Materials and Metallurgy - September 26, 2001
Official Transcript of Proceedings
NUCLEAR REGULATORY COMMISSION
Title: Advisory Committee on Reactor Safeguards
Materials and Metallurgy Subcommittee
Steam Generator Action Plan
Docket Number: (not applicable)
Location: Rockville, Maryland
Date: Wednesday, September 26, 2001
Work Order No.: NRC-032 Pages 1-166
NEAL R. GROSS AND CO., INC.
Court Reporters and Transcribers
1323 Rhode Island Avenue, N.W.
Washington, D.C. 20005
(202) 234-4433
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
+ + + + +
ADVISORY COMMITTEE ON REACTOR SAFEGUARDS
MATERIALS AND METALLURGY SUBCOMMITTEE
STEAM GENERATOR ACTION PLAN
(ACRS)
+ + + + +
WEDNESDAY
SEPTEMBER 26, 2001
+ + + + +
ROCKVILLE, MARYLAND
+ + + + +
The ACRS Materials and Metallurgy
Subcommittee met at the Nuclear Regulatory Commission,
Two White Flint North, Room T2B3, 11545 Rockville
Pike, at 8:31 a.m., Dr. F. Peter Ford, Chairman,
presiding.
COMMITTEE MEMBERS PRESENT:
DR. F. PETER FORD, Chairman
DR. MARIO V. BONACA, Member
DR. THOMAS S. KRESS, Member
DR. DANA POWERS, Member
DR. WILLIAM J. SHACK, Member
DR. JOHN D. SIEBER, Member
ACRS STAFF PRESENT:
NOEL F. DUDLEY,
ACRS Cognizant Staff Engineer
I-N-D-E-X
AGENDA ITEM PAGE
Opening Remarks by the Chairman. . . . . . . . . . 4
Introductory Remarks by Mr. Sullivan . . . . . . . 6
Presentation by M. Banerjee on Status of . . . . . 8
Steam Generator Action Plan
Presentation by Steve Long on. . . . . . . . . . .23
Status of Action Plan DPO Issues
Presentation by Ted Sullivan on NEI 97-06. . . . .26
Presentation by Kenneth Karwoski on. . . . . . . .70
Overview of South Texas Steam Generator
Tube Integrity Issues
Presentation by Joseph Muscara on SG . . . . . . 105
Action Plan
Presentation by Charles Tinkler on . . . . . . . 135
Overview of Severe Accidents
Presentation by Stephen Bajorek on . . . . . . . 145
Thermal Hydraulics
Presentation by Christopher Boyd on. . . . . . . 153
CFD Predictions
P-R-O-C-E-E-D-I-N-G-S
(8:31 a.m.)
CHAIRMAN FORD: The meeting will now come
to order. This is a meeting of the ACRS Subcommittee
on Materials and Metallurgy. I am Peter Ford,
Chairman of the Subcommittee.
ACRS Members in attendance are William
Shack, Mario Bonaca, Thomas Kress, John Sieber, and
Dana Powers, and hopefully Steve Rosen.
The purpose of this meeting is to discuss
the status of the staff's Steam Generator Action Plan
and South Texas, Unit 2, steam generator tube leakage,
and to decide what further ACRS reviews should be
scheduled.
The Subcommittee will gather information,
analyze relevant issues and facts, and formulate the
proposed positions and actions, as appropriate, for
deliberation to the full Committee. Noel Dudley is
the Cognizant ACRS staff engineer at this meeting.
The rules for participation in today's
meeting have been announced as part of the notice of
this meeting previously published in the Federal
Register on September 11th, 2001.
A transcript of the meeting is being kept,
and will be made available as stated in the Federal
Register Notice. It is requested that speakers first
identify themselves and speak with sufficient clarity
and volume so that they can be readily heard.
We have received no written comments or
requests for time to make oral statements from members
of the public regarding today's meeting.
The staff issued the Steam Generator
Action Plan on November 16, 2000. The Action plan
consolidated half a dozen or more staff regulatory
activities related to steam generator tube integrity.
The staff updated the Action Plan on May
11th, 2001, to include items associated with the
differing professional opinion associated with steam
generator tube integrity.
After hearing the staff's presentation, we
will develop recommendations on what activities we
want to review and comment on, and when we should
schedule those reviews.
We will now proceed with the meeting, and
I call upon Maitri Banjeree, of the Division of
Engineering, Office of Nuclear Reactor Regulation, to
begin.
DR. SHACK: Before we start, Mr.
Chairman, I should mention that I have a conflict of
interest here because Oregon is doing work on steam
generators for the NRC.
CHAIRMAN FORD: A;ll right.
DR. POWERS: Is that why they keep falling
apart all the time?
CHAIRMAN FORD: Oh, I'm sorry.
MR. SULLIVAN: My name is Ted Sullivan,
and Maitri is the next speaker. I will just take a
minute and spent a little bit on the introduction to
give you a little bit more information on what we are
going to be doing this morning.
Maitri is our first speaker, and she is
going to be giving an introduction to the steam
generator action plan, and basically tell you what
some of the early activities were that led to the
development of the action plan, and what it considers,
and what it doesn't consider.
One of the major elements in that action
plan is NEI 97-06, which is our steam generator
regulatory framework initiative that we have been
working on for quite some time.
So I am going to get up after Maitri and
give a presentation on the status of that, and the
issues that we are currently dealing with that are
holding us up from completing that initiative.
After the break, Joe Muscara is going to
give a presentation on the DPO related issues in the
action plan. His focus is not going to be going
through the entire set of issues in that portion of
the plan.
Rather, he is going to focus more on the
near term activities. We thought that would be of
more benefit. And then after that, Ken Karwoski is
going to do two things. Basically, he is going to
discuss two of the action plan items related to the
DPO that are NRR responsibilities, as opposed to
research.
And then he is going to transition into a
discussion of what has been going on in the past
couple of intervals related to the South Texas use of
voltage based repair criteria.
And I agree with what you had to say in
terms of the objective. I think that we are not going
to get into a tremendous amount of detail, as we are
covering a lot of material here. So I think it would
be good to decide what additional briefings you would
like.
And certainly in the area of NEI 97-06, we
are prepared to get into more detail if you are
interested in a subsequent briefing.
DR. POWERS: And in what phase of the
briefing will we discuss the iodine spiking issue?
MR. SULLIVAN: It should be covered in the
DPO portion, but Joe, can you address that?
MR. MUSCARA: I will have one view graph
on the status of the operation.
DR. POWERS: Okay. An in-depth
discussion, I can tell. This is an easy issue to
solve, Joe.
MR. MUSCARA: That's what they tell me.
CHAIRMAN FORD: Thank you, Ted.
MR. KARWOSKI: The first question is are
you related to Sanjo Banerjee?
MS. BANERJEE: Not that I know of.
MR. KARWOSKI: Okay. Then you are okay
then.
MS. BANERJEE: That's reassuring. My name
is Maitri Banerjee, and I am the NRR lead project
manager for the steam generator action plan, and I
will provide you a short background and overall status
of information on the action plan. Can everybody see
this slide?
All right. Here is a historic overview of
the --
DR. POWERS: History begins with an IP2?
MS. BANERJEE: And of significant actions
taken, and that led to the issuance of the steam
generator action plan, and this kind of explains
itself.
The purpose of the plan. As Chairman Ford
pointed out the plan was originally issued in November
of 2000, and it was issued keeping the NRC performance
goals in mind, and in maintaining safety in the IP2
area, and renewing public confidence, and also using
NRC and stakeholder's resources effectively and
efficiently.
And the purpose of the plan is to direct,
monitor, and track NRC's activities to completion so
that we get to an integrated steam generator
regulatory framework.
DR. POWERS: Can I ask what an integrated
regulatory framework means?
MS. BANERJEE: Well, I am going to defer
answering that question to Ted Sullivan, who is going
to talk about NEI 97-06 activities that are going on.
DR. POWERS: Well, maybe you can give me
an idea of what we are integrating with what.
MR. SULLIVAN: My name is Ted Sullivan,
and I think what we are trying to do is to make sure
that all of the various elements involved in ensuring
tube integrity are integrated into a steam generator
regulatory framework that considers more than just,
say, inspection and repair issues.
But that goes beyond that into all the
other disciplines that are involved in ensuring tube
integrity. Disciplines related to doing risk
assessment, and the research that is developed that
feeds into that, and that sort of thing. I think that
is the general idea, and the radiological issues.
MS. BANERJEE: Do you have any other
questions? If not, the action plan consolidates a
number of activities, including Indian Point 2 Lessons
Learned Task Group Report, and the OIG report that was
issued subsequent to that, and then it was revised in
May to incorporate the steam generator DPO related
issues.
And obviously the milestones related to
the staff review of NEI 97-06 is in there, and we will
make revisions in the future to incorporate milestones
for resolution of GSI 163.
We also anticipate revisions to
incorporate GSI 188 and Draft Guide 1073.
CHAIRMAN FORD: Could I just for clarify?
The resolution of the steam generator DPO, that is
essentially the output from the ad hoc committee,
subcommittee from ACRS?
MS. BANERJEE: Yes, that's correct, from
the NUREG requisition. The steam generator action
plan also includes some non-steam generator related
issues that came out of the OIG report. They had
issues in the EP area, and also that task group's
report.
And the second bullet is sort of a
disclaimer. It says that the action plan doesn't
address any plan-specific reviews or industry efforts
related to voltage-based tube repair criteria.
CHAIRMAN FORD: Is there a reason for that
disclaimer? Why the disclaimer?
MS. BANERJEE: I guess these are plan-
specific issues that are not addressed in the action
plan. The action plan is basically what came out of
the Indian Point 2 lessons learned task group, and
what came out of the OIG report subsequent to Indian
Point 2, and also the DPO related issues.
And so we didn't go into addressing
Generic Letter 95-05, any kind of industry work being
done in that area, or any kind of plant-specific
licensing work related to voltage-based tube repair
criteria.
CHAIRMAN FORD: But surely as you go
through the action plan, which is your calculations,
experiments, and studies, there has got to be a
feedback into what the plant is actually doing.
MS. BANERJEE: Ultimately, yes.
CHAIRMAN FORD: And so when does that
occur? That second bullet is saying, hey, we stopped
short of actually calibrating our calculations against
what is in fact happening. Isn't that a over
simplification of what that statement is saying?
MR. SULLIVAN: I think what we were trying
to say is that there is a lot of plant-specific
reviews that are going on. They continually go on.
They might have to do with ultimate repair
criteria that we maybe reviewing, and what we are
basically saying is that they are tracked in other
systems, and so we weren't going to track them in the
action plan.
