exactly correct.
17
The regulations
18
in Part 52, Subpart B for design certifications --
19
it's 52.47 and a bunch of numbers, and it talks about
20
a
21
results.
description
of
the
design-specific
MEMBER APOSTOLAKIS:
22
Yes.
PRA
and
its
But once you
23
start questioning the results, presumably they will
24
give you the reasons why the results came out the way
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
280
1
they did.
MR. OESTERLE:
2
I understand.
And what I
3
am sensing is some frustration on the part of the
4
Subcommittee members in that you don't have the same
5
tools available to you to review this information as
6
the staff does.
I
7
mean,
we
PRA
down
to
the
GEH
and
all
of
the
facilities
and
9
supporting
documentation
10
like that.
And perhaps you don't.
11
don't know if there is anything preventing you guys
12
from going down there.
14
the
go
8
13
audit
can
behind
it
and
everything
I'm not sure.
I
MEMBER APOSTOLAKIS:
It will be worse in
CHAIRMAN CORRADINI:
Let me just ask you,
the future.
15
16
Eric.
I think I understand Rick's point and the
17
staff's in agreement with it relative to the level.
18
So without an audit, without a rev. 4, does the PRA
19
satisfy the DCD at this point?
20
sort of view, does it or doesn't it?
As a progress report
21
I guess my feeling is given what the RAIs
22
have been -- there's no other full RAIs, then I get
23
the impression that this is the level of detail that
24
is good enough.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
281
MEMBER APOSTOLAKIS:
1
We don't know.
At
2
least I don't know what it means to use a PRA in the
3
design certification process.
4
mean?
What on Earth does it
5
CHAIRMAN CORRADINI:
Well, a couple --
6
MEMBER APOSTOLAKIS:
Wait.
7
it, and I say, "Okay.
8
mean?"
And I look at
There is a PRA.
What does it
We don't know, I don't think.
MEMBER
9
STETKAR:
We've
heard
10
that.
11
fashion to identify a lot of RTNSS systems.
some
of
We've heard that the PRA has been used in some
MEMBER BLEY:
12
This morning we heard that
13
significant sequences all had melt greater than six
14
hours, which led to the criteria for running the
15
experiments.
16
PRA, it seems to me you've got to go a level deeper
17
than just saying there is one and that the results
18
don't --
Once things like that are hooked to the
MEMBER APOSTOLAKIS:
19
20
that,
21
requirements that says requirement number 4 says that
22
you should use your PRA to prove this or to do that.
I
don't
think
MEMBER BLEY:
23
24
but
Yes, I agree with
No.
there
is
Stop.
a
list
Stop.
of
There
isn't.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
282
MEMBER
1
2
APOSTOLAKIS:
So
it's
kind
of
fuzzy.
MEMBER ABDEL-KHALIK:
3
But how can you
4
review something if you don't have a complete set of
5
documents?
MEMBER APOSTOLAKIS:
6
I agree with that.
7
All I'm saying is maybe the root cause of the problem
8
is that there isn't a clear role of the PRA in the
9
design certification process.
MEMBER SHACK:
10
11
up -MEMBER APOSTOLAKIS:
12
13
Well, if I could just back
Just put PRA, they
say.
MEMBER SHACK:
14
If you are even going to
15
get a summary of the results, George, you would like
16
to have confidence that it is a summary of -MEMBER APOSTOLAKIS:
17
Somehow you got the
18
impression that I am against that.
19
-MEMBER SHACK:
20
21
I am not saying
You are saying you only
have to submit the summary of the results.
MEMBER APOSTOLAKIS:
22
23
Rick says for our purposes.
24
the purposes?
What I am saying is
And I'm asking, what are
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
283
1
MEMBER SHACK:
2
MEMBER APOSTOLAKIS:
3
4
Well, yes.
As I understood --
But, I mean, what does that mean?
MEMBER SHACK:
6
MEMBER APOSTOLAKIS:
Do a PRA.
-- do a PRA.
Yes.
Now what?
MR. CARUSO:
8
9
Part
52 says --
5
7
Certify the design.
This is Mark Caruso with the
staff.
We discussed this.
10
the
Commission
There are a number of
11
objectives
has
let
12
papers about what you should use the PRA for.
13
mean, there are not regulations, but the requirement
14
is have a PRA and do it.
Well, I agree with you.
15
out
in
policy
I
It talks about
16
how they should use it.
17
know, make sure there is a balance between prevention
18
and mitigation.
19
when you look at those objectives and you say to
20
yourself, "Well, you know, I've got to have a PRA of
21
pretty good level to do this stuff.
22
be able to give me answers," I think we struggle a
23
lot with what level of quality do you need for this.
24
Look for reliability.
You
And I think from my perspective,
And it's got to
We don't have a guide for that or a standard.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
284
MEMBER APOSTOLAKIS:
1
Draw the contrast.
2
There is a contrast between, say, risk-informed ISI
3
and what we are doing here.
4
MR. CARUSO:
Right.
5
MEMBER APOSTOLAKIS:
I know what I should
6
be doing for risk-informed ISI.
7
Westinghouse has guidance.
8
them.
9
mechanisms.
The staff has reviewed
So it says you go and look at the potential
So you develop a matrix, and you say
10
there are consequences.
11
to use the PRA.
Here
12
13
EPRI has guidance.
I have guidance.
I know how
I know what the requirements are.
it
says
do
a
PRA.
Submit
the
results.
MEMBER
14
SHACK:
There
15
agreement we want a category 1 PRA.
16
MR. CARUSO:
17
agreement on category 2.
18
MEMBER SHACK:
19
the plant, no matter what.
seems
to
be
Well, maybe you don't have
You still have to reflect
20
MEMBER APOSTOLAKIS:
21
MEMBER APOSTOLAKIS:
No matter what.
So it's clear to
22
you, the, what the role is of PRAs in certification?
23
I mean, let me ask the question because maybe I am
24
missing the point.
It is not clear to me at all.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
285
1
And Part 52 is very vague.
2
MR. OESTERLE:
3
Do it.
I believe it's been left
vague intentionally so.
4
MEMBER APOSTOLAKIS:
5
MR. OESTERLE:
6
I know.
Eric Oesterle from the
staff again.
Because,
7
by
the
again,
PRA.
the
And
design
8
informed
that
is
9
risk-informed nature of this process.
is
to
part
of
be
the
Also, what we
10
understand is that the PRA is commensurate with the
11
level of design completion of the certification.
12
I
13
risk-informed
14
And design certifications still have some detailed
15
engineering to be completed before we can get to the
16
level
17
risk-informed --
ISIs
you
with
you're
MEMBER
18
19
that
mean,
were
comparing
completed
talking
plants
about
APOSTOLAKIS:
I
already.
using
guess
these
I
these
can't
express myself clearly today.
CHAIRMAN CORRADINI:
20
George, can I try an
21
analogy on the level 2, which is what I was looking
22
for this morning?
23
MEMBER APOSTOLAKIS:
24
CHAIRMAN CORRADINI:
Yes.
I was listening to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
286
1
the discussion.
And there are only two things that I
2
was watching for.
3
features didn't work, could I still feel comfortable
4
that nothing happens within 24 hours?
5
features, what is the chance that it would work one
6
out of ten times?
I wanted to understand, if certain
And given the
7
MEMBER APOSTOLAKIS:
I heard --
8
CHAIRMAN CORRADINI:
That is the only two
9
things that I was looking for for the level 2.
MEMBER APOSTOLAKIS:
10
I heard both from
11
Mr. Wachowiak and Mr. Caruso for the purposes of the
12
certification, it was good.
13
that.
14
the purposes of certification, but for the purposes
15
of Stetkar, it is not good.
I don't know what the hell that means.
MEMBER
16
17
For
They are different.
STETKAR:
Let
me
bring
up
What
is
the
something specific here.
18
19
And I don't understand
MEMBER
APOSTOLAKIS:
MEMBER
STETKAR:
difference?
20
I
am
a
specific
21
detail-focused guy.
And since we have got this part
22
of the fault tree up here, it is relevant to give you
23
a little bit of my concern.
24
drawing that you brought up, Mike.
It gets back to a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
287
At the outlet of each GDCS pool, there is
1
2
a manual isolation valve that is normally open.
3
characterized
4
whatever.
5
normally open.
Why
model?
a
maintenance
isolation
valve
It is a manual isolation valve.
6
7
as
are
those
valves
not
in
It's
or
It is
the
PRA
those?
I
That's a question.
8
MR. WACHOWIAK:
9
MEMBER
Right.
STETKAR:
Why
Okay.
are
10
brought it up earlier in June, but we said we would
11
get to it.
CHAIRMAN CORRADINI:
12
13
Rick acknowledged it
and said he is going to have to get back.
MR. WACHOWIAK:
14
Yes.
And we looked at
15
this.
16
remember back to this morning, we said we originally
17
didn't have back in rev. 0 and rev. 1 of the PRA a
18
BiMAC model.
19
that gave us a 10-3 failure rate or better back then.
20
21
There are a couple of things.
First, if we
We said it had to have a reliability
It was a single point thing.
And so there was no
BiMAC model.
Then
22
when
we
were
looking
or
we
were
23
modeling GDCS, we hadn't decided yet how the deluge
24
lines were going to work and those sorts of things.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
288
1
There were some concepts for it, but it wasn't fully
2
formed yet.
One
3
of
our
screening
rules
for
these
4
manual valves was, does it affect any other system?
5
And at the time it didn't.
6
sure where in the line the -- well, I said that --
7
where in the line the BiMAC pipe was going to take
8
off.