And then related to the second half of
that, they are a number of issues that industry has
been asking us to take on, their proposed
modifications to GL 95-05, and that in a sense would
be relaxations.
And the staff's view was that the priority
effort should be on the action plan when resources are
available, and we will get back to taking those kinds
of reviews on. So for the second half of that, it was
really more of a priority of resources matter.
MS. BANERJEE: Thank you, Ted. This slide
presents an overall status of the action plans.
Currently, we have 40 major items, milestones, in the
action plan, 11 of which consist or came out of DPO.
And 20 of the 40 major milestones are
completed, and there is one milestone with a schedule
to be determined. This has to do with how we
communicate risk to the public.
The agency has done some work in the area
of communication plan and currently NRR is looking at
ways to improve that. And that is the overall status.
This slide lists some of the significant
activities in the action plan. A regulatory summary
was issued in November of 2000, with experience from
Indian Point 2 and ANL, and a number of issues were
raised by both task groups, and the OIG related to
steam generator inspections, GSI inspections.
And in response to that the base line
inspection procedure was revised. It focuses on the
steam generator ISI inspector, in terms of how the
licensee is going condition monitoring, and how they
are meeting the performance criteria, versus looking
at any current testing.
A risk informed significance determination
process is being developed for ISI inspection results,
and NRC's findings related to that, and with inspector
training, we will be providing written material,
written packages, for inspector training related to
the new inspection program in October.
And formal training will be provided to
the regional inspectors in February. In terms of
steam generator tube leakage, technical guidance is
being developed and will be provided to the regions
some time in the very near future.
And this has to do with helping the
regional inspectors oversight of PWRs with steam
generator tube leak, and help them understand the role
of the primary to second leaking monitoring in
assuring steam generator tube integrity.
And in the area of steam generator
performance indicators, we have done some review, and
a decision was made not to add any new PI related to
steam generators.
And our next bullet has to do with
conference calls during outages. The NRR staff will
continue doing the conference calls with the licensees
during -- the selected licensees during the outages,
and we will docket the telephone summary.
And we will also formally review their ISI
results report, which sometimes is called the 90-day
report. A steam generator workshop was held with
stakeholders in February, and the regulatory
information conference also had discussions on steam
generator issues.
The next slide is a continuation of this
slide. Both the task group and the OIG made
recommendations for some improvements to NRR's process
for license amendment reviews, and changes were made
in response to that.
As I mentioned before, NEI 97-06, Ted
Sullivan will provide a detailed discussion on that.
Subsequent to Indian Point 2, as you all know, the
staff stopped its review of NEI 97-06, and we
recommenced in January of this year.
And so a lot of activities are going on in
that area. And then a web page was developed and
being maintained for internal and external access.
And risk communication, that has already been
mentioned on what we are doing.
And milestones for ACRS' recommendation on
the DPO, and we have a much more detailed presentation
by Jim Muscara as Ted mentioned; and the last bullet,
as I mentioned before, are future activities.
CHAIRMAN FORD: Is there a particular
reason why this NEI 97-06 was put on hold?
MR. SULLIVAN: Dr. Ford, I am going to be
getting into that. I plan to cover your question.
CHAIRMAN FORD: Okay.
MS. BANERJEE: This slide is on the
management of the action plan. We will formally
document completion of each major milestone, and we
will be coordinating a resolution of issues with
external and internal stakeholders. Like all of our
meetings with NEI, they are open to the public.
And the status of the milestones are
updated, and a complete copy of the milestones is
maintained in NRR's Director's Quarterly Status
Report, and an abbreviated version in is the CTM.
The CTM is updated monthly and the QSR is
updated quarterly. And the overall management of the
action plan is the responsibility of the projects in
NRR. This completes my presentation.
CHAIRMAN FORD: Maitri, as I look through
all the milestones and their completion dates,
starting back from the earliest of these action plans,
a tremendous number of them are way, way behind, a
year behind in completion. Is there a reason for
this?
MR. SULLIVAN: When you say behind, do you
mean delayed or do you mean scheduled for some time?
CHAIRMAN FORD: Well, in these lists here,
I see the targeted completion date, and you are way,
way beyond. Like NEI 97-06, there is just one, but
there are many others.
MS. BANERJEE: Like DPO has a lot of
milestones.
CHAIRMAN FORD: Well, I am just putting
this in general. All of them are way, way behind on
schedule. Is there a particular reason for this
delay?
MS. BANERJEE: As far as I can tell, some
of the actions are a little bit behind, but in terms
of scheduling those milestones into the distance or
future is because of all the activities that needed to
be completed before we can get there.
And that is a considerable amount of work
that needed to be done, especially in the area of the
DPO recommendations.
CHAIRMAN FORD: So it is manpower and
dollar constriction on completing those?
MR. SULLIVAN: I think that is true, along
with all the other work that was already in place
before we developed the action plan.
CHAIRMAN FORD: Okay.
MR. SULLIVAN: I think the major delays
are in the NEI 97-06. A number of other items -- you
are right -- they did slip, but usually on the order
of not too many months; and the DPO work, I wouldn't
characterize it as having been slipped.
The schedules were based on the research
plans that were pretty much in existence when the ACRS
report came out.
CHAIRMAN FORD: Okay. Thank you.
MS. BANERJEE: Any other questions?
DR. POWERS: I am curious about the train
of reasoning that went about to decide that there
would be no performance indicator for steam generator
tubes.
And I am perplexed in this area because I
remind myself that steam generator tube rupture
accidents are risk dominant for a number of plants;
and bypass accidents in general are risk dominant.
And seldom do you have a more direct
indicator of risk than steam generator performance.
So what was the rationale that went about not having
a PI for steam generator performance?
MS. BANERJEE: The way I understand it is
that the staff considered three potential Pis. One
had to do with tube degradation, and one had to do
with integrity of the tube integrity; and another one
had to do with primary to secondary leakage.
The purpose of a PI is to give you only
indications of things going south, and in the case of
the first two, they are only information or new
information is only available during outages, which
happens every 18 to 22 or 24 months.
So the staff concluded after a lot of
consideration that it doesn't really provide you with
an indicator in all cases. And then in terms of
primary to secondary leakage, the relationship of the
steam generator performance with the leakage is not
very clearly established, and we don't even know that
it could be established.
Because like in the case of Indian Point
2, we have not seen a tremendous amount of leakage to
happen before an event occurred. So considering all
of that, a conclusion was made that at this point we
don't have a real good parameter which we can use as
an early indicator of problems. Does anybody on the
staff want to add more to that?
MS. KHAN: I think that summed it up
pretty well. By the way, my name is Cheryl Khan, and
I work in materials in the chemical engineering branch
in NRR.
But that pretty well sums it up as far as
the main viewpoints, and as Maitri indicated, the
first two that she mentioned didn't really fit the
typical type of performance indicator, the parameters.
It needs to be an ongoing parameter that
you are monitoring continuously; and with respect to
the third one, as she indicated, leakage is not
necessarily correlated to the real condition of what
is going on, and to generate as far as how significant
the issue is.
And in fact the issue may be more
significant compared to the leakages. So it was not
felt that that really was an appropriate term to
monitor a performance indicator.
The ones that we took beyond that was that
what the performance indicators would have provided to
us was the capability to take some type of actions if
there were signs of degradation occurring in the steam
generators or issues of significance occurring in the
steam generators.
And so the way that we tried to address
that is through the inspection process in lieu of
using performance indicators, because it is typically
an either/or.
And so through the inspection process the
intent is that there are periodic inspections that are
being performed under in-service inspection procedure,
and it incorporates with the in-service inspection
program, as well as steam generator inspection
activities.
And there are -- there is a means, that
dependent on the outcome both of the inspection, the
NRC's inspection, as well as what the licensee is
finding, that there is the potential to take immediate
action, meaning further NRC inspection and
involvement.
And we felt that was more appropriate,
because that is when the degradation and issues would
be clearly identified, and then we would be able to
take immediate action if they were significant enough.
DR. POWERS: So from that I conclude that
the first decision was that since we couldn't get
information, except for every 18 months or every
outage, we would take no PI at all.
And that the second one is that because
the correlation between leakage and tube condition,
which is good enough for the alternate criteria, is
not good enough for monitoring the plant?
MR. SULLIVAN: Excuse me, but what do you
mean by good enough for alternate repair criteria?
DR. POWERS: Well, it's used. The
correlation is used as part of the alternate repair
criteria.
MR. SULLIVAN: I think that one of the
factors that we considered in terms of primary to
secondary leakage was that the information that was --
we had originally proposed that we go down that road
and look, and what we were advised was that it wasn't
necessary to put this in as a performance indicator in
order to get that information.
We get that information on a daily basis
from plants that are experiencing leakage. And we are
involved in it in the sense that the regions will
typically inform us of when the leakage is increasing,
and they are going to have phone calls with licensees,
and we get involved in those phone calls.
So we really felt that adding a
performance indicator in this arena wasn't really
going to substantially add to our ability to conduct
oversight.
MS. BANERJEE: That is one thing that the
resident inspectors review in their daily status
inspections.
CHAIRMAN FORD: Thank you very much.
MR. LONG: This is Steve Long, and I am in
NRR in the risk assessment group, and I just wanted to
add something on the relationship for the performance
indicators and the parameters we measure.
When the reactor is operating the only
thing we are really getting information on is leakage
during normal operation. We don't know what the
leaking would be if there was an off-normal condition
because the off-normal condition isn't there.
So it is very hard to relate a very small
operational leakage number to anything that will help
us figure out what the actual risk at that time is.
When we shut down the plants and inspect the plants,
then we have good information.
And the thing that was not mentioned here
that I want to add is that at that point, if there are
findings of degradation, we are developing a
significance determination process for those findings.
Those actions go into the action matrix,
like the performance indicators go into the action
matrix, for making a decision about how we are going
to inspect and regulate the plant.
So instead of having a performance
indicator that is being updated every three months,
and that only be tied to an observation every three
months that is not necessarily in any quantitative way
tied to the risk, we decided to go with the
determination of significance of inspection findings
when something is determined not to be needing the
performance -- you know, the performance on tube
integrity, leak tightness and structural integrity,
and that sort of thing.
But that information is still going under
the action matrix, just like a performance indicator
would, and we are still making regulatory decisions on
that information. It is a timeliness thing.
DR. POWERS: And we don't have a SDP for
these findings right now?
MR. LONG: That is one of the action
matrix -- excuse me, but that is one of the action
plan items, and where that stands at the moment is we
are just signing out a review of what needs to be
done, and some suggestions that are going down to the
branch that is responsible for implementing that into
procedures. So that is in the process.
DR. DUDLEY: Do you have a feel for when
that might be available for ACRS review?
MR. LONG: It is supposed to be in ADAMS
now, but we had a little glitch. It is going to be in
ADAMS by the end of the month I promise.