Okay?
Also, we weren't
Was it before that manual valve or after
9
10
the manual valve?
I think on some of the cartoon
11
P&IDs that we've had later, they started to show up
12
after the maintenance valve.
13
This is one of those areas where as that
14
detail was being filled in, we hadn't caught yet the
15
fact that our initial screening process, that this
16
locked
17
indicated and alarmed in the control room that didn't
18
affect any other systems now affects other systems.
manual
valve
was
going
That's a four-legged stool, I guess, there.
21
one.
22
see what to do about it.
to
be
We lost
And we needed to go back and look at that and
So you're right.
23
24
that
So we lost one of our legs of our stool.
19
20
open
A design detail got
filled in that we didn't catch in the modeling of the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
289
1
PRA.
So after that came up, a couple of things that
2
we have done.
3
looking at whether or not adding that part to the
4
model is going to make any difference.
We have done a sensitivity now at
Okay?
And I think we have got the results here
5
6
somewhere in one of our files.
7
have to have the valve, all four of the valves, be in
8
the closed position with the indication showing that
9
they were open and a separated stem and disk sort of
10
In essence, you would
thing.
So in the end, it turns out that yes, we
11
12
missed that.
It doesn't affect the model.
It goes
13
in the PRA maintenance process, as described in the
14
ASME standard, as something you fix the next time you
15
update the PRA.
16
The other thing that we look at with that
17
is we are supposed to have two independent systems
18
here:
19
BiMAC line taking off after the maintenance valve.
20
would rather have it take off before the maintenance
21
valve.
mitigation and prevention.
I don't like that
I
22
So the second piece of this is just not
23
probablistically but deterministically, I can use my
24
influence on the design, whenever that might be, to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
290
1
try to get those guys when they fill in the design to
2
put that line before the maintenance valve because
3
the BiMAC is supposed to have its own maintenance
4
valves anyway.
MEMBER STETKAR:
5
You know, Rick, I hear
6
you say all of this.
And you have spent ten minutes
7
justifying in an ad hoc basis why the PRA did not
8
include those valves when it would have taken an
9
analyst all of a minute to put the valve in the PRA
10
model initially and we wouldn't have this discussion.
And the valve has been in the design.
11
12
It's always been there.
13
seen
14
labor-intensive to put those valves in the PRA model.
15
If they were in there, their importance indeed would
has
had
these
Every picture that I have
valves
in
there.
It
is
16
be explicitly quantified.
17
and deluge would, in fact, be explicitly modeled.
not
Their effect on injection
The fact that one valve, in fact, affects
18
19
four
injection
valve
lines,
20
mentioned but one valve being closed affects, well,
21
four injection valves, if that valve is closed, your
22
m out of n injection valve opening criteria changed
23
dramatically, injection valves.
MR. WACHOWIAK:
24
which
you
have
not
If two --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
291
MEMBER STETKAR:
1
2
If the sample is four
injection valves.
MR.
3
WACHOWIAK:
Two
4
There are four maintenance valves.
5
-MEMBER STETKAR:
6
7
MR.
Each one has two
There are three back at
WACHOWIAK:
About
the
GDCS
line
Not
the
GDCS
line
injection.
MEMBER
10
11
injection.
STETKAR:
The GDCS pool isolation valves.
12
MR. WACHOWIAK:
13
MEMBER STETKAR:
14
talking about valve F004A, for example.
MR.
15
16
valves.
the -- are we talking about the same valves?
8
9
injection
WACHOWIAK:
Okay.
Pull up a drawing.
Yes.
So
open
up
I'm
in
chapter 4 the -MEMBER STETKAR:
17
If you have a simplified
18
diagram of GDCS.
And there are only three of those
19
because there are only three pools.
20
MR. WACHOWIAK:
21
MEMBER STETKAR:
One of them, I believe
MR. WACHOWIAK:
The larger pool has two
22
--
23
24
But there are four lines.
injection lines.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
292
MEMBER STETKAR:
1
2
But I thought there was
only a single valve.
MR. HOWE:
3
Each maintenance valve is on a
4
line that then branches into two injection lines.
5
one valve takes up two.
6
front of me.
I don't have the drawing in
MR. WACHOWIAK:
7
So
When we build a PRA, we
8
set up some initial ground rules for what things get
9
put in and what things don't get put in.
10
MEMBER STETKAR:
I guess that's what I'm
MR. WACHOWIAK:
Manual valves that have
11
asking.
12
13
indication in the control room.
Unless they affect
14
more than one system, they're not modeled.
15
MEMBER STETKAR:
And that's --
16
MR. WACHOWIAK:
That's one of our basic
17
assumptions.
MEMBER STETKAR:
18
I really question that
19
assumption because:
20
a nuclear plant know or any kind of facility know
21
that those valves do indeed fall apart, that the
22
indication
23
walked around and closed it, but they do fall apart.
control
And
24
if
a) people who have ever operated
room
tells
there's
you
a
very
nobody
long
randomly
exposure
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
293
1
period to that failure, like, for example, you don't
2
put flow through that line for, oh, several years,
3
the likelihood that that thing is closed when you
4
need it can get rather large.
5
So leaving it out of the PRA regardless
6
-- the whole point is that we're having a discussion
7
about why you did not put something in the PRA that
8
would have taken an analyst one minute to put in the
9
PRA.
This is not a huge philosophical issue.
10
11
It's
not
12
anything.
13
says,
14
takes.
a
labor-intensive
issue.
It
is
not
It is putting a basic event in there that
"Valve
closes
spuriously."
That's
all
it
An analyst knows how to do that that quickly.
In fact, you have the data for it already
15
16
in the PRA model.
So you don't need extra data.
17
It's that level of effort that we're talking about,
18
and it is part of the design.
19
design.
That valve is in the
Now, where is the pool BC valve?
20
21
the A.
22
And
23
affects the A and whatever it is, the A and E.
that
That's
That's the valve that I was talking about.
indeed
only
affects
one
injection.
It
What about the pool B/C valve, though?
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
294
1
Isn't there just a single valve from the discharge of
2
pool B/C and then the discharge breaks?
MR.
3
doesn't just have one, as we said before.
CHAIRMAN
out
of
just
5
6
coming
B/C
equivalents
CORRADINI:
the
has
two
pool.
It
Because
it's
a
bigger pool?
MR. WACHOWIAK:
8
9
this
Pool
4
7
of
HOWE:
My understanding is that
there are a total of four injection lines --
10
MR. HOWE:
Right.
11
MR. WACHOWIAK:
-- that do this.
And one
12
pool just has two of those injection lines coming out
13
the bottom.
MEMBER STETKAR:
14
You know, I thought that
15
the drawing that I saw that had all the pools on it
16
-- does your drawing show all the pools?
17
MR. WACHOWIAK:
18
MEMBER STETKAR:
I don't remember any -Because, you know, I was
19
concerned about it in a bigger picture sense, not the
20
details of level of effort and screening criteria.
21
was concerned about it in a bigger picture sense that
22
here is a single failure that affects two functions,
23
both deluge and injection.
I
And also I was concerned that on the B/C
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
295
1
line, if there is a single isolation valve there, it
2
affects four injection valves and six deluge valves,
3
which is a big deal if that one is -MR. WACHOWIAK:
4
5
would be a much bigger deal.
6
MEMBER STETKAR:
Much bigger deal, and it
And it would be.
It's
7
still a single failure that affects one train of
8
injection and one train of GDCS deluge.
9
MR. WACHOWIAK:
MEMBER STETKAR:
10
Right.
You know, I don't want
11
to make value judgment.
The reason I do a PRA is not
12
to make pre-decisional value judgments about what is
13
going to be important and what is not going to be
14
important.
15
The reason I do a PRA is to model the
16
plant and the plant design and let the PRA tell me
17
what is going to be important, what is not going to
18
be important.
19
So if spurious closure of that valve, if
20
it's in the PRA, if it's in the PRA and it's not
21
important in the PRA results, fine.
22
problem.
23
determined that that is not a big deal.
I've
indeed
modeled
I've solved the
the
plant
and
If it's not in the PRA model at all, I
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
296
1
just don't know.
2
might be.
I have no idea how important that
3
CHAIRMAN CORRADINI:
4
MEMBER APOSTOLAKIS:
George?
I am wondering now
5
where we're going with this discussion.
6
to make a recommendation of some sort?
7
very unusual Subcommittee meeting.
8
review in real time.
CHAIRMAN CORRADINI:
9
Are we going
This is a
We are doing the
The purpose of the
10
Subcommittee meeting was to give ample time to ask
11
questions about level 1.
MEMBER APOSTOLAKIS:
12
Yes.
We have been
13
talking now for about 50 minutes about what is in
14
chapter 22.
15
John probably has no examples of what he just gave
16
us.
17
recommendation regarding what should be in the PRA or
18
are we going to write a letter in March, when we see
19
it?
Where are we going with this?
So where is this going?
I mean,
Is it going to be a
I'm trying to figure out where we are
20
21
going
with
all
of
this.
And
I
still
have
this
22
question in my mind when the applicant and the staff
23
say for our purposes this is good enough and we seem
24
to disagree because I don't know what those purposes
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
297
1
are.
MEMBER
2
was
STETKAR:
everybody
But
3
discussion
agrees
4
satisfy, should be at least --
5
MEMBER APOSTOLAKIS:
6
MEMBER STETKAR:
7
that
of
it
your
should
Category 1.
-- at least category 1,
--
8
MEMBER APOSTOLAKIS:
9
MEMBER STETKAR:
Yes.
-- which says that valve
10
should be in there, shouldn't it?