DR. POWERS: Yes, but when can we get it?
MR. SULLIVAN: I previously introduced
myself as Ted Sullivan, and I am going to be talking
about NEI 97-06.
DR. KRESS: And you are still Ted
Sullivan?
MR. SULLIVAN: Yes. We have had a number
of briefings with the ACRS, and I am going to actually
go through that towards the end of this view graph a
little bit.
I had gone over this, but I thought it
would be worth it to spend a very brief time on some
background, starting with something that I think we
have started all these briefings with, which is to
state that the current requirements, particularly as
imbedded in the text specs, are prescriptive and out
of date.
They go back to the '70s. These
requirements are not focused on the key objective of
ensuring tube integrity for the entire period between
in-service inspections.
Rather, they are inspection and repair
oriented, and they don't focus on the time that steam
generators can operate between inspections and
maintain safety margins.
And recognizing that the staff began
initiatives in probably the early '90s, beginning with
a rule making initiative in the mid-1990s that turned
out not to be a vehicle that we could use.
We briefed the ACRS on that in '96, and
several times in 1997. We discussed with the ACRS in
1997 a change in strategy to a generic letter. We
proceeded down that path for probably a year or a
year-and-a-half.
And at the same time as that was going on,
NEI was developing its 97-06 steam generator program
guidelines initiative, and I believe in the '98 or
early '99 time frame -- I think the '98 time frame --
we began discussions with NEI regarding putting the
generic letter on hold, and switching our focus to a
new regulatory framework based on NEI 97-06.
Throughout a lot of 1999, we held meetings
and discussions with NEI and other industry
counterparts on a generic change package that was
being developed. The generic change package is kind
of a centerpiece of proposed technical specifications.
And we had reached some tentative
agreement on drafts of the generic change package in
late '99, and NEI then went through its process of
issuing it. It was issued on February 4th of 2000,
shortly before the Indian Point-2 tube rupture, less
than two weeks before that.
I think as Maitri mentioned, we suspended
are review after the Indian Point-2 rupture for
basically two reasons. One was that our resources
were devoted or diverted to Indian Point-2 recovery.
A lot of staff resources went into
reviewing the restart plans and the operational
assessment that Con-Ed was producing and working on.
Prior to that, we were reviewing and participating in
NRC inspections related to the Con-Ed steam generator
inspections.
And also some of our staff was diverted to
the lessons learned task force. So that was sort of
reason number one. Reason number two was that we
really wanted to wait and see what came out of the
Indian Point-2 lessons learned, and factor them back
into the review.
So we didn't want to really make a false
start. It wasn't that we had a lot of time that we
were sitting anyway. The two things came together
nicely, but we did deliberately indicate to various
constituents that we weren't going to do the review,
or commence the review, until the lessons learned
study was finished and until we had a chance to look
at it.
DR. SHACK: Ted, every time we look at a
license renewal with a steam generator and we look at
GALL, everybody seems to be using 97-06. So that
means that they are under a dual sort of system. They
use 97-06 for their own tracking and monitoring
purposes, and yet they still meet their tech specs
also? Is that the way that the system is working now?
MR. SULLIVAN: That's correct. Licensees
have all committed in a manner that I think Jim Riley
could elaborate on if you want, but it is basically an
internal industry arrangement that every PWR licensee
is committed to implement NEI 97-06 for a couple of
years now. And I think it was at the first refueling
after January of 1999.
DR. SHACK: Now, how many PWRs are
actually running under 95-05? That is, at least for
their tube support plate degradation, and they are
really controlled by 95-05 rather than the old 40
percent through wall kind of thing.
MR. SULLIVAN: For that mode of
degradation, yes. If the controlling document is tech
spec amendment dealing with 95-05, and it is on the
order of a dozen plants, I am not sure if that is
accurate.
DR. SHACK: So there is still 600 mil
anneal plants that don't use 95-05?
MR. SULLIVAN: Yes, there are quite a
number, probably on the order of about two-thirds of
them, I guess. I mean, I think about half of the
plants have replaced roughly, and so that is on the
order of about -- between 30 and 35.
DR. SHACK: Yes, I was just looking at the
Mil Anneal 600 plants, yes.
MR. SULLIVAN: And that is what I am
talking about. About half still have Mill Anneal 600,
and half have replaced, and a dozen of that 30 to 35
reactors use generic letter 95-05 for ODSCC tubes or
plates.
The staff review of the generic change
package when we commenced that review included a
consideration of issues associated with the lessons
learned report.
A regulatory issue summary of 2022, which
Maitri mentioned, but I will just elaborate very
briefly to say that it described technical issues that
came out of the staff review of Con-Edison's Indian
Point-2 restart assessment, as well as an operational
assessment of Arkansas Nuclear Unit-2.
And that basically led to a mid-cycle
inspection. It was not exactly mid-cycle literally.
It was sort of late cycle inspection, an additional
inspection, during the summer of 2000.
And then we have also considered the DPO
action plan issues that were developed in response to
the ACRS report. I will go over this briefly as it is
nothing new.
And even as far back as the rule making,
our intent was to put in place a new regulatory
framework that has these features that are in bold.
That is, that it is performance based, and it
establishes performance criteria for ensuring tube
integrity and leaking integrity under normal and
accident conditions.
So I am going to elaborate a little bit
more on that later when I get into a brief discussion
of performance criteria. Performance criteria are in
terms of parameters that are measurable and tolerable.
The framework is supposed to be flexible,
in that the methods for meeting the performance
criteria are up to the licensee. It should be
adaptable to changing mechanisms and technology which
a prescriptive framework would not be.
And it is risk-informed to ensure that no
-- that there is no significant increase in risk
associated with operational steam generators.
CHAIRMAN FORD: If you could just go back
to that last slide.
MR. SULLIVAN: Sure.
CHAIRMAN FORD: Industrial parlance, would
you say that this is a stretch goal given the fact
that you no longer -- that you don't currently have
Pis, forced steam generators as I understand for
reasons that were just enunciated.
So this is really a wish list, and if I
look at the timing on your latest action plan, the one
that takes into account the NUREG 17.40
recommendations, you are looking several years out.
You are looking 2, 3, 4 years out --
MR. SULLIVAN: Well, in terms of the
framework --
CHAIRMAN FORD: before you can have this.
MR. SULLIVAN: In terms of the framework,
not exactly. I will try and capture the time frame
that we have in mind. In terms of the framework
itself, we are -- and as I will discuss a little bit
later, we are probably not going to completely capture
the performance-based element.
We will incorporate it, but it won't be
strictly non-prescriptive. We still have to work this
through with NEI, and that is -- our current target
date for completion is April, and that is probably
optimistic.
CHAIRMAN FORD: After discussing it with
NEI?
MR. SULLIVAN: Well, our target date for
reaching resolution of NEI 97-06 is April, and I am
saying that may be optimistic. After we reach
resolution, which will entail some things that I am
going to talk about later having to do with issuing a
generic safety evaluation and so forth, the individual
plants have to send in tech spec amendments to put
this in place.
The tech spec amendment process could take
up to an additional year. So just that alone could
potentially take a year-and-a-half to two years. In
terms of the risk issues, I don't think we will
consider that we fully understand or more completely
understand risk until the other issues associated with
what I refer to as the 3.X items in the action plan
are completed.
And the action plan has 1.X, and 2.X, and
3.X items. The 1.X are steam generator related, and
the issues that came out of the lessons learned
report.
The 2.X items are the non-steam generator
related items that came out of the report; and the 3.X
items are the ones that basically relate to the ACRS
report on the DPL. So I am not sure if I have
confused things by that answer.
CHAIRMAN FORD: And I am sure it is
because of my lack of understanding of this whole
process. But standing back, as I understand it, we
have got a whole lot of reactors out there with steam
generators that are demonstratively cracking.
We are not too sure how to quantify the
progress of this cracking because of monitoring
discrepancies or restrictions, et cetera, and modeling
restrictions all go into this NUREG 17.40.
We don't have any Pis to tell us right now
on an ROP basis as to how we are doing. And what you
are just saying is that this is the wish list of where
you want to go, but it is going to be the middle of
next year before we have got the NEI thing reviewed,
and 97-06 reviewed, and signed off.
And the information for this is not going
to be around and approved without being used legally
if you like until another 5 or 6 years. So what
happens in the meantime? What is our backup plan?
MR. SULLIVAN: The intent is to put into
place a new regulatory framework which I am going to
cover in subsequent slides and describe in subsequent
slides.
CHAIRMAN FORD: I'm jumping in. Sorry.
MR. SULLIVAN: And the intent is to get
that in place for every PWR within about a year-and-a-
half, assuming -- and that schedule is contingent on
reaching resolution of the outstanding issues with NEI
and the industry. I noticed Jim Riley from NEI is
interested in adding to what I have been saying.
MR. RILEY: Hi, I am Jim Riley from NEI,
and I am NEI's project manager for steam generator
issues. I think a real important aspect of what we
are doing here is Ted's illusion to an NEI initiative
that is set in place.
So even though the regulatory framework
isn't there right now, and we are all working towards
it, the fact is that the plants are inspecting their
steam generators to a performance based program based
on NEI 97-06, which involves basically all these
things that Ted is talking about, the differences, and
we don't have the tech specs in place yet that give
the regulatory aspects of what we are doing some
substance.
But in fact the plants are all committed,
all the PWRs, to implementing NEI 97-06 and its
guidelines that are associated with it.
DR. POWERS: And Indian Point-2 was one of
those plants that followed this 97-06?
MR. RILEY: That's correct. I would like
to point out though that at the time that Indian
Point-2 did their inspection previous to their problem
was 1997, and at that point in time they had not
implemented 97-06 because it wasn't in place at that
time.
CHAIRMAN FORD: Could I ask my colleagues
have we seen 97-06?
DR. POWERS: Yes.
DR. SHACK: Yes.
DR. SIEBER: Before you take that slide
down, on the second bullet there, how does one
determine whether the value of some parameter is
tolerable or not tolerable?
MR. SULLIVAN: The basic concept there is
that we have in place concepts -- and as Jim said, in
NEI 97-06, of being implemented -- related to specific
performance criteria.
For example, the structural integrity
performance criteria is that there should be a factor
of safety of three times normal operating pressure
against burst, and 1.4 times main steam line break
pressure.
In terms of measuring, the basic concept
is that you have a qualified NEI sizing technique, you
assess -- and with suitable uncertainties, you assess
the condition of the tubes against that criteria.
If you don't believe that you have a
sufficient understanding of NDE uncertainties, the
approach is to prioritize the tubes that are most
damaged by this degradation mechanism and do institute
testing against those factors of safety, and determine
whether or not the performance criteria are being
satisfied.