11
plant design.
That's part of the
MEMBER APOSTOLAKIS:
12
13
part
faithful model of the plant.
MR. WACHOWIAK:
14
Well, should be a
There's no -I think you've got an
15
assumption that open manual valves that don't affect
16
more than one system is typical -MEMBER STETKAR:
17
18
typical
and
typically
19
typical.
20
again, for top --
Bring up your -- it's
applied
but
it's
Bring up your fault tree, if you could,
CHAIRMAN CORRADINI:
21
wrong,
Can we just get back
22
to George's question?
I guess when we left the June
23
3rd meeting, there was uncomfortableness from most of
24
the Committee that was in the room when we had the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
298
1
Subcommittee meeting.
2
MEMBER APOSTOLAKIS:
3
CHAIRMAN CORRADINI:
4
understand
some
specific
5
details of the level 1.
MEMBER APOSTOLAKIS:
7
CHAIRMAN CORRADINI:
time
for
9
piece.
the
level
2
But we wanted to
sequences
6
8
Okay.
and
dig
into
Right.
And we ran out of
severe
accident
management
So we wanted to have the Subcommittee meeting
10
for two purposes.
One was to go over the severe
11
accident management.
12
And secondly is to dig deeper into the
13
road three of the PRA to get a warm and fuzzy feeling
14
that
15
robustness that we should feel good with an interim
16
letter that said they looked like they were on the
17
right path, you know, keep going.
things
were
of
a
level
of
MEMBER APOSTOLAKIS:
18
reliability
Yes.
and
And what came
19
up today is that documentation is not to the level
20
where we can actually draw conclusions.
21
that's what really is happening here.
22
CHAIRMAN CORRADINI:
23
MEMBER APOSTOLAKIS:
24
I think
I guess I -We are having some
logic diagrams that John, for example, who obviously
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
299
1
looked at it in detail, sees for the first time.
2
Right?
MEMBER BLEY:
3
Could I say something on
4
that?
This came up last time.
5
John
6
decided to just look at little bit into the model.
7
And each of us picked a little different part of the
8
model.
and
I
before
we
had
And it began because
that
last
meeting
had
And I went into the fault tree for the IC
9
10
system and kind of found the same thing John did.
11
found four or five things that were just wrong with
12
respect to the system description right there at the
13
top, not spending hours and hours, just looking where
14
you would, top event, the first couple of pages.
15
there they were.
16
that's the only few there are."
But
17
I
And
And I said, "Well, you know, maybe
how
could
we
get
confidence
that
18
that's the only few there are?
19
could think of was what if we pick a sequence and
20
just work all the way through, see the event, look at
21
the data, look at the model, and see if this is an
22
anomaly or if there is a lot of it.
Now, the things that I had found in the
23
24
And the only way we
IC
tree,
it
looks
like
most
of
them
you
guys
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
300
1
identified
2
uncomfortable that if we could look in 15 minutes and
3
see things like that, that the staff hadn't flagged
4
those, those kind of levels in the tree.
5
we were told at the last meeting, well, you know, the
6
trees are too complicated for us to dig very hard
7
into.
fixed.
I
am
also
a
little
And I think
So we were left with this uncomfortable
8
9
and
feeling.
We talked as a group.
And I'm just hit
10
with -- you know, I'll stay with the correlation.
If
11
you found the correlations that were used were wrong,
12
would you be comfortable?
And we had this feeling.
13
So now I've got a lot more confidence in
14
the review process that's going on at GEH, but we're
15
still finding a few things that -MEMBER APOSTOLAKIS:
16
But the real issue
17
that we have been discussing is this documentation
18
because you say, "I've heard them, and now I feel
19
better."
20
MEMBER BLEY:
21
MEMBER
22
APOSTOLAKIS:
Yes.
But
you
see
the
couldn't find it in the written document.
MEMBER
23
24
Well, some better.
BLEY:
And
I
couldn't
trees to check and see how the model was really done.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
301
1
It fits great now.
MEMBER
2
are
we
APOSTOLAKIS:
3
where
going
4
recommend something?
5
two more hours.
with
this?
MEMBER BLEY:
7
MEMBER APOSTOLAKIS:
going?
We're
going
is,
to
Yes.
Is that where we're
To try to formulate a recommendation?
CHAIRMAN CORRADINI:
9
question
Because we can discuss this for
6
8
My
I guess I actually
10
look to you guys to get what you might be suggesting.
11
A recommendation might be that until there is an
12
audit, a letter can't be, an interim letter can't be,
13
issued
14
misunderstand it, the staff has no way of verifying
15
some of this stuff because they don't have it.
we
have
no
way
MEMBER APOSTOLAKIS:
16
17
because
and
unloose
I
An audit or until we
see four or five of the PRA, which would be it.
CHAIRMAN CORRADINI:
18
Well, I don't view
19
our job -- this is just for me because I'm not sure
20
since you want to use the heat transfer.
21
to use these sorts of analogies -- and I'll go back
22
to the level 2.
23
transient response of the BiMAC, I am not going to
24
buy off that the BiMAC is any better than just having
If you want
Until I hear something about the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
302
1
the ABWR approach.
But
2
I
hear
that
they
have
done
an
3
analysis and are giving to the staff analysis as for,
4
essentially, severe accident management as the ABWR
5
approach
6
certification
7
accident management.
8
implications relative to the transient analysis for
9
the BiMAC, that might be fine.
will
probably
criteria
satisfy
for
the
design
essentially
severe
So if I see that and I see some
Similarly here, I am listening to you
10
11
guys.
And you are not feeling comfortable.
But I
12
don't think it's our job to feel comfortable.
13
job is to feel comfortable with the staff's review,
14
not with their level, rev. 3 PRA.
Our
15
So if the staff has an audit, they look
16
at things in detail, and they feel comfortable and
17
they come back to us and say in October or whenever,
18
that we went down and we visited, we checked, we did
19
this sort of stuff, and now we feel comfortable that
20
all of the things in the rev. 3 have satisfied us,
21
that's what we -MEMBER APOSTOLAKIS:
22
Why, then, don't you
23
say also that, instead of you wanting to see the
24
transient analysis, if the staff comes back and tells
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
303
1
you, "We looked at it.
2
it?
3
analysis?
Why
you
want
to
look
CHAIRMAN CORRADINI:
4
5
do
It's fine," then you accept
at
the
transient
I want to talk to
the staff when they look at it.
MEMBER APOSTOLAKIS:
6
Then other people
7
will talk to them when they come back and talk about
8
-CHAIRMAN
9
CORRADINI:
What
I'm
hearing
10
from you guys, just so we're clear, what I'm hearing
11
from you guys, at least that's the impression I get,
12
is if this is -- I don't know what you say -- a
13
category 1 PRA.
14
If you're not comfortable with this at
15
the point that you have the ability to look at it, it
16
doesn't exemplify the characteristics of a level 1
17
PRA.
Then you don't seem comfortable going forward.
MEMBER APOSTOLAKIS:
18
19
And in order to do
that, we have to see.
MEMBER BLEY:
20
So parallel to what you
21
just said, I would like to see the staff come back,
22
say they have looked through the details.
23
also like to see the fault trees that go with the
24
event trees that were provided --
I would
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
304
CHAIRMAN CORRADINI:
1
2
questions.
MEMBER BLEY:
3
4
-- and be able to ask a few
questions.
5
MEMBER APOSTOLAKIS:
6
MEMBER BLEY:
7
Exactly.
Very similar to what you
are saying.
MEMBER APOSTOLAKIS:
8
9
And be able to ask
So we cannot really
reach any conclusions until these things happen.
MR. WACHOWIAK:
10
I think you'll find it
11
very difficult, then, to get through the EPR review
12
because they're not submitting any of this stuff.
MEMBER APOSTOLAKIS:
13
14
I'm telling you the
future is even bleaker.
MEMBER BLEY:
15
16
I
don't
know
17
strikes me.
18
say much more.
what
I haven't seen theirs.
they're
using
it
for,
but
So
it
The rule says you need a PRA and doesn't
But when you start using the PRA to make
19
20
decisions, I think that is great.
I think that is
21
important.
22
of that.
23
being comfortable with what is in the PRA.
24
representation as a safety model.
I think we'll have safer plants because
That puts a little higher level need on
It's a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
305
MEMBER STETKAR:
1
And because when you
2
have a passive design that relies, in large part, on
3
non-active equipment with multiple redundancies that
4
result in extremely low numbers, results, the bar is
5
then raised when you start talking about relative
6
contributors.
7
completeness and consistency and things like that.
8
You know, it would be great.
The bar is then raised in terms of
MEMBER BROWN:
9
How do you test them?
10
MEMBER STETKAR:
11
That's what we're doing now.
MEMBER BROWN:
12
You don't test them.
This is it.
I know.
That's why they
13
look marvelous.
You can't test them.
14
know they're going to operate?
15
more active systems is you can test them.
16
them do something.
MEMBER BLEY:
17
How do you
The nice thing about
You see
The trouble is you've got
18
to test them, you know, 10,000, 100,000 times to get
19
the kind of confidence we need to -CHAIRMAN CORRADINI:
20
So have we satisfied
21
George because I don't think we have yet?
22
make sure.
MEMBER
23
24
APOSTOLAKIS:
And
I want to
my
second
conclusion is since we have reached this question
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
306
1
today, do we need to be here tomorrow?
2
will be more of this.
CHAIRMAN CORRADINI:
3
4
7
8
9
MEMBER APOSTOLAKIS:
I mean, yes.
Okay.