In terms of tolerable, the basic concept
there is to set the performance criteria such that
there is some leeway that if the performance criteria
aren't satisfied, you are not falling off a cliff in
terms of safety.
And in terms of leading to spontaneous
tube ruptures or being vulnerable to main steam line
break. Do you want to add to that?
MR. MURPHY: Yes, I can add to that. This
is Emmit Murphy from the Materials and Chemical
Engineer Branch of NRR. I might also add that when
considering appropriate performance criteria, we did
consider the available information on risk.
And we considered some of the findings in
NUREG 15-70 pertaining to risk, and which also
included an early look at tube rupture accident
sequences and their impact on risk.
And the conclusion based on the
information available at the time was that for plants
maintaining margins at the performance criteria that
were being proposed that there was not a significant
risk issue at that point.
So whether you were just slightly below
the performance criteria, or you were right at the
performance criteria, there is not going to be -- you
don't cross a critical risk threshold.
DR. SIEBER: Thank you.
MR. SULLIVAN: I think one of the major
elements of the NEI 97-06 generic change package is
the revision to the text spec that is being proposed,
and we have worked quite a bit with industry to sort
of get on the same page on this issue, and on this
part of the change package we are all in agreement on.
And that is that it would contain
basically three new elements that I have outlined on
this view graph. The first is to revise the existing
operational leakage tech spec downward from this
standard of 500 gpd, which is in the improved
standard, to 150 gpd, which a lot of plants already
have in their tech specs.
And then secondly there would be a new
limiting condition for operation, entitled, "Steam
Generator Tube Integrity," and that would have a
surveillance requirement to verify that the structural
integrity and accident leakage integrity performance
criteria are met in accordance with the steam
generator program.
And then a new administrative text spec
called "The Steam Generator Program," which I am going
to talk about on the next view graph. The new
administrative tech spec basically has four elements,
or maybe five, but over five different elements.
It starts out by saying that a steam
generator program shall be established and implemented
to ensure tube integrity and performance criteria are
maintained. It goes on to require that condition
monitoring assessments of the as found condition of
tubes be performed to verify that the tube performance
criteria that I mentioned previously, the structural
integrity and the accident leakage integrity
performance criteria, are being maintained.
Then it goes on to say that licensees have
to use NRC approved performance criteria, even though
those performance criteria are located in the industry
steam generator program, they have to be ones that are
reviewed and approved by the NRC, either generically
or plant specifically.
And in a similar fashion, the tech spec
goes on to say that licensees can only use approved
tube repair criteria, and NRC approved repair methods,
whether they are again approved generically or plant
specifically.
And the last section of this tech spec
deals with tube inspection reports, and that is not on
the view graph, and that has to do with when reports
have to be submitted, and what triggers their
submission, and what they are to contain.
As I mentioned, the details of a steam
generator program would be located outside of the tech
specs. The tech specs basically say what I just went
through.
As Jim Riley indicated, licensees -- well,
actually this isn't what Jim indicated. This is
something different. As part of submitting the
generic change package, licensees will commit to
developing the steam generator program in accordance
with NEI 97-06 guidelines.
The difference here between this and what
Jim Riley said is that this is a commitment to us, as
opposed to an internal industry commitment. The top
tier of 97-06 guideline document provides general
guidance for a performance based programmatic strategy
for ensuring tube integrity.
And it includes the elements that I have
towards the bottom of the view graph. It includes
performance criteria, tube integrity assessment, in-
service inspection elements, tube repair limits and
repair methods, and leakage monitoring.
Not the details, but a description of
those elements of a program, and it is our intent to
review NEI 97-06 for endorsement as part of the NEI
97-06 generic change package.
CHAIRMAN FORD: And all of these, the sub-
bulleted performance criteria and in-service
inspection, the metrics for all of those come out of
the latest action plan that we have got, the
integrated NRR for such programs?
MR. SULLIVAN: No.
CHAIRMAN FORD: Where do the metrics come
forth? For instance, in the in-service inspection or
leak monitoring? Well, specific data and specific
numbers?
MR. SULLIVAN: The specific approaches are
in guideline documents that I am going to talk about
on the next page. In terms of inspection, for
example, since you mentioned that, there is a
guideline document that contains details on matters
such as what sort of degradation to look for, what
sort of probes to use.
CHAIRMAN FORD: All right.
MR. SULLIVAN: What type of qualifications
the inspectors need to have. In terms of limits,
limits are in the performance criteria that the
inspection program will develop the information to
apply through integrity assessments to determine
whether or not the performance criteria are being
satisfied.
Actual limits are in the guidelines with
respect to primary to secondary leakage monitoring and
the actions that need to be taken.
CHAIRMAN FORD: I understand.
MR. SULLIVAN: So I mentioned NEI 97-06 as
a top tier guideline, but here are subtiered
guidelines that are on this view graph, and I thought
I would give you a little bit of a flavor of the age
of those documents, because they do vary quite a bit.
The steam generator examination
guidelines, and examination being another word for
inspection, currently licensees are using Rev. 5,
which came out in 1997, and Rev. 6 is being developed.
And I am going to talk about Rev. 6 a couple of view
graphs hence.
I believe those guidelines first came out
in the '80s. They have been around quite a lot time.
The tube integrity assessment guideline is the most
recent, and I believe that came out in February of
2000. So that is only a little over six months old,
in terms of it actually being issued to licensees.
The in-situ pressure test guidelines has
been around about a year longer than that. The
guidelines for monitoring primary to secondary leakage
came out I believe in the early '90s. I think they
are up to Rev. 2 of that.
The water chemistry guidelines we believe
came out or first came out in the late 1970s. And the
EPRI sleeve and plug assessment guidelines have been
around for 4 or 5 years.
DR. BONACA: I have a question. Going
back to actually slide seven, when you talk about
performance criteria in '97 or '96, and this is more
for information, could you give me a feeling for what
is involved in that performance criteria?
Is it just simply the number of tubes, or
leakage, or is it also for example the prediction or
the ability to predict?
MR. SULLIVAN: There are three performance
criteria. The operational leaking is probably the
easiest because that already exists. The structural
integrity criterion says that no tube should have --
I don't know if this is literal in this, but this is
actually something that we need to discuss further
with NEI.
But the gist of it is that no tube should
have less than a margin of three against bursts, and
the margin of three is against normal operating
pressure, and 1.4 against main stream line break.
The accident leakage integrity criterion
is again something that you have to calculate, and the
idea of it is that under accident conditions the total
primary to secondary leakage under accident conditions
should not exceed one gallon per minute. Does that
answer your question?
DR. BONACA: Yes. I guess what I am
looking for is there some element that measures the
ability of the inspections to predict, for example,
the growth of the number of defects, as well as the
severity of the indications?
Is there anything, any element, that does
that in this program?
MR. SULLIVAN: Well, I think I can address
that, and if I can't, maybe Emmit can add to it. I am
trying to figure out where this comes up or whether I
have already covered it.
I think I already covered it when I talked
about in-situ, and talked about the administrative
tech spec requires that licensees perform condition
monitoring of as found condition of the tubes.
In a similar fashion, while it is not
embedded in the administrative tech spec itself, the
bases as it is currently written in draft in NEI 97-06
talks about the basic understanding that licensees
perform what is called operational assessments.
And I had talked about that previously in
the context of risk 2022, where licensees do
predictions through calculational techniques, which
would involve things like growth of degradation, to
determine how far out in time they can operate and
still maintain those safety margins.
DR. BONACA: Well, the reason that I am
asking the question is that to me that is an element
of performance that I don't measure in leaking, but I
have a statement on the part of the utility that
performs these inspections that says based on what we
do, we predicted that we will not have more than X-
number of additional tubes, nor more than this number
of severe laceration.
Now, if I get to the next cycle and I find
that these predictions are good, it gives me
confidence in the process. I could say that that is
a good performance element in their program if
conversely they come back and they are totally off,
and there is a much faster growth, and they cannot
predict, and I would expect that I would measure that
as an element of performance in their ability to
support programmatically the steam generators.
Do you see where I am going? I am trying
to understand how that --
MR. SULLIVAN: One of the reporting
requirements that I didn't mention is that when
licensees don't satisfy their performance criteria,
they have to report that to us on a pretty short
schedule. I am not sure exactly what the timing is.
And our intent if that were to occur would
be to devote additional resources over what we planned
to understand what is going on with that particular
plant, and to work with the licensees.
They may not express it exactly the same
way, but to work with the licensees to make sure that
we agree with what their plans are for the next
operating interval.
In the case of ANO-2, we had observed that
they didn't satisfy performance criteria on a number
of occasions going back as far as, I think, 1992. And
ANO-2 had been on several occasions between then and
when they replaced their steam generators last
October, I believe, had done a number of mid-cycle
inspections.
They had planned to only do one mid-cycle
inspection in their last operating interval, and
basically because of disagreements that we had with
the licensee, they agreed to do two mid-cycle
inspections.
So it is not formalized in terms of some
sort of performance indicator or performance monitor,
but it is where we devote our resources when we
observe that licensees are having problems.
DR. BONACA: I still feel that performance
criteria here focuses -- or I thought, focused
specifically on the performance of the steam
generator. I think that I would like to look at
elements of the steam generator program, and among
those there is also this ability of predicting the
future leakage and somewhere they must be, and I am
sure that NEI --
DR. SHACK: But you do that, right,
because he has to do the performance assessment which
sort of predicts where he is going to be. And then he
does the condition monitoring to find out how well his
prediction worked.
I was curious that when he misses that
prediction, there is a discussion of why he missed it,
and the result is a change in his assessment
procedures, or the mid-cycle inspection, or that is a
kind of an ad hoc thing that you go through when the
two don't agree?
MR. SULLIVAN: Right. I mean, one way to
put it is that we don't typically review operational
assessments. That's not something that we do in
detail, particularly in headquarters.
But if there is a missed performance
criterion, we would at least review elements of the
operational assessment, and maybe not take it under
formal review, but in the sense that we would want to
approve it.
But we would probably ask that it be
submitted, and we would ask the licensees to give us
briefings on what their understanding is of why they
missed it, and what their corrective actions are.
DR. BONACA: It seems to me that if you
really miss it -- I mean, what you are trying to do in
this performance is to predict if you really meet in
fact this criterion leakage, and accident leakage, and
so on and so forth, all through the period of
operation that they are allowed to go before
inspection.
And if your predictive models are
incorrect, then you are violating this criterion by
definition, simply because they have no basis and no
foundation.
So there has to be some -- and you are
right. The real problem or has to be a fundamental
element of performance, I think.
MR. RILEY: This is Jim Riley again of
NEI. Let me see if I can explain how the whole
process fits together. There is really three
assessments associated with the steam generator
inspection.
The first is called the degradation
assessment, and that is done prior to the inspection.
And the utility takes a look at what has transpired in
their steam generator to this point, and evaluates
what kinds of degradation they have going on, and
where it is going on, and they plan their inspection.
They figure what they are going to see,
and they plan what probes they are going to use, and
what places in the steam generator they are going to
look, et cetera.
And that's all based on previous history
and anticipated degradation. They then do their
condition monitoring, which is the actual inspection
of the steam generator. They look at what they
actually have in place.
If they find in their condition monitoring
that things are going on that they did not predict in
their degradation assessment, they revisit the
degradation assessment during the inspection to see
does this affect my inspection plans, and do I need to
look in new places, and do I need to use different
kinds of probes, and what do I need to do to account
for this.
When they finish their condition
monitoring, the last thing they do is an operational
assessment, and that is a prediction forward. If they
look at what they have got, and what growth they
experience, and they predict as Ted indicated how far
can I operate and still be able to ensure that I will
meet my performance criteria when I next shut down and
inspect.
And that process repeats itself the next
time they shut down and do a degradation assessment.
So there is a feedback mechanism that makes sure that
they are accounting for what they are seeing with
respect to what they are predicting, and influencing
their inspection program accordingly.
MR. SULLIVAN: Okay. I am going to kind
of shift focus in a sense for the rest of the
presentation and start to try to give you some
insights into what is currently going on with NEI
97-06 and some of the problems that we have been
encountering.
At the time that we made the transition
from the generic letter and fully understood where we
were going with respect to setting up a regulatory
framework that was based on an industry initiative, it
had not been our intent to review and endorse the
subtier guidelines that I put up a couple of view
graphs ago, the detailed subtier guidelines.
Based on the guidelines that were
available at that time, we expected significant
enhancements to industry efforts to ensure tube
integrity under this program.
The staff's expectation was that the
guidelines would be sufficiently well developed to
lead to improved tube integrity performance under the
new framework, bearing in mind that we didn't have all
the guidelines. They had not all been issued at that
time.
And we had expected, and continue to
expect, that the guidelines will evolve over time in
response to technology changes, lessons learned from
operating experience, and results from various
studies.
The staff developed a couple of concerns
more recently though, and in just this past year, and
I will try to lay out without getting into too much
gory detail how they came about.
The first one is related to an action plan
item having to do with conducting a steam generator
workshop, which we did in February of this year. And
in that workshop some of the industry representatives
discussed draft revisions to the EPRI steam generator
examination guidelines, Rev. 6 basically, to permit
inspection intervals for steam generators with
improved materials, which we didn't have an issue with
in particular.
But we noticed that at least that draft
has since been revised substantially, but the draft
had inspection intervals that would go significantly
beyond Rev. 5, as well as what is in the tech specs.
And bear in mind if this has not been
clear that the approach under the new frame work would
be to lift the maximum intervals between inspections
that is in the tech specs, and rely on the performance
based strategy instead.
In one scenario, as I have on that second
bullet, it would have permitted inspection intervals
ranging to 22 full power months. I am not trying to
put that there as characterizing the proposals. I
want to put out kind of one of the extremes that was
in that proposal, at least that we considered an
extreme.
We also began to have concerns about
condition monitoring being implemented, and these grew
out of questions that we were asking licensees in our
outage phone calls about their bases for performing
in-situ testing of tubes.
We had some concerns that at least in our
view that in-situ testing wasn't being performed as
routinely or under situations that we think they
should have been performed, at least in some cases.
And I am not saying that they weren't
being performed. Lot of utilities did institute tests
last outage, but there were some plants that generated
some concern in our minds who weren't performing any.
These concerns basically could be
characterized as concerns whether or not the tube
integrity performance criteria would continue to be
met, and whether conditions not meeting the
performance criteria would be detected.
DR. SHACK: What control do you have when
they do a tube test that they pick the worst tube? I
mean, I can always pass it by picking the right tube
to test.
MR. SULLIVAN: Right. Well, if the key
work is control, we don't have any. But we have I
think some influence. Usually when it is evident --
well, first of all, we only pick the licensees for
phone calls that we think have the most degradation,
or that we are particularly curious about.
For example, we are going to have a phone
call with Turkey Point-3 this season. They have got
improved materials, but they have been operating for
quite a long time.
We go over the results, and licensees
generally characterize their worst tubes, and that
gives us a sense for whether we agree or want to
discuss further the in-situ testing that they are
going to do.
They also frequently provide us with lists
of any current measurements, bearing in mind that
there is uncertainty, but they give us those
measurements in tables that they are using themselves,
and they tell us which tubes they are going to test.
We have had occasion, and one that comes
clearly to mind --
DR. SHACK: But you see that list before
they do the tests?
MR. SULLIVAN: Yes, generally before.
Does that answer your question or should I elaborate?
DR. SHACK: That answers my question.
MR. SULLIVAN: We have had some influence
in the past. And in the case of ANO-2, for example,
in the '98 or '99 time frame, there were four tubes
that we questioned why they weren't going to test.
They indicated that they thought they were
unbrellaed by previous tests. We had given the
uncertainties and we didn't agree with that. They
subsequently ended up testing all four tubes, and
discovered that one of them was at least questionable,
or inclusive, regarding whether or not they could
conclude that they had satisfied the performance
criteria.
Okay. What I wanted to say that is that
out of the latter concerns having to do with the in-
situ tube testing, we took on kind of an initiative if
you will to spend more time studying those portions of
the EPRI guidelines dealing with condition monitoring,
and generated a number of concerns.
Those concerns are developed in a letter
that we sent to NEI. They knew that it was coming.
We had had some discussions with them, and it is dated
August 2nd.
My understanding is that you were provided
with that many sometime last week. I'm sorry that we
didn't get that to you sooner. The issues relate to
industry practice that exist under the current
regulatory framework.
But these are not brand new issues. They
are concerns that we have recently generated in our
own minds. But they are existing -- they would exist
under the new framework, assuming that we were to go
forward with the new framework, which is our intent.
These are not issues that we think we can
settle in a real court time frame, and that's why I am
talking about it in this kind of context. We don't
think that the existence of these issues, particularly
given the remarks that Emmit just last made, would
reduce assurance of tube integrity or increase risk
under a new framework, assuming that the inspection
intervals don't increase relative to the current
requirements.
And I am going to get into that in a
little bit, as that is not quite as hard and fast as
that may make it sound. In kind of a parallel
fashion, at least in terms of the bottom line of this
view graph, we have reviewed most of the industry
responses to issues identified for the industry in the
NRC IP-2 lessons learned study.
I am sure that you have glanced at that at
least and noticed that there were quite a number of
issues in there for industry, as well as for the NRC
staff.
And likewise for the review that we have
done, we included some write-ups on those industry
sponsors in that same letter that I just mentioned of
August 2nd.
These issues primarily relate to EPRI
guidelines and some of the issues overlap what I have
been discussing in terms of condition monitoring and
inspection intervals. But some of them go beyond
that.
A number of those issues still remain
unresolved, including the issues that extend beyond
condition monitoring and inspection intervals. But
likewise, those issues exist under the current
framework and will likely continue to exist under a
new framework.
And we don't think that the existence of
those issues reduce assurance of tube integrity or
increase risk under a new framework. Again, assuming
inspection intervals don't increase relative to
current requirements.
And again I will repeat that is pretty
hard and fast, and I am going to explain that a little
bit more in the last two view graphs. So, in terms of
conclusions, pending resolution of these guideline
issues, the staff has concluded preliminarily that it
can proceed with review and approval of a generic
change package provided that there are licensing
restrictions on inspection intervals.
And what I mean by that is that we would
have in mind that the generic change package
incorporate agreements with industry on appropriate
prescriptive intervals for inspections that would be
tailored to the specific material in the tubing, Mill
Annealed 600 thermally treated and Inconel 690
thermally treated.
And then the idea behind the words
licensing restrictions would be that changes to those
agreements would be likewise to performance criteria
and repair methods, either generically or plant
specifically, they would need to be approved by the
NRC. That is the proposal that we are working on with
industry right now.
DR. POWERS: I got a little confused. You
said Mill Annealed, and then you said thermally
treated. Did you mean just thermally treated?
MR. SULLIVAN: No, I meant three different
materials. I'm sorry.
DR. POWERS: Oh, so three different
things.
MR. SULLIVAN: The Mill Annealed 600,
thermally treated 600, and the thermally treated 690.
With this approach, we believe that the generic
package -- I'm sorry. The generic change package
would reduce the assurance of tube integrity only in
cases where longer inspection intervals than currently
permitted would be implemented without adequate
justification.
That is just another way to say what I
have just been saying. I think the rest of this,
except for the last bullet, is kind of repeating what
I just said. I wanted to go on to a different concept
to kind of tie a little bit of this together.
And to note that on the last bullet that
we are working with industry to establish a protocol
agreement resolving outstanding technical issues. It
would formalize an approach for interactions between
NRC and industry when resolving technical issues that
exist and that will continue to arise.
This is not just something to settle NEI
97-06, but it would be a long term protocol. Examples
of the types of issues that we currently would deal
with under that protocol would be the lessons learned
issues, and the condition monitoring issues that we
have been talking about, the risk 2022 issues, and
that sort of think, and any new issues that might come
up over time.
DR. POWERS: Could I go back to the next
to the last bullet.
MR. MURPHY: Yes.
DR. POWERS: And you say you were
exploring alternatives with the industry, particularly
for improved tube materials.
MR. SULLIVAN: I think what we mean there
is that the proposal that we most recently have been
discussing with industry would require that -- and
correct me if I am wrong, Jim, but the Mill Annealed
600 tubing plants would basically have to inspect
every refueling outage.
And longer intervals that follow a more
elaborate scheme, depending in part on what the
material is, and how long the plant has been
operation, would have maximum intervals longer than
that, up to three intervals between inspections, or
three outages or three cycles of inspections.
DR. POWERS: This is what I am struggling
with, is that -- well, it is very simple. People say
690 is a better material. As far as I can tell, that
is what they thought about 600, too.
I mean, do we have any confidence that
this material is really that much better, and that it
is not going to start cracking?
MR. SULLIVAN: Well, I think that there is
a lot of evidence in this country that 690, which has
been in plants for close to 10 years, is performing
much better than the Mill Annealed 600. But I am not
sure if that is what you are driving at though.
DR. POWERS: Well, what I am going to say
is that 10 years ago we probably could have said the
same thing about 600. Well, maybe not. Maybe it had
to be 20 years ago. But at some time we would have
said that.