So Stetkar comes up with ten
We had the three that we didn't show you.
No.
Here it
is, this and that.
MEMBER STETKAR:
No, George, it's not all
of that.
12
MEMBER APOSTOLAKIS:
13
MEMBER STETKAR:
14
It's a big question.
more examples, and we always get the same thing.
10
11
Well, let's not go
to that question just yet.
5
6
I mean, that
What is it?
If you just let me get
to a couple of more examples.
15
MEMBER APOSTOLAKIS:
16
MEMBER STETKAR:
Okay.
Because some of these
17
things have come up.
So far I have been surprised
18
because they made changes to the models in areas that
19
I had problems with, but I couldn't see the changes,
20
which is the documentation, the thing you've got.
21
MEMBER APOSTOLAKIS:
22
MEMBER STETKAR:
23
24
Right.
There are still things
that I would like to understand a little bit better.
Okay?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
307
MEMBER APOSTOLAKIS:
1
2
That's fine.
That's
fine.
MEMBER STETKAR:
3
just
And I hope that they're
4
not
5
something that is perhaps documentation-related, if
6
you drop -- do you see where your dotted black line
7
is there?
In
particular,
Drop down to something that says, "GDCS
8
9
documentation-related.
injection line break."
I don't recall.
I don't ever
10
recall.
11
recall seeing that before.
12
wasn't in the rev. 2 model.
13
change since between rev. 2 and rev. 3.
14
would like to know where it is because I --
15
CHAIRMAN CORRADINI:
16
funny.
I just tried in real time to -- I don't ever
At least I
You guys are nodding
MEMBER APOSTOLAKIS:
What is true, that
CHAIRMAN CORRADINI:
That it's not in
rev. 2.
21
MR. HOWE:
22
CHAIRMAN CORRADINI:
23
MEMBER STETKAR:
24
So this is another
it's not --
19
20
So that
Is that true?
17
18
Well, okay.
Yes, that's true.
That is true?
Okay.
Then if that's true,
then if you could go back to that --
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
308
MR. HOWE:
1
2
It is described right here in
22.
MEMBER STETKAR:
3
Okay.
But I didn't see
4
the fault tree.
5
to
6
sometime somebody made a value judgment that spurious
7
closure of a normally open manual valve, everybody
8
knows
9
contributor.
it,
That's documentation.
please,
that
added
because
that's
not
right
here
significant
Everybody knows that.
10
have
in
11
multiplied
12
have explicitly modeled this thing.
by
an
some
and
gate
MEMBER BLEY:
14
MEMBER STETKAR:
terms
of
a
break
of
a
pipe
things.
I
A very unlikely event.
I don't know what the
numbers are, pretty -10-5, 3 minus 5.
16
MEMBER BLEY:
17
MEMBER STETKAR:
18
numerically
And, yet, here I
factor-apportioning
13
15
a
in
But go back
Multiplied by something
else.
19
MEMBER BLEY:
Times .4, .4.
20
MEMBER
21
contributor is what they have.
22
have modeled a pipe break, but I haven't modeled --
23
and pipe breaks don't occur very frequently. I
24
seen valves fall apart.
STETKAR:
So
roughly
1-5
a
But a pipe break.
I
have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
309
1
They occurred countably more frequently
2
than pipe breaks, but somebody made a value judgment
3
not to put the valve in there.
4
judgment to put the pipe break in.
MR.
5
But they made a value
WACHOWIAK:
It
is
not
a
value
6
judgment to put the pipe in.
7
technique such that only initiating events that have
8
medium
9
failure.
liquid
will
get
MEMBER STETKAR:
injection
line
This is just a switch
for your -MR. WACHOWIAK:
12
A switch where our code
13
calculates multiple initiators.
14
that.
MEMBER STETKAR:
15
16
that
So what that does is --
10
11
LOCA
This is a modeling
So that's filter for
This is just a filter
for that initialing -MEMBER BLEY:
17
While you have it up there,
18
though, I saw the same thing.
19
something wrong.
20
for the IC.
21
there.
And maybe I'm reading
I saw the same thing in the tree
And that is you have a pipe rupture
22
But when I go to the data point for the
23
pipe rupture in the data tables, it is a different
24
number, not a lot different but different, same thing
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
310
This is a 3.22-5.
1
here.
2
data table, it says 7.5-5.
Why isn't the data in the data table the
3
4
And when I run over to the
same as the data in this tree?
5
MR. WACHOWIAK:
6
MEMBER
7
BLEY:
It
MR. HOWE:
9
MEMBER BLEY:
that
it's
I missed that.
Initiating events changed a
little bit.
MEMBER BLEY:
12
13
long as you documented it.
14
--
I missed that.
Okay.
As
But I didn't see it as I
MEMBER APOSTOLAKIS:
15
I have found they
are the same type, though, where the table has -MEMBER BLEY:
17
18
them,
19
changed the data.
but
I
didn't
Yes, I found several of
notice
that
20
MEMBER APOSTOLAKIS:
21
MEMBER BLEY:
22
us
Yes.
MR. HOWE:
10
16
tells
different data?
8
11
Is that in section 2.22?
you
told
us
you
Right.
So I missed that.
So it's
that documentation thing that's pretty tough.
23
You are frowning.
24
MEMBER STETKAR:
Yes.
I am still -- the
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
311
1
valves bother me because the problem is that if it is
2
value that is spurious with a pump, that is not a
3
problem.
4
identify the fact that spurious closures of manual
5
valves whose status was not verified by actual flow
6
tests can be visible contributors.
There have been many and early PRAs done to
MEMBER BLEY:
7
In fact, there's a data
8
point, for example, in certain kind of PWR back 20
9
years ago after WASH-1400 was done 30 years ago that
10
made them take the internals out of some valves like
11
that, "they" being NRR, to make them a piece of pipe,
12
instead of a valve, because they thought it was one
13
valve that could take out a system, so same kind of
14
thing.
MR. WACHOWIAK:
15
16
can take out a system here.
17
take out --
18
MEMBER BLEY:
19
MR. WACHOWIAK:
It's not one valve that
It is one valve that can
That's true, yes.
So one of four trains
20
that are needed.
21
that blanket assumption and there are some reasons
22
why that you don't think that that is appropriate.
A couple of things, though, with that.
23
24
So I understand that you don't like
I
mean, we have since you asked that question, since
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
312
1
the time that you additionally asked that question,
2
we did look at what would happen to the results if we
3
did change this assumption and explicitly model that
4
valve using the data that was suggested in June that
5
we
6
failed-to-remain-open of the manual valve.
7
we suspected, it didn't make much difference to the
8
answer.
already
had
in
the
document
9
MEMBER STETKAR:
10
MR. WACHOWIAK:
11
MR.
HOWE:
to
model
that
And, as
Much difference.
What was the result?
The
results
I
have
here,
12
actually, I have quantified with a large failure rate
13
just so I could always back it down and to do it
14
over.
15
But, even using a very high number, which
16
per valve was like 5E-2, it still showed a reasonable
17
impact.
18
there for spurious closure of a manual valve, each
19
valve would be on the order of about E-6.
20
take a look at that phenomena.
And I think if we used our data that's in
CHAIRMAN CORRADINI:
21
So we did
Can I just make sure
22
I understood what you just said because this is not
23
my area again.
24
something very low, but you upped the value up to
So you're saying the deluge suggests
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
313
1
5E-2.
And at that point, it started being noticeable.
2
MR. HOWE:
Right.
3
CHAIRMAN CORRADINI:
Okay.
4
MEMBER APOSTOLAKIS:
But not notable.
5
CHAIRMAN CORRADINI:
Not notable?
6
MR. WACHOWIAK:
In PRA, we continue to
7
have to come up with new words for things because
8
people keep codifying the words that we used in the
9
past.
10
11
12
ASME standard ties significant to a specific number.
So we can't say "significant" anymore.
We're
going
to
use
"notable"
until
somebody defines that.
15
(Laughter.)
16
MR. WACHOWIAK:
17
We've got to
use something else.
13
14
We can't say "significant" anymore because the
You can't use the English
language anymore.
CHAIRMAN CORRADINI:
18
I guess, John, do
19
you need more questions now or can we at least plot a
20
path forward for tomorrow?
MEMBER STETKAR:
21
22
things.
CHAIRMAN CORRADINI:
23
24
I've got a couple of
Because George has
that second question that I refuse to answer, but I
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
314
1
need to answer that to plan tomorrow.
MR. WACHOWIAK:
2
Two other things that I
3
want to bring up on this that should get up getting
4
reflected into the rev. 4 description.
5
plan for testing that check valve is to use flow
6
through
7
valve, each outage, it is expected that we will get
8
flow through that valve.
that
pipe.
So
when
they
check
9
MEMBER STETKAR:
10
because in the documentation we have --
11
MR. WACHOWIAK:
12
MEMBER STETKAR:
13
MR. WACHOWIAK:
test
That's really important
It doesn't say that.
-- the flow through that
That is correct.
That is
what it says there.
MEMBER
16
17
the
valve would be tested once every ten years.
14
15
Currently the
unavailability.
STETKAR:
That
is
a
huge
It dominates that line, each line.
MR. WACHOWIAK:
And now because of some
20
MEMBER STETKAR:
Individually.
21
MR. WACHOWIAK:
-- issues with how do we
18
19
other --
22
test the check valve, the process for testing the
23
check valve is going to rely on flow through that
24
valve.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
315
MEMBER STETKAR:
1
But the -- well, I've
2
said enough.
3
bring up, then, before you get to the bigger-picture
4
stuff is that the current models for GDCS and, in
5
fact, I think most of the systems do not include any
6
contributions from maintenance unavailability.
7
is that?
8
you
9
service indefinitely, there is no limit on me, Tom.
to
The only other thing that I want to
Why
Especially because your tech specs allow
have
one
I
10
of
complete
can
safety
operate
safety
this
division
plant
operating,
out
with
three
11
divisions
12
continuously.
13
for that possibility.
14
are always 100 percent perfectly available to operate
15
except for hardware failures.
inoperable
The PRA model does not account at all
Those four safety divisions
MR. WACHOWIAK:
16
one
of
And you are correct that
17
that is the way it is modeled, that we don't have
18
maintenance unavailability for many of the systems.
One
19
of
the
things
that
you
have
to
20
recognize for the passive systems, we don't wire it
21
up the way that you were saying.
22
1 electrical system that controls division 1 valves.
23
All four electrical divisions control all eight of
24
the mechanical valves.
There is a division
So there really isn't a div.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
316
1
1 valve.
2
So what we have looked at in our detailed
3
I&C model is what happens if we assume one of the
4
digital I&C trains, if you will, or divisions is out
5
of service?
6
answer is it doesn't make any difference because the
7
common
8
failure
9
systems.
Does that make any difference?
cause
of
failure
the
systems
of
software
by,
MEMBER STETKAR:
10
Tom,
And the
dominates
the
digital
Let me back off.
the
I&C
In the
11
tech specs, it allows me to have three or more GDCS
12
injection lines.
13
MR. WACHOWIAK:
14
MEMBER STETKAR:
15
MR. WACHOWIAK:
16
Okay.
This is not divisions.
Okay.
I thought you were
talking about -MEMBER STETKAR:
17
Well, I was trying to
18
generalize it to the fact that no maintenance is
19
modeled, but I will be specific.
20
injection lines can be inoperable for 12 hours.
21
can be inoperable for 14 days.
22
MR. WACHOWIAK:
23
MEMBER
24
Two
Right.
STETKAR:
inoperable indefinitely.
Three or more GDCS
And
one
can
be
Those are the way the tech
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
317
1
specs are written.
2
lines.
3
into them.
5
I don't care how many signals or things come
MR.
4
And those are GDCS injection
WACHOWIAK:
I
know
what
you
are
talking about.
MEMBER STETKAR:
6
Well, I was talking in a
7
broader context, but if we want to get specific, why
8
are those allowed conditions not included in the PRA
9
model?
MR. WACHOWIAK:
10
Because we wouldn't be
11
doing maintenance on GDCS valves.
12
tech spec allows that, what are we going to do?
MEMBER STETKAR:
13
14
Even though the
Why is it in the tech
spec?
15
MR. WACHOWIAK:
16
MEMBER STETKAR:
I don't know.
We aren't going to be
17
doing maintenance on the valves because they are in
18
the drywell, but I could certainly disable power to
19
them if I am going to do stuff.
20
signals to them.
21
plant.
You know, I could do things in the
MR. WACHOWIAK:
22
I could disable
You could do that, but,
23
as I said, actually, we have four divisions of power
24
going to every one of those valves.
And you're only
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
318
1
allowed to remove one of those divisions of power
2
from anything in the plant anyway.
You
3
can't
get
to
a
state
in
the
I&C
4
system that allows you to take four power sources
5
away from that valve.
MEMBER STETKAR:
6
7
division out.
8
out.
You can have all of one power division
MR. WACHOWIAK:
9
You can have all of one
Right.
But the valve
10
gets power from four different divisions.
11
MEMBER STETKAR:
12
That's right, but the
changes -MR. WACHOWIAK:
13
And we looked into that
14
with taking one of the I&C systems out indefinitely
15
and looking to see if it affected our model.
16
turns out that it did not.
And it
17
So the question would be, do you model it
18
with them all in service or do you model them with it
19
all out of service?
20
-- or is one of them always out of service?
21
answer turned out to be that we didn't get any change
22
from doing that.
So we looked into it.
23
24
In the end, the answer doesn't
And the
Should we have put
all of those maintenance terms on there?
That is a
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
319
1
good question because typically in a PRA, you don't
2
use
3
maintenance.
4
And you look at how long something is out of service.
the
tech
spec
outage
time
for
your
You take actual maintenance records.
And
5
allowed
when
we
looked
at
what
kind
of
6
maintenance you would routinely do in the plant, not
7
where you would have, you know, someone using an
8
event but what kind of actual maintenance the plant
9
would
10
typically
do,
they
don't
see
them
doing
anything with those valves, calculations.
MEMBER ABDEL-KHALIK:
11
But if tech specs
12
allow you to operate with one system out of service
13
indefinitely,
14
problem that results in a leakage from one of the
15
tanks.
16
that tank until the next outage to find out what is
17
going on and fix it?
say,
you
know,
you
have
some
Wouldn't a prudent operator sort of drain
MEMBER SHACK:
18
19
let's
I'd keep the water there.
It depends on the point of prudence you're -MR. WACHOWIAK:
20
the
GDCS
pools,
that
If there is a leak in one
21
of
would
be
a
GDCS
pool
22
unavailable.
And they are probably only a 12-hour
23
LCO on that.
And it would show up as unidentified
24
leakage anyway.
And they would probably be shutting
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
320
1
the plant down as a LOCA precursor.
What we are trying to do here -- and this
2
3
is
hard
in
the
design
PRA
phase,
when
we
don't
4
actually have a lot of this stuff -- we are trying to
5
extract from design documents what would go into a
6
value in the PRA that is normally generated from
7
historical data.
8
on this.
And so we have to make the judgment
9
Are we going to say that everything is
10
going to be maintained, like it says in the tech
11
specs, or are we going to look at each specific piece
12
of
13
non-outage maintenance to go on with this valve?"
equipment
and
Now,
14
say,
it
I
"Okay.
doesn't
agree
that
Do
we
cover
it
expect
any
corrective
15
maintenance.
doesn't
cover
16
corrective maintenance.
17
think we're going to expect to see a huge fraction of
18
corrective maintenance on these.
19
data for the valve failure itself that is somewhat
20
based on the interval time where it could have failed
21
since the last outage.
But, once again, I don't
And we do have the
22
So I think we kind of sort of pick up the
23
corrective maintenance piece in that because we have
24
got it in the valve data.
They are not going to
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
321
1
close that maintenance valve because the maintenance
2
valve is going to shut off the other GDCS injection
3
line.
4
LCO that is going to have to shut them down.
Once they shut that, they're in the shorter
MEMBER STETKAR:
5
It's a shorter LCO, but
6
still it might be the most important contributor.
7
they only do it once every ten years and have it out
8
for a couple of weeks, it might be the most.
See,
9
it
the
is
And
can't
don't
know
how
understand
how
important
11
important it might be because I don't have the volume
12
control
13
adjust it.
built
my
sound
system
to
even
It's just not in there.
You
14
into
I
I
10
knob
is.
problem
If
say
studies
that
and
you
16
which you say for everything.
17
in the 10-8 world, it is really easy to get factors of
18
2, 3, 5, or 6.
20
MEMBER
not
some
sensitivity
MR. WACHOWIAK:
it's
done
15
19
that
have
important,
And my problem is that
Yes.
STETKAR:
That's
really
easy.
21
Now, is that important relative to a 10-4?
22
not.
23
the plant and the risk assessment of the plant design
24
and the way we expect it to operate and identifying
No, it's
Understanding, however, the vulnerabilities of
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
322
1
those vulnerabilities is not an absolute.
2
relative thing.
It's a
3
In other words, if 90 percent of my risk
4
comes from the fact that I allow a single division of
5
safety-related
of
service
6
indefinitely, I would like to know that.
I would
7
like to know that even if it was 90 percent of the
8
10-6
9
criteria.
number.
power
I
still
to
need
be
all
out
of
my
acceptance
I still need all of that stuff.
But that
10
tells me something about the plant design and its
11
interaction with the tech specs that I can't learn
12
from this.
MR. WACHOWIAK:
13
I fully understand what
14
you are saying, but I think the main issue that I
15
have is that you're looking at what the plant needs
16
to do with its operational PRA and how they use the
17
maintenance rule and how they do their performance
18
monitoring.
They will need to know that when they do
19
20
that.
That's
21
requires them to update the PRA and keep it for doing
22
that sort of thing.
of
the
reasons
that
Part
50
So if the question is, do we have 90
23
24
one
percent
of
CDF
associated
with
this
so
10-8
now
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
323
1
becomes 2 times 10-8, is that going to say that we
2
shouldn't certify this plant?
3
no.
4
should require the operators to model those things so
5
that when they put their plant into a maintenance
6
configuration
7
detect those kinds of changes.
We should still certify the plant.
that
includes
that,
that
And we
they
can
And so this is why it's really hard with
8
9
The answer to that is
this,
because
everybody
is
used
to
dealing
with
10
operating plant PRAs.
11
"Where
12
We're trying to say, "Is this plant safe enough to
13
operate in the U.S. given the rest of the body of
14
regulations that we have to follow?"
is
all
the
And we're not trying to say,
risk
from
15
MEMBER STETKAR:
16
CHAIRMAN CORRADINI:
17
plant
coming?"
Let me ask -I'm sorry.
I need
to go to George's question.
18
MEMBER APOSTOLAKIS:
19
CHAIRMAN CORRADINI:
20
this
the moment?
Go on.
Are you happy for
Okay.
21
MEMBER STETKAR:
Sure.
22
CHAIRMAN CORRADINI:
Okay.
So I guess,
23
with that, I want to get back to your question.