MR. SULLIVAN: We would have said that at
the outset, but the Mill Annealed 600 tubing started
performing badly from the very outset. I mean, plants
were in their first inspection and performing their
inspections after the first -- maybe you can elaborate
on this more, Emmit. You were there at the time.
MR. MURPHY: Well, in fact -- this is
Emmit Murphy again. In fact, plants developed leaks
during the first operating cycle of operation just as
an illustration of how quickly the problems developed.
DR. BONACA: Well, that was much to do
with chemistry.
MR. MURPHY: Well, I can think of one case
where the crack involved was primary water cracking
that occurred in the first operating cycle.
DR. POWERS: I guess what I am driving at
is how does one go about arguing that 690 allows you
to go three operating cycles between inspections?
Now, it seems to me that if you can say, well, it has
operated for 10 years, and no problems. That's a
pretty good argument for longer cycles.
I mean, if it is that empirically based,
then it is pretty inarguable. The trouble that I see
is the potential for it just suddenly starts leaking
because of this long induction period it takes for
cracks to suddenly show up on the detection device.
MR. MURPHY: This Emmit Murphy again, and
I think we shared that concern, and I think that some
of the operations that we are exploring with the
industry here that would provide opportunities for
materials, for plants with the newer tubing material
to implement longer inspection intervals.
And that these prescriptive limits on
cycle length would give us the level of assurance
maintaining the tube integrity margins set that we
have historically enjoyed, and certainly can do better
than that hopefully by virtue of the expected and
improved performance of these new materials.
MR. SULLIVAN: Another thing that I might
add is that the plants -- you know, this is a little
bit of an elaborate strategy, and we have not tried to
get into particulars here.
But I think if you take some of the plants
with Inconel 690 that have been operating the longest,
the current proposal wouldn't allow them to go three
cycles. The current proposal would allow them to go
two cycles, which is basically what the current text
specs already allows.
So it would only be -- I mean, the basic
idea is that the licensees do a pre-service inspection
at the first refueling outage, and they would have to
do another inspection to monitor for things that --
you know, like wear.
That in loose parts, you can't just say,
well, that is not going to happen. And then they
would move on to a strategy of thee cycles. I think
that it factors that in, as well as being based on
some of the empirical observations that we have had.
MR. RILEY: If I could say something
again. This is Jim Riley again from NEI. Another
consideration that we have put into our guidelines
again is this degradation assessment that I mentioned
the last time.
The plants, even though they wouldn't have
to inspect every outage under our scheme, would be
required to do a degradation assessment every outage,
and that degradation assessment needs to take a look
at what has been happening at their plant, as well as
what has been happening in other plants around the
industry and around the world.
And if there are things going on in these
other plants with Inconel 690 that wasn't anticipated,
that has to be taken into account and it has to be
taken into account from the perspective of how well it
pertains to their design steam generators, their
materials, their chemistry, et cetera.
But if they feel that this is challenging
what otherwise would have been their inspection
interval, they need to be reacting accordingly.
DR. POWERS: It is an encouraging thought,
but what is discouraging is when I look at the
assessments under the maintenance rule, one of the
areas that the licensees found most challenging was
the ability to take into account experience within the
industry, and not at their own facility.
So, pardon me, but I would be just a
little skeptical that they will -- that in the
assessment that they won't be looking for ways to
argue what is going on some place else just doesn't
relate to my plant.
MR. RILEY: That's difficult to argue. I
mean, obviously it depends on an individual plant, but
I will say this. That there are plenty of information
available to the licensees, in terms of what is going
on elsewhere.
We have an industry organization that
meets three times a year and shares operating
experience. We have a steam generator degradation and
steam generator database that EPRI maintains that
keeps track of what is going on at different places,
in terms of tube degradation, and tube pulls, and tube
information, et cetera.
We have organizations within the industry
that do reviews of steam generator programs at various
-- well, they rotate through all the plants, and do an
evaluation of how well they are conducting their
program with respect to what the requirements are in
NEI 97-06 and other places.
And we have internal peer reviews that are
done between organizations, and all these things are
intending to look at how a particular utility is
conducting its steam generator program with respect to
the norm and the expectations.
And sharing with plant management cases
where they feel that they are not meeting the industry
standards on these issues.
MR. SULLIVAN: One thing that I might add
for what it is worth is that over the years when a new
degradation mechanism is identified, or not
necessarily a mechanism, but a new location, and we
learned about it in a phone call.
And we might be on the phone call at the
same time, or the next day, or whatever, with a
similar plant, and we would bring it up in the phone
call, and I can't remember a single time that the
licensees weren't already aware of it.
And I think as Jim indicated, the
networking is pretty strong, and had modified their
inspection plans to look for it if it was applicable,
just in the "for what's it is worth department."
CHAIRMAN FORD: Can I just ask a question,
more on a technical management aspect? Do I
understand that right this instant, in terms of
monitoring the performance of the steam generator
tubing, that we are essentially using NEI 97-06
procedures, regardless of how they stand within the
regulatory framework right now?
And that in very short order that you are
going with this generic change package, which is based
on NEI 97-06, but with modifications associated with
its memo that you sent out on the 2nd of August?
And that would give some regulatory aspect
to approval if you like. It may not have gone through
all the sign-offs, et cetera, et cetera, that you may
have to do. But essentially you have got regulatory
approval for the NEI 97-06 procedures, et cetera, and
that is in the short term.
MR. SULLIVAN: Yes.
CHAIRMAN FORD: And for the longer term,
as we go through the question of brisk assessment of
the delta-LOCA and the delta-LERFs, and modifications
to your current understanding of those parameters, and
that will come out in later years as a result of this
joint NRR research program. How I got the sequence of
events right?
MR. SULLIVAN: I think that's correct, and
then depending on what comes out of that, we may have
to factor it back into our understanding, and/or our
regulation of the steam generator programs.
CHAIRMAN FORD: Now, is it appropriate
therefore in the short term, if you are going to have
this model one of this generic change package in
place, is it appropriate to have a presentation to
this subcommittee -- and let's say in December -- so
that we understand at least the technical pros and
cons of this process?
MR. SULLIVAN: I think it is a good idea.
CHAIRMAN FORD: And I stress the technical
aspects. For instance, what the pre-inspection
assessment methodology is, and what is the
uncertainties in it, et cetera, so that we understand
the impacts on safety.
MR. SULLIVAN: I think coming back for
another presentation is a good idea. The only thing
that comes to mind is that we are also making a
presentation to the Commission on December 4th. So we
want to make sure that we don't have a conflict there.
CHAIRMAN FORD: I have no idea what the
constraints of this particular aspect is --
MR. SULLIVAN: As a concept, I think it is
a good idea, and we did anticipate that you want more
technical details than what we are talking about
today.
CHAIRMAN FORD: Okay.
MR. SULLIVAN: This is just kind of an
introduction.
DR. DUDLEY: Just thoughts. Would it be
more appropriate for an ACRS presentation before or
after the presentation made to the Commission?
MR. SULLIVAN: Can I get back to you on
that later? I would like to talk to my colleagues.
DR. DUDLEY: Yes, that is something that
you need to work out.
CHAIRMAN FORD: But this is a joint
NEI/NRR?
MR. SULLIVAN: Yes. Sure. We will have
to coordinate with Jim, of course. I can't speak for
them.
CHAIRMAN FORD: Excellent.
MR. SULLIVAN: But they have been willing
in the past to come and make presentations like this.
MR. RILEY: Jim Riley again. We would be
happy to join your presentation on the technical
aspects of the program.
MR. SULLIVAN: Okay. I just have a couple
of comments. I have kind of covered this, but I just
wanted to make sure that it is clear that we do plan
to develop a safety evaluation on this whole generic
change package.
The vehicle for issuing it would be a
regulatory issue summary, and the proposal would be to
put it out for public comment before we finalize it.
There are some specific reasons that we want to do
that that we can get into now or in the next
presentation.
Our target date had been the end of next
month, and we clearly see that we are not going to
make that. We are hoping that we can get this done in
April of 2000, although I have to admit that was kind
of an arbitrary projection that we could get it done
within about six months.
We are still working with NEI on technical
issues, as well as the regulatory issue having to do
with regulatory controls. And so I am not sure just
how optimistic or achievable the April date is.
And as I mentioned before, this same sort
of data is contingent on coming to terms with this in
the pretty near term, because there are a lot of steps
that we need to go through, in terms of things like
issuing a risk for public comment, and finishing the
safety evaluation, and so forth. So that concludes my
presentation.
CHAIRMAN FORD: Thank you very much. I
would like to put this on hold for 15 minutes, and I'm
sure that on hold isn't the right word, but we will
take a tea break.
(Whereupon, the meeting was recessed at
10:08 p.m. and resumed at 10:25 p.m.)
CHAIRMAN FORD: Okay. We are back in
session, and we are reversing the order. Ken is going
first, and Joe is coming second. So, Ken will be
talking about the South Texas project.
MR. KARWOSKI: I am going to stand during
this, just because I need to point to some of the
stuff on the view graph. My name is Ken Karwoski, and
I am with the Materials and Chemical Engineering
Branch in NRR.
My presentation is broken into two parts.
The first part will be the overview of the South Texas
steam generator operating experience, and the second
part will get into the last part of the presentation,
which is some of the issues on the -- with respect to
the differing professional opinion.
So the slides are in the opposite order
that I had anticipated. So we will skip the first two
slides, and I will come back to those at the end of
the presentation, and I will start with South Texas.
South Texas is a four loop pressurized
water reactor. It has a model E-2 steam generators
and there is about 4,900 tubes in each of those steam
generators. They have Alloy 600 mill annealed tubing,
with the exception that there is 15 tubes in one of
the steam generators that is made of Alloy 600
thermally treated.
They did that, I believe, to test for
whether or not this material would be any better.
They have three-quarter inch diameter tubes, which is
important for generic letter 95-05. The tubes are
supported at various elevation by drilled holes
stainless steel tube support plates.
That is a little different than most of
the mill annealed plants. Actually, it is the only
plant in the country that has drilled hole stainless
steel tube support plates.
The bulk of the plants that use generic
letter 95-05 have carbon steel drill holes tube
support plates, and I will talk a little bit about
that later on.
DR. POWERS: What is the potential
difference between the stainless steel and the Alloy
600, the electrical-chemical potential differences?
DR. SHACK: There's not much.
DR. POWERS: But just about everything is
though.
DR. SHACK: Well, it is certain less than
carbon steel.
MR. KARWOSKI: But the key with the
stainless steel, which I will get into, is that it is
less corrosion resistant in a steam generator
environment. So what you have with the carbon steel
tube support plates is those tend to corrode and tend
to fill the crevice with magnetite, which tends to
impact the tubes, and actually cause corrosion-induced
bending.