So
24
what are we looking for tomorrow from these gentlemen
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
324
1
and staff to discuss?
2
And I guess I would like to turn first to
3
Dennis and see what you are looking to ask them or
4
wanting to -- because this is an open discussion
5
tomorrow.
6
looking to us.
They
have
MEMBER
7
nothing
BLEY:
prepared.
Let
8
wanted the open discussion.
9
point right now.
me
tell
They're
you
why
I
I guess I am at the
I know there were a lot of errors
10
in the rev. 2 fault trees everywhere we have looked.
11
I wanted to go through and track and see the data
12
and see the fault trees and see how they worked.
13
we don't have them in front of us.
14
real hard up here.
I
15
think
a
process
But
It's going to be
that
waits
for
the
16
audit and allows us to see the fault trees and look
17
for the new ones and see if we find problems like we
18
did
19
rummaging through that tomorrow.
in
the
others
would
be
more
effective
than
I had one last comment I had wanted to
20
21
make.
And that is when I read through chapter 22, I
22
see an awful lot of "This is conservative," "That is
23
conservatively
24
assumed."
assumed,"
"This
is
conservatively
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
325
1
And I remember coming here and talking
2
when Hal Lewis was here, who would just, you know,
3
"Why is it conservative?
4
5
Is it conservative always?"
There is no discussion of why it is conservative.
And is it conservative under all conditions.
6
That leaves me a little empty as well.
7
So I think that goes in with the other kinds of
8
things.
9
this
There is not enough information to know if
is
really
conservative,
if
there
are
some
10
branches in the event trees for which this assumption
11
of conservatism is backwards, you know.
And I think you need some statements of
12
13
that sort.
And I hope after the audit, staff can
14
probably tell us if these are really conservative.
15
CHAIRMAN CORRADINI:
16
what you said so I get it right.
17
MEMBER BLEY:
18
CHAIRMAN CORRADINI:
So let me sharpen
Yes.
Are you saying that
19
-- you said "audit," but I am going to change the
20
word just so I get it.
21
no need necessarily to talk more in detail about
22
specific sequences tomorrow and to ask more specific
23
questions?
Rather,
24
you
Are you saying that you have
would
want
to
wait
until
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
326
1
staff has looked at rev. 3, either by traveling down
2
there or by getting more information on a chapter 22
3
revision or by some mechanism and have them come to
4
us so we can ask penetrating questions of staff now
5
or them, GEH, rather than stay tomorrow and do more
6
of this.
MEMBER BLEY:
7
With one more addition,
8
that we also get the new fault trees so we can look
9
at them.
10
CHAIRMAN CORRADINI:
11
MEMBER BLEY:
Okay.
I think if we go through it
12
tomorrow, it will be really hard to do it up on the
13
board.
MEMBER APOSTOLAKIS:
14
15
It will be the same
as now.
16
CHAIRMAN CORRADINI:
17
DR.
18
accidents, rather than PRA.
KRESS:
My
area
19
CHAIRMAN CORRADINI:
20
DR. KRESS:
21
questions
22
being.
I'm
going
Tom?
was
the
severe
I know that.
And I basically asked all the
to
ask
on
23
CHAIRMAN CORRADINI:
24
MEMBER SHACK:
this
Okay.
for
the
time
Bill?
Well, I am here to learn
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
327
1
and listen about the level 1 PRA.
2
John still has questions, you know, I think it might
3
be worthwhile staying around.
4
lot today.
5
conclusions, but I learned a lot.
CHAIRMAN CORRADINI:
7
MEMBER ABDEL-KHALIK:
Dennis said.
Said?
I agree with what
This is not really very productive.
CHAIRMAN CORRADINI:
9
10
I think I learned a
I'm not sure I've changed any of my
6
8
So, you know, if
I am skipping you.
Charlie?
MEMBER BROWN:
11
I only had one observation
12
out of this.
They made statements about where they
13
got their failure and other results data.
14
your June status meeting for both component failures
15
and human probability stuff.
16
here, I haven't opened my mouth.
It was in
And since I wasn't
17
They didn't say why that data was valid
18
for unique components that are in this plant that may
19
not be in others.
20
or something like that in this plant that we -- do we
21
have those in others?
I guess we've got explosive valves
I don't know.
22
MEMBER BLEY:
23
MEMBER
24
BROWN:
Yes.
So
I
had
some
--
I'm
always suspicious of component failure data because
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
328
1
my experience is you can report it and it's not very
2
accurate.
3
program and found that about a fourth to less than a
4
fourth of the data was valid.
5
spurious reporting.
I did it for 22 years in the nuclear
And you got so many
I mean, you take five things.
6
It turned
7
out it was the fifth one.
8
back and put those original of the first four in to
9
see if it still works.
10
11
Well, but you never go
But we had to report it all.
They got all five failure reports.
And it just
skewed the data.
However, after 10 or 15 years, the skew
12
13
is pretty consistent.
So you could kind of evaluate
14
a float as to where it is kind of okay.
15
My four-star admiral didn't really like
16
that explanation too much but managed to sell him all
17
three of them on it over that period.
18
only point.
MR. WACHOWIAK:
19
That was my
And I think we agree with
20
what you are saying, that the data is what it is.
21
were using the data that was provided by the utility
22
requirements document.
23
MEMBER BROWN:
24
MR. WACHOWIAK:
We
That's fine.
And it compares across
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
329
1
the different plants.
2
we use our PRA in the design is that we're trying to
3
make it so that any individual component failure rate
4
isn't the key to why the core damage frequency is the
5
way it is.
6
existing
7
accomplished that, not in all cases.
9
We want it to be less sensitive than
plants.
There
8
However, one of the ways that
And
are
in
still
many
cases,
some
we
components
have
where
there are a few things that lead you -- luckily,
10
they're -- not luckily.
By design, they're things
11
that tend to be needed late after 24 hours.
MEMBER BROWN:
12
I've gotten something out
13
of this, but I agree with Said and Dennis.
14
don't
15
overwhelmingly
16
opinion.
think
grinding
through
productive.
these
right
That's
my
CHAIRMAN CORRADINI:
17
I just
now
is
personal
So I am going to
18
look to John and George to end this off because I
19
have some concluding things I want to get clear from
20
the staff.
George and John?
MEMBER APOSTOLAKIS:
21
Well, first of all,
22
I have a comment on common cause failure model, which
23
is not a matter of documentation.
24
ask it.
Okay?
So I can
In ten minutes, we will have it resolved.
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
330
1
The way I see tomorrow is the following.
2
If John says that he has other issues that will not
3
lead us again to a question of documentation, then I
4
see us coming in the morning and adjourning at noon.
5
If, on the other hand, John says, "Well,
6
you know, more or less most of them will end up like
7
this," GEH will say, "We have this new figure.
8
hasn't seen it" and all of that, then I don't think
9
we should come at all.
He
10
And we should wait, as Dennis said, for
11
the audit and the new documentation to be given to
12
us.
So it's up to you.
MEMBER STETKAR:
13
Well, I think, George,
14
that, unfortunately, the way we started -- and I'm
15
open.
16
and as we got into some of the detail that's making
17
your eyes glaze over, unfortunately, some of that is
18
indeed related to documentation.
19
have said enough about that.
You know, the way we started this afternoon
And, you know, we
20
The thing that I just mentioned regarding
21
the non-modeling of any maintenance contribution is
22
not related to documentation.
23
there.
It's an active decision.
There
24
That's simply not in
are
other
parts
of
the
model
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
331
1
dealing
with
2
that I have a lot of detailed comments on that, as
3
best
4
documentation.
5
are relatively subtle.
6
time?
as
I
physical
can
and
tell,
functional
are
also
dependencies
not
related
They are relatively detailed.
to
They
Is it worth spending people's
I think that's the judgment of the Committee.
7
Part of my concern, quite honestly, is
8
that we have been talking about now relying on the
9
staff's audit of the rev. 3 PRA, whatever that audit
10
means, as a way of resolving all of these concerns if
11
they're going to do an audit and the result of that
12
audit will be a determination that, indeed, the PRA
13
is either acceptable or it needs yet more changes.
14
MEMBER APOSTOLAKIS:
15
MEMBER STETKAR:
are
we
only
We will see also.
I know, but the question
16
is,
17
Because part of my agenda for this discussion is to
18
try to make the accumulated wisdom here sensitive to
19
some of these issues.
Now,
20
postponing
if
the
the
staff
same
discussion?
determines
that,
21
indeed, it is okay that maintenance is not modeled in
22
the
23
determination.
24
recognize
PRA
at
this
that
stage,
that
should
be
an
active
There should be something saying, "We
maintenance
is
not
modeled.
And
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
332
1
that's okay."
To
2
this
point,
there
has
been
no
3
statement of anything regarding that.
4
these kind of dependencies are not modeled.
5
aware of it, and it is okay that they are not modeled
6
at this stage.
MEMBER BLEY:
7
8
We are
Would they not be -- it's
kind of a question.
MEMBER STETKAR:
9
MEMBER BLEY:
10
11
It's okay that
To whom?
It is going through John,
but it's really aimed at the staff.
Would those not be things -- I'll say it
12
13
the other way.
14
fact, they are okay to be things that have DAC items
15
associated with them.
16
get the issues on the table.
17
MEMBER STETKAR:
18
I would expect those things if, in
And I think it is important to
But there aren't DACs
and ITAACs on the PRA.
MEMBER BLEY:
Is that true?
21
MR. OESTERLE:
True.
22
MEMBER STETKAR:
19
20
23
There won't
be?
It's true.
The PRA is
done.
MR.