The stainless steels are less susceptible
to corrosion in the steam generator environment, and
you don't get that type of corrosion product build up
in the crevice which could restrict leakage and can
bend the tubes.
There have been other plants in the
nuclear industry, particularly Doel 4 and Tihange 3,
which ave these types of tube support plates. The
steam generators at those plants have been replaced.
At South Texas the tubes have been
hydraulically expanded into the tube sheet, and the
expansion transitions were shortened to reduce
susceptibility of corrosion.
R-1 and 2 of the steam generators went
through a U-bent heat treatment to also reduce the
suspectibility of corrosion of the R-1 and 2 U-bends.
South Texas, coming on line later, implemented several
enhancements to their steam generators in order to
reduce the susceptibility of the tubes to --
DR. SHACK: Well, it is awful late for a
Mill Annealed plant though?
MR. KARWOSKI: I think they started
commercial operation in like '89, but when they
ordered their steam generators and when they planned
that, I don't have that information.
But, yes, in the overall sequence of
events, if you look at some of the earlier
replacements, they were thermally treated in the early
'80s. And so I am speculating that they must have
ordered them.
DR. SHACK: They must have decided that
they didn't need to do that.
MR. KARWOSKI: Yes. Of particular
interest here is that in their pre-heater area, they
expanded several tubes as a result of a concern of
tube wear that had been observed in Westinghouse Model
D steam generators, and I would just point this out
there because they have observed some corrosion there
or some damage at that location.
And that is because of the cross-flow of
velocity of the feed water entering the steam
generator. South Texas has a T-hot of approximately
625 degrees fahrenheit, and that is one of the higher
ones in the country, which just exacerbates some of
the corrosion problems that they may be observing.
At the end of Cycle 8, which was in March
of 2001, they had approximately nine effective full
power years on their steam generators, which is not a
lot of time.
The primary degradation mechanism is
actually oriented outside diameter stress corrosion
cracking at the tube support plates, the focus of
Generic Letter 95-05.
I just briefly want to discuss some of the
other degradation mechanisms that they have been
observing. They have detected some free span outside
diameter stress corrosion cracking, primarily
associated with dings.
I use the term "dings" because instead of
corrosion-induced denting, it is more damage as a
result of fabrication.
DR. POWERS: What is the gap width for
this drill hole plate in the tube wall roughly?
MR. KARWOSKI: I think the exact value is
proprietary, but it is on the order of less than a
tenth of an inch for the normal support plates. They
have a flow distribution baffle, which I think is on
the order of a tenth of an inch, which has an enlarged
tube hole opening.
And that is the first support plate
elevation, and in general they have not observed as
much degradation at that location than they have at
the higher locations, where the diametrical clearance
is less.
DR. POWERS: What I am trying to
understand is that because we don't have this included
hole in the plate are we getting what would be crevice
type chemistry changes in there, in that hole region?
MR. KARWOSKI: Can I answer that a little
later on?
DR. POWERS: Sure.
MR. KARWOSKI: But that is one of the
theories that might be happening with respect to the
operational leakage. But I will touch upon that later
on.
So they have had free spanaxial outside
diameter stress corrosion cracking, and they have also
detected some free span volumetric indications, and
they have detected some of these over the course of
the last cycle or the cycles prior to that.
CHAIRMAN FORD: I wonder if you could just
mention -- and maybe you will mention it later on, but
the question of the difference between the stainless
steel and the carbon steel floor plates, the fact that
there is generally less corrosion product, and
therefore that would have an impact on leak rates.
MR. KARWOSKI: I will get to that in
probably 3 or 4 more slides.
CHAIRMAN FORD: Okay. Good.
MR. KARWOSKI: New mechanisms that they
observed during the March 2001 outage, they detected
some indications that the hot leg expansion
transition, that's not unusual for a plant with Alloy
600 Mill Annealed.
The indications were primarily OD. They
did find some ID indications of one ID indication.
The licensee speculates that the shop cleaning may
have been effective in reducing some of the ID
cracking.
Some of the dings in their steam generator
are basically separated by about three-quarters of an
inch, which is the thickness of the tube support
plate.
They believe that as they inserted the
tubes into the steam generator that there was some
bending moment that caused what they called paired
dings. At one of those paired dings, they observed
circumvential cracking at one location and axial
cracking at the other.
They found a Row-1 new bend indication,
which was outside diameter stress corrosion cracking.
They also found cracking at the U-Bend transition, and
they found a volumetric indication at the expansion
transition of one of those tubes expanded in the pre-
heater.
Most of these degradation mechanisms are
common among plants with 600 Mill Annealed tubing.
The licensee has currently plugged about 9 percent of
the tubes. Their licensing basis limit, I believe, is
10 percent. The are scheduled to replace in December
of 2002 at the end of the present cycle.
DR. SHACK: Do they sleeve or do they just
plug every one?
MR. KARWOSKI: I think they just plug.
DR. SHACK: With respect to the voltage
based repair criteria, I did mention that that is
their primary degradation mechanism, and they first
implemented Generic Letter 95-05 during Cycle 7, which
was in the '98-'99 time frame.
They were approved for a one-volt repair
criteria at that time. As a result of that amendment,
they analyzed for 15.4 gallons per minute primary to
secondary leakage during a steam line break to
demonstrate that the off-site builds consequences
where acceptable.
And during this review that the staff
approved that limit. Cycle 8, the licensee also
implemented the one volt repair criteria, and in Cycle
9, which is the cycle that they are presently
operating in, they recommended a 3 volt repair
criteria.
That repair criteria had been used at
Braidwood and Bryon, and evasively what it involves is
demonstrating that the motion for the tube support
plants is limited such that the degradation at the
support plate will not be exposed during a steam line
break.
And which allows them to go to a larger
voltage limit because the probability of burst will be
less.
DR. SHACK: Now, I noticed that South
Texas gets the benefit from IRB technology, as well as
the three volt limit. Did Braidwood and Bryon get the
IRB technology, or did they just live with 00 votes.
MR. KARWOSKI: By the IRB, the indications
are that that methodology, although the value of what
we assigned to those --
DR. SHACK: The probabilities, the 10 to
the minus 5?
MR. KARWOSKI: Right. Both South Texas
and Braidwood, and Bryon had to model URDs in their
methodology to account for the potential that a tube
attempts to burst, but can't because of the presence
of the plate, and therefore the leakage could be
higher. Braidwood and Bryon had to model that and
South Texas also.
DR. SHACK: They got to use 10 to the
minus 5th, first, and then two, as well as the three
volts? When we say the three vote criterion, I never
realized that you got a double-benefit.
MR. KARWOSKI: Okay. Let me take a step
back. When you implement this methodology,
essentially by locking the support plate in place, you
have essentially -- for an axial crack, you basically
prevented it from fully opening or fully achieving
burst because of the diametrical clearances.
Because of that the probability of an
axial rupture, that could be on the order of 10 to the
minus 5th. I don't recall what the actual number is,
but they basically modeled what the probability is for
a burst given the amount of displacement of the plate.
In addition, they have a correlation which
they say, okay, now that I can potentially go to
higher limits, what is the probability that I tear
this tube and get a circumvential break?
And that's how they would -- they would
generate a limit for that. The industry would claim
that that limit, that you could tolerate 10 volt
indications, and the staff said 3 volts based on that
correlation.
And so they also modeled the probability
that you would get a circumvential failure of the tube
at the location. So there is two parts of that
methodology.
Now, the URDs, that is basically a leakage
model aspect, and basically in the leaking
correlation, basically they don't have indications
which try to burst and actually leak excessively.
So as part of the three volt amendment,
Braidwood and Bryon embarked on a testing program to
figure out, okay, how that I have got these higher
voltage limits, if this tube starts to open up how
much will it leak given that the plate is there.
And that is what the URDs do, is that it
is another leakage correlation that is tacked on above
the normal free span leak rate correlation. So in
Cycle 9, basically in February or March of this year,
we approved this 3 volt criteria, and the licensee
expanded tubes at tube support plates 2, 3, and 4 in
order to limit the motion.
They only chose these lowe support plates
because that is where most of the degradation is
occurring. And I will talk a little bit more of how
they actually implemented that repair criteria during
this last outage.
During their past cycle, Cycle 8, prior to
implementing this 3 volt repair criteria, the licensee
was observing primary to secondary leakage in all four
steam generators, for a total of about --
DR. SHACK: Excuse me, but can I just --
this IRB is confusing me again, because as I read this
thing, when they do what I thought was a 95-05
methodology, which ignores the restricting from
bursting, they exceed the 10 to the minus 2
probability of failure.
Then they have to go to the IRB thing, and
that gets them down to 1 times 10 to the minus 3. So
it is not an additive thing. They don't use the 3
volt criteria for the plates that are locked; is that
the way that I am interpreting this?
MR. KARWOSKI: For the plates that are
locked, basically they say how far will the plates
move, or could they potentially move, and if I were to
expose a crack of that length throughout that plate,
and for all the plates in the steam generator which
have applied that criteria, what is the probability of
burst of that axial crack.
DR. SHACK: Okay. So that is saying that
we understand the movement of this plate well enough
that 10 to the minus 5th is the product of the
probability that the tube will burst without the plate
times the probability that it will be uncovered,
right?
MR. KARWOSKI: It is more of just the
materials issue. It is just that you have to
understand how much the plate is going to move. So
that aspect is in there.
You have to know how much of the crack
will be exposed or could potentially be exposed,
because we are postulating that the crack is at the
tip of the support plate, and as the support plate
moves it exposes the entire flaw over that length.
DR. SHACK: But they had to calculate that
probability somehow from their fluid mechanics
calculation.
DR. KRESS: They just assumed it happened.
MR. KARWOSKI: But they assume all -- they
calculate the maximum displacement of the plates.
DR. KRESS: And then they assume it
occurs.
MR. KARWOSKI: And then they assume it
occurs over the entire plate, and so basically they
are saying, okay, I have exposed -- I think in their
case they postulated that -- or they determined that
it would meet something on the order of .15 inches.
And so they said .15 inches for every tube
at that plate. They didn't say that the plate is
going to move .15 inches here, and .12 inches here,
and .02 inches here. They just assumed that the
maximum displacement for every intersection.
DR. KRESS: How did they make that
determination? Do you know?
MR. KARWOSKI: The determination of how
much it would move?
DR. KRESS: Yes.
MR. KARWOSKI: That is by thermal
hydraulic modeling.
DR. KRESS: So you don't have a
probability associated with that then?
MR. KARWOSKI: There is no probability
associated with that.
DR. KRESS: So the probability of the
materials isn't --
MR. KARWOSKI: Right.
DR. SHACK: So what you are saying then is
that if I uncover a tenth of an inch, say, I can
somehow calculate then the probability that he burst
will be 10 to the minus 5?
MR. KARWOSKI: Yes. I think in general
they say less than 10 to the minus 5.
DR. SHACK: And how do I do that?
MR. KARWOSKI: Well, basically you have a
crack that extends outside the plate, and so the plate
is constraining the crack, the bolt of the crack.
Let's assume it is a three-quarter inch long crack for
simplicity.
And I move the plate .15 inches, and so I
have got 6/10ths of an inch crack within the plate,
and .15 inches outside.
DR. SHACK: Do I do this on a mechanistic
fracture mechanics basis rather than on a voltage
basis?
MR. KARWOSKI: Yes. Yes. Basically, how
much support does the plate give, and what the vendor
would argue is that the plate basically -- that the
length of the exposed crack is what is dominating the
probability of burst.
So basically you can say, well, what is
the probability of a .15 inch long flaw bursting. It
is based on mechanistic and it is not voltage. It is
not voltage.
CHAIRMAN FORD: So can I have just a time
sanity check here? We are required to have a letter
on the DPU issue at the next ACRS meeting. How long
do you think at this current rate of progress do you
think it will take? Can you be finished by 11
o'clock?
MR. KARWOSKI: Yes.
CHAIRMAN FORD: Provided that we don't ask
too many more questions.
MR. KARWOSKI: Right.
CHAIRMAN FORD: Okay.
MR. KARWOSKI: Okay. So they were
observing leakage in all four steam generators, and
when they came into the outage, they did a secondary
side pressure test, where they filled the secondary
side up with water, and pressurized it to something on
the order of 600 pounds.
And then they monitored for leakage on the
primary side of the tubes, and looked for drippage
from the tubes. What they found was that none of them
were leaking excessively, but there were some tubes
approximately that were damp.
The leakage was attributed to outside
diameter stress, corrosion, cracking, at the support
plates, and that is important because no other
domestic plant has ever observed operating leakage as
a result of cracking at the tube support plate
locations.
And that gets back to various theories of
why we haven't observed leakage, and one of the
theories is that as the carbon steel support plates
corrode, they form magnetite, and the magnetite gets
into the crevices and impinges -- well, impinges isn't
the word.
But it forms magnetite and the magnetite
fills the crevice, and it will start denting the tube
and basically or essentially would seal the crack.
That is one theory.
So that the crack tries to leak, and it is
not very porous, and it doesn't get out. That is one
of the theories that has happened. And the stainless
steel tube support plates situation in South Texas,
you don't have that magnetite filling the crevice, and
you have might scale on the outside of the tube, and
still have a crevice.
And so you are still observing the
corrosion, but in this case it is not impeding the
flow of the crack. That is a theory. As I mentioned
before, South Texas, too, is the only domestic plant
with stainless steel tube support plates, drilled hole
stainless steel support plates.
And Doel-4 and Tihange-3 had that. Doel-4
had exhibited leakage coming from the support plates
during a similar secondary side pressure test in the
early '90s.
Because of the concerns on operational
leakage, although the licensee was authorized to
implement a three vote repair criteria, they
preventively plugged down to approximately 1-1/2 volts
because of those concerns.
They did some depth-sizing of some of
these flaws to determine which ones that they thought
may have been most likely to leak, and they prevently
plugged those.
After the outage and these results became
available, the license submitted their 90 day report.
It is basically a summary of inspection activities
primarily related to Generic Letter 95-05.
The staff reviewed that report and we
identified several issues that we asked the licensee
to address. And the issues are on this view graph,
and I would just like to illustrate them.
One of them is the ability to predict end
of cycle conditions, which I believe was one of the
concerns raised earlier this morning. There are two
things that we look for during these reviews, and that
is the number of indications predicted, reasonable,
and is the severity, and in this case is the voltage
of the indications reasonable.
What this table shows is that it shows the
four steam generators and also the total, and it shows
the three cycles where they implemented the voltage-
based repair criteria.
For each one of these cycles, they show
the projected number of indications that they
determined, and then the actual. In this first cycle,
you will notice that they under-predicted the actual
in one of the four steam generators, but in general
they were conservative, with the exception of Steam
Generator C.
DR. POWERS: And before I leap to that
conclusion, I guess I would ask you how many
indications were in these steam generators that they
failed to detect?
MR. KARWOSKI: This actual number does not
include any account for the probability of detection.
So this number here and the assessments that they do
is basically assuming that you are finding the more
severe flaws.
And that the flaws that you are not
detecting are not of structural leakage significance
even now, and that they would not be of structural
leakage significance at this point. This number does
not account for that.
DR. POWERS: Okay. But if I take my
probability of detection at .6, and they then do it
for everything?
MR. KARWOSKI: Right. But this is more of
a condition monitoring assessment. This number here
would be -- would include the .6 from the prior cycle,
but yes, you are right. The value of .6, remember, is
to account for two things.
It is not only to account for indications
which we missed during the inspection, but also for
new indications which may develop or initiate over
that cycle. So to adjust these by .6 in a condition
monitoring system --
DR. POWERS: It is not quite fair, but to
adjust it by some number is fair.
MR. KARWOSKI: Yes, but what we would
argue is that what they missed is probably not --
DR. POWERS: I don't think you can do
that. I mean, I think you have a database that says
there are flaws of substantial size --
MR. KARWOSKI: That's true.
DR. POWERS: And you have a plant up in
New York where that is definitely true.
MR. KARWOSKI: That is true. That is
true. So this number does not include any -- it is
basically what they found in the steam generator
during that inspection, and it does not account for
any improbability of detection.
CHAIRMAN FORD: All those numbers, the
right or actual numbers, should be multiplied by 1.4
or whatever the number is?
DR. POWERS: I don't think it is quite
fair to do it that way.
MR. KARWOSKI: No, no.
DR. POWERS: As he pointed out the .6
counts for other things. But there is some number
that they should be multiplied by.
CHAIRMAN FORD: Correct.
MR. SULLIVAN: And that multiplication
factor is used in the projections forward.
MR. KARWOSKI: Right. So to arrive at
these projected numbers, what they did is they took
the actual, and divided by .6, and subtracted off the
number that they repaired, and that's how many they
got.
The purpose of this is just to show the
number of indications and the probability of
detection, and you need both the numbers and the
severity of the degradation.
DR. SHACK: So when we see these cases
where the actuals exceeded the projected that is
extremely distressing
MR. KARWOSKI: Let me phrase it this way.
In general, for Generic Letter 95-05, one of the
criticisms that the industry has always said is the
POD of .6 is excessively conservative, excessively
conservative.
So when you typically look at these 90 day
reports, you typically see numbers like that. In the
case of South Texas --
DR. POWERS: You see numbers like C.
MR. KARWOSKI: Right. And if you just
look at the total numbers, you start saying that
things are getting pretty close, and if you look at
the last cycle, they under-predicted the number of
indications in two of the four steam generators.
Now, that may not be bad in and of itself,
because if I am just finding a bunch of low voltage or
indications which have no structural or leakage
significance, that may not be a problem.
But this is just one piece of the puzzle.
Next, the next graph addresses the severity of the
indications, and basically it is a similar table to
the previous one.
It shows the steam generators, and as
voltage goes up the severity of the indication
increases and we compare it projected to actual. And
in general if you just look at the totals, in this
case they under-predicted the number of larger voltage
indications in the first cycle, but the number was
minimal.
The second time they also under-predicted
and the same thing for this third cycle. As a result
of this, we are pursuing discussions with the licensee
to ask them to address it.
And in the interest of time, this last
view graph just shows that the average growth rate,
that if you look at Cycles 6, 7 and 8, the growth rate
has been increasing the average growth rate, and that
pretty much is supported by the previous table.
There are some other issues that we have
asked the licensee to address regarding leakage
observations. During the inspections, they had done
some in-situ pressure testing, where they insert a
device inside the tube, and pressurize it to determine
whether or not it is going to leak and/or burst.
And they observed some leakage during
those tests, and given that the in-situ tests are
typically done on the worst tubes, from the
information that we were provided, it doesn't seem
like those results indicate or could account for all
the operational leakage that they observed.
And so we have asked them to take a look
at that. So basically the last view graph, here the
next step is that we post these issues to the
licensees, and they are monitoring for operational
leakage.
And there has not been any observed
presently and the licensee plans to replace their
steam generators at the end of the current cycle.
DR. BONACA: Well, you started to say
something about after you looked at the severity of
indication, because of this, we asked the licensee --
and then you didn't complete the phrase.
MR. KARWOSKI: We have asked the licensee
that in light of these results, basically tell us why
the methodology is working for your plant. What
confidence do you have that we will be able to
actually project what is going to be on this steam
generator at the end of the next cycle.
DR. BONACA: Well, it seems to me that
they are under-predicting both, in terms of severity.
MR. KARWOSKI: That's true, and in some
cases that may not be a concern. If I am calculating
leakage of a 10th of a gallon per minute during
accident conditions, and I have under-predicted the
number of severity, that may not be a problem in and
of itself.
But in this case, they are, and in one of
their generators they are projecting leakage which is
approaching that 15.4 gallon per minute.
DR. BONACA: Plus, there are a number of
indications that are going so fast and that is really
what we are transmitting. And at that point you begin
to wonder about when do you get to that point where
you have a critical change in the leakage, for
example.
MR. KARWOSKI: Right.
DR. SHACK: Now, when they do the
operational assessment what will they use for the
average growth rate? Will they project that
increasing curve, or will they use the observed --
MR. KARWOSKI: They will use the
methodology that is called for them to use, and the
most conservative over the last two cycles, which I am
assuming was the last cycles, and so they will use the
observed.
And the reason for showing you the tables
of the -- of what I will call the increase in growth
rate is that that is certainly one of the issues that
the staff would like addressed, which is, is the
methodology working.
And that is basically the reason or could
be a reason why they have under-predicted the severity
of some of those indications. At this point, I would
like to move to the second part of the presentation,
which basically addresses two of the ACRS'
recommendations on the differing professional opinion.
The two recommendations that I want to
discuss are the seven-eighths inch diameter leak rate
database, and the recommendation with respect to flaw
growth.
With respect to the seven-eighth inch
diameter leakage database, the ACRS indicated that the
database needs to be greatly improved to be useful,
and that the staff should consider requiring a near
term expansion of that database.
The staff agrees that the seven-eighth
inch database does not exhibit as strong a correlation
as the three-quarter inch. To refresh everybody's
memory the three-quarter inch database has
approximately 50 pull tubes, and about half of which
come from pull tubes.
The seven-eighths inch database on the
other hand only has approximately 30 data points, of
which only around 25 percent, or seven or eight data
points are from pull tubes.
So the staff agrees that this seven-
eighths