24
OESTERLE:
There
are
DAC
items
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
333
1
associated with certain elements I think that Rick
2
talked about that factor into the PRA but associated
3
with the actual physical design.
MEMBER STETKAR:
4
5
But there will be no DAC
items saying when the COL comes forward -CHAIRMAN CORRADINI:
6
Before we spin out
7
of control, so your summary is?
8
hold a tight leash.
9
some things that are probably not documentation but
10
are specific that you would like to go over tomorrow
11
to sensitize the rest of the -- at least I'll call
12
myself an educated member into the joys of all of
13
this.
going to be a level 1 analyst, never in my life.
MEMBER STETKAR:
16
17
Your summary is that you do have
I think you've convinced me I'm never
14
15
I'm going to now
That's not a bad thing,
by the way.
CHAIRMAN CORRADINI:
18
of
this,
19
all
20
documentation?
21
and this one I guess I would like to ask the staff to
22
think about to have an answer tomorrow because your
23
silence
24
agreement with what Rick has said, which is it's kind
leads
right,
To the subtleties of
that
aren't
That's point one.
me
to
believe
necessarily
Point two is --
that
you
were
in
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
334
1
of good enough.
That is, you're not dealing with
2
maintenance.
3
mistakes.
4
significance, whatever the right word is.
5
notable.
That's okay.
That's okay.
Thank you.
There may be some small
They are one level of no
Thank you.
It's not
Thank you.
6
I guess what I'm asking is if we do these
7
detailed things, I guess I am going to turn to the
8
staff sometime tomorrow morning and ask.
9
like some sort of discussion as to since in 19.1.1
10
and 19.1.2 of the DCD, it gives the laundry list of
11
things the PRA is to be used for to trundle down that
12
list and say, at this point in kind of a progress
13
status, are you happy with all these things?
I would
Does the PRA as you see it at this moment
14
15
satisfy
these
things?
16
extemporaneously
17
things do you need to feel good about so that we know
18
where you are going to be?
or
if
it
And
doesn't,
if
what
it
does
sorts
of
19
Because when you said "audit" and you
20
said, "audit," I think the generalized review, it
21
could be a review that they're going to come in.
22
could be a visit.
23
questions and get answers.
It
It could be they just ask more
I would like to ask the staff to be ready
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
335
1
tomorrow for that so that we can get a feeling of
2
what is good enough, which goes back to George's
3
first question.
MEMBER APOSTOLAKIS:
4
5
So now the
last question to John, how much time do you need?
MEMBER
6
STETKAR:
7
whenever we want to.
8
we can finish by noon.
We
can
cut
that
off
If you want to finish by noon,
CHAIRMAN CORRADINI:
9
10
Okay.
We will be finished
by noon.
11
MEMBER STETKAR:
Finished by noon.
12
MEMBER APOSTOLAKIS:
13
MR. WACHOWIAK:
That is good enough.
I would like to jump in
14
on this as well.
15
this with my staff yet, but I think it would be a
16
good
17
because
18
December somebody decide that we need to do a rev. 5
19
of the PRA.
idea
to
what
I think that -- I haven't discussed
go
I
through
don't
want
those
to
20
MEMBER APOSTOLAKIS:
21
MR. WACHOWIAK:
details
have
tomorrow
happen
is
in
That's a good point.
I would like to make sure
22
that if there's anything that needs to be addressed,
23
that we get it -CHAIRMAN CORRADINI:
24
That's fine.
So
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
336
1
let's just push that point to make sure I -- so
2
Dennis had a certain way of doing it, and John has.
3
But they have additional specific questions.
The
4
only
thing
that
I
have
as
an
5
uneducated listener here, we did have in front of us
6
-- now I have lost it -- four sequences from you and
7
from
8
benefit
9
walking through it in a forward fashion tomorrow and
Dennis
us
and
any
from
more
John
by
and
George.
taking
one
10
draw out further questions and details?
11
the feedwater, the FDW-0050 or -MEMBER
12
13
APOSTOLAKIS:
Would
of
these
it
and
For example,
I
think
you're
getting into too much detail.
14
CHAIRMAN CORRADINI:
15
MEMBER APOSTOLAKIS:
I know, but -The important thing
16
in my mind is for the applicant and the staff to know
17
by the end of the meeting what the concerns of the
18
members are.
19
sequence or to start saying, like he did today -CHAIRMAN
20
21
Now, whether John wants to go through a
CORRADINI:
What
about
this?
What about that?
MEMBER
22
APOSTOLAKIS:
--
and
23
that, leave it up to him.
24
important thing because I agree with Rick.
this
and
But that is the most
We have
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
337
1
to start reaching closure.
So he has to know, he and
2
his team have to know, what issues the members raise.
3
So next time they come back, they will have answers.
4
They will have done something.
Right?
And I think the staff is in the same
5
6
position.
7
some members have some issues in their minds that
8
were not aired.
morning.
MEMBER
Give us everything you've
APOSTOLAKIS:
No,
no,
no.
We
never give you everything we've got.
MR. WACHOWIAK:
15
16
Let's come in the morning.
got.
13
14
Okay.
MR. WACHOWIAK:
11
12
That should be the goal.
So it seems we all agree we need the
9
10
We cannot finish this meeting and have
I know.
Then we can
never answer everything.
CHAIRMAN CORRADINI:
17
I'm going to put you
18
guys a bit on the spot tomorrow, but I guess if you
19
could give that some thought because I guess we need
20
to
21
wrap-up relative to chapters 19 and 21 so we know is
22
an interim letter in October reasonable given what
23
you're planning to do, is it unreasonable?
24
not, what needs to be done to get it wrapped up
get
feedback
from
you
as
to
how
you
see
the
And if
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
338
1
because I know I agreed with Amy ahead of time that
2
we would talk about this and plan it.
MEMBER
3
APOSTOLAKIS:
GEH
has
also
4
responded to an RAI regarding passive systems in a
5
certain way.
We haven't discussed it at all.
6
CHAIRMAN CORRADINI:
We haven't seen it.
7
MEMBER STETKAR:
8
CHAIRMAN CORRADINI:
We got it.
9
MEMBER APOSTOLAKIS:
A 207 response.
Have we got it?
10
Harold said it was ten days ago.
11
means this is the current position.
12
MEMBER BLEY:
13
MEMBER SHACK:
14
That's why I was -We had that basically at
MEMBER APOSTOLAKIS:
16
more than a year.
17
This is the current position.
Correct, Harold?
MR. VANDER MOLLEN:
19
MEMBER APOSTOLAKIS:
Yes.
Okay.
So can we
spend half an hour tomorrow on this?
MEMBER ABDEL-KHALIK:
21
Good idea.
Yes.
We should, yes.
CHAIRMAN CORRADINI:
23
24
We have had it for
And then Harold sent it again.
18
22
And I assume that
the end of the last meeting it was around.
15
20
And
You guys are all
right for tomorrow?
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
339
1
MR. CARUSO:
2
CHAIRMAN CORRADINI:
3
MR. WACHOWIAK:
we
were
going
to
We'll address it tomorrow.
fit
Okay.
Good.
We tried to figure out
4
how
5
sequences, but we are prepared to talk about the
6
TRACG runs and things.
MEMBER STETKAR:
7
it
into
one
of
your
Part of the reason for
8
identifying the sequences was actually a context.
9
thought that there were some very specific -CHAIRMAN CORRADINI:
10
So we're adjourned.
I
So are we all set?
11
All right.
We're back tomorrow.
12
(Whereupon, the foregoing matter was recessed at 4:53
13
p.m., to be reconvened on Friday, August
14
22, 2008, at 8:30 a.m.)
15
16
17
18
19
20
21
22
23
24
NEAL R. GROSS
COURT REPORTERS AND TRANSCRIBERS
1323 RHODE ISLAND AVE., N.W.
(202) 234-4433
WASHINGTON, D.C. 20005-3701
www.nealrgross.com
ESBWR Severe Accident Management
Rick Wachowiak
Technical Lead ESBWR PRA
Glen Seeman
Senior Engineer ESBWR PRA
August 21, 2008
Scope of Severe Accident Analyses
Discussion of severe accident prevention
• Examples: ATWS, SBO, Fire Protection & ISLOCA
• Covered in previous meetings
Discussion of severe accident mitigation
• Examples: Hydrogen control, debris coolability,
high-pressure melt eject, containment performance,
containment vent, equipment survivability
Severe accident mitigation design alternatives
Contained in DCD Ch 19, NEDO-33201 Ch 21, and
NEDO-33306
2
Severe Accident Evaluation Process Overview
For Each Threat
Failure
Modes
no
Bounding
Loads
Probabilistic
Treatment
Mitigating
Features
yes
Fragilities
Success Criteria
Fault Tree Model
ROAAM Review
Containment Event Tree
Results
3
Hydrogen Generation and Control
Detonation
• Inert containment atmosphere precludes H2 burn
• No credit for containment while deinerted
Overpressure
• Containment ultimate strength fragility
• Reacting all Zr surrounding fuel does not fail
containment
4
Zr Surrounding
Fuel Reacted
Containment Fragility (500 oF)
5
Core Debris Coolability – Ex-Vessel
Basemat Melt Penetration
• Flood lower drywell
– Fault tree model for actuation
• Large spread area
– No guarantee that debris is coolable from above
– Because of this significant uncertainty, ESBWR PRA does
not credit this cooling mechanism which was found
acceptable in previous certified designs
• BiMAC
– Local burnout
Confirmatory testing
– Water depletion
PCCS fault tree model
– Local melt through
Sacrificial layer
6
The Basemat internal Melt Arrest and Coolability
(BiMAC) device
7
BiMAC Configuration
All Numerical Values are Preliminary
8
Pipes cover entire floor
BiMAC Flow Path
9
Natural Convection in BiMAC
BiMAC Range
10
Wetting of BiMAC Horizontal Channels
Dryout
(approximate)
BiMAC Range
11
Thermal Loads against Coolability Limits in
BiMAC Channels
12
BiMAC Thermal-Hydraulic Testing
Results provided in NEDE-33392P
• Demonstrates that the analytical results presented
on the previous slides are bounding
• Even a few degrees of subcooling greatly enhances
the performance of the BiMAC
• Staff is reviewing this document to close a
significant open item
13
Test Overview
14
High Pressure Melt Eject
Direct containment heating
• Assume bounding physical parameters for HPME
• Pressure suppression containment absorbs dynamic
load
Local liner failures
• Liner anchorage prevents release path
15
Quantification of DCH Loads
Identified three dynamic regimes
Used complete space (up to all fuel, Zr, and SS) to bound independently each
failure mode
12
Uppe r drywell
Lowe r drywell
We twell
P re s s ure , ba r
P re s s ure , ba r
10
6
8
6
4
Regime I
HYPOTHETICAL
2
0
1
2
3
Time ,s
4
4
3
2
5
Uppe r drywe ll
Lowe r drywe ll
We twe ll
5
1
0
Regime II
Creep Rupture Bounding
5
10
Time ,s
15
20
Regime III Expected (not shown)
16
Minimum (Bounding) Margins to Energetic DCH
Failure
Upper Bound Load
Lower Bound Fragility
17
Fuel Coolant Interaction
Ex-vessel steam explosion
• Damage pedestal wall
– Very deep, subcooled pool in LDW
• Damage BiMAC pipes
– Deep, subcooled pool in LDW
• Minimize water in LDW prior to vessel breach
– BiMAC does not require pre-flooded LDW
18
Pedestal Failure Margins to EVE
1 to 2 m Subcooled Pools
Upper Bound Load
Lower Bound Fragility
19
BiMAC Failure Margins Due to EVE
1-2 m subcooled pools
Upper Bound Load
Saturated Low Level
Upper Bound Load
Subcooled 1-2 m
20
Containment Overpressurization – Long Term
Containment systems mitigate this threat
• Passive Containment Cooling System
• Vacuum breakers
• Active RHR systems
• Venting – treated as large release in Level II
Treated probabilistically using fault tree models
21
Conclusions
Most open items have been resolved
BiMAC test report still under review
~ 30 questions received last week
With closure, ESBWR severe accident evaluations
expected to be determined acceptable for design
certification
22
ESBWR PRA Focused Review
GEH PRA Team
August 21 - 22, 2008
GEH PRA Team Representatives
Rick Wachowiak
Gary Miller
Glen Seeman
Jonathan Li
Justin Howe
Lou Lanese
2
Meeting Purpose
Gain an understanding of the technical quality of the
ESBWR PRA
Review the degree of completeness of the ESBWR
PRA
Investigate details of the ESBWR PRA
Accomplished through a detailed review of four
selected sequences
3
Quality and Scope
Required elements in a PRA depend on the application
RG 1.200
• “… the staff’s recognition that the PRA needed to support
regulatory decisions can vary (i.e., that the “scope, level of detail,
and quality of the PRA is to be commensurate with the application
for which it is intended and the role the PRA results play in the
integrated decision process”).”
Interim Staff Guidance
• “PRAs that meet the applicable supporting requirements for
Capability Category I and meet the high level requirements as
defined in the ASME PRA Standard (ASME-RA-Sb-2005) should
generally be acceptable for DC and COL applications.”
4
Design Certification PRA Objectives
Identify vulnerabilities
Reduce/eliminate risk contributors in existing plants
Select among design and operational features
Confirm design robustness
Identify risk significance of operator actions associated with
design
Demonstrate that the plant meets the Commission’s safety goals
Show a balance of prevention and mitigation
Show a reduction in risk in comparison to existing plants
Address known design issues with respect to core and
containment heat removal systems
5
Design Certification Not the Last ESBWR PRA
10 CFR 50.71(h)(1) states that no later than the scheduled
date for initial loading of fuel, each holder of a COL shall
develop a level 1 and a level 2 PRA. The PRA must cover
those initiating events and modes for which NRC-endorsed
consensus standards on PRA exist one year prior to the
scheduled date for initial loading of fuel.
It is not required to submit this PRA to the NRC, but instead
should be maintained by the licensee for NRC inspection.
The need for any such submittal or review would be
determined by any risk-informed application for which the
licensee might wish to use this PRA, such as in support of
licensing actions.
6
Ongoing PRA Upgrade Requirements
10 CFR 50.71(h)(2) states that each COL holder must
maintain and upgrade the PRA required by 10 CFR
50.71(h)(1). The upgraded PRA must cover initiating events
and modes of operation contained in NRC-endorsed
consensus standards on PRA in effect 1 year prior to each
required upgrade. The PRA must be upgraded every 4 years
until the permanent cessation of operations under 10 CFR
52.110(a).
PRA maintenance and PRA upgrade will be consistent with
how they are defined in the American Society of Mechanical
Engineers (ASME) “Standard for Probabilistic Risk
Assessment for Nuclear Power Plant Applications”
7
ESBWR Design Certification PRA
Meets the scope and quality for certification
Meets the scope and quality for COL given no
significant departures from the certified design
Provides a starting point for operating plant PRA
8
Presentation to the ACRS Subcommittee
ESBWR Design Certification Review
Chapter 19.3 of DCD, Tier 2
Presented by NRO/SPLB
August 21, 2008
1
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
Purpose:
• Brief the Subcommittee on the results of the staff’s
review of the ESBWR DCD application, Chapter 19.3,
Severe Accident Evaluations
2
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
Outline of Presentation:
• Applicable Regulations
• SER Technical Topics
• Significant Open Items
3
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
Regulatory Requirements:
•
•
•
•
10 CFR 52.47(a)(8) – comply with TMI requirements
10 CFR 52.47(a)(21) - resolve USI/GSI
10 CFR 52.47(a)(23) – provide description and analysis of design
features for prevention and mitigation of severe accidents
10 CFR 52.47(b)(2) – provide an environmental report, as required by
10 CFR 51.55, that addresses the costs and benefits of severe
accident mitigation design alternatives, and the basis for not
incorporating these in the design to be certified.
Regulatory Guidance:
•
•
•
Policy Statements on Severe Accidents and Use of PRA
SECY-93-087, SECY-96-128, and SECY-97-044 - guidance for
implementing features in new designs to prevent or mitigate severe
accidents
Regulatory Guide 1.206 and SRP Chapters 19.0 and 19.1
4
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
19.2 of SER: Severe Accident Evaluations
Technical Topics:
•
•
•
•
•
19.2.2
19.2.3
19.2.4
19.2.5
19.2.6
Severe Accident Prevention
Severe Accident Mitigation
Containment Performance Capability
Accident Management
Severe Accident Mitigation Design
Alternatives (GEH documentation in
NEDO-33306, Rev 1, August 2007)
5
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
19.2.3 Severe Accident Mitigation
Significant Open Items:
• BiMAC performance test report
– Response to RAIs19.2-23 S02 and 19.2-25 S02
included a topical report documenting the results
of the BiMAC tests.
– Topical report NEDE-33392 has been reviewed
and 27 RAIs issued.
6
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
•
BiMAC Performance RAIs
The review focused on:
– Adequacy of test facility scale for applicability to ESBWR configuration and
design.
• Generally, the test facility adequately scales expected prototypical conditions
• RAI on the scaling basis of the multi-channel tests.
– The range of measured test data as compared with severe accident loading
conditions.
• RAIs on relevant tests for near-edge tubes, range of heat fluxes chosen for tests,
and other issues.
– Adequacy of the theoretical predictions as compared to the data.
• the single tube independent (NRC) theoretical predictions seem to support the
experimental measurements.
– Implications of the BiMAC design on ESBWR operational safety.
• RAIs on thermal load boundary conditions, the use of CFD simulations to obtain
boundary conditions, the structural integrity of Zirconia, effects of crusts on heat
loads, and other issues.
7
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
19.2.3 Severe Accident Mitigation
(cont.)
Significant Open Items:
• Vacuum breaker performance
– Further information was requested on vacuum breaker
design (including isolation valves), coverage in DCD and
ITAAC, and on emergency procedures related to failed
vacuum breakers.
– Responses to RAIs 19.2-6, 19.2-10, and 19.2-11 have
recently been received and are acceptable.
8
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
19.2.4 Containment Performance Capability
Significant Open Items:
•
Calculated upper drywell liner strain exceeds Level-C limit under
conditions of 100% metal/water reaction
Response received from GEH for RAI 19.2-86 and issue is
resolved.
•
Temperature boundary condition for drywell head in finite element
model set incorrectly at 110 °F versus drywell air space temp of 500 °F
Response received from GEH for RAI 19.2-41 Supplement 2 and
issue is resolved.
9
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
19.2.5 Accident Management
Significant Open Items:
• Description of the process for developing Severe
Accident Guidelines
– The staff requested additional information on the process
that will be used by GEH to develop the Severe Accident
Guidelines (SAGs) in RAI 19.2.4-1 and its supplements.
– A new supplemental RAI has been prepared, asking for the
technical basis for ESBWR severe accident management.
10
ACRS Subcommittee Presentation
ESBWR Design Certification Review
Chapter 19
Discussion / Questions
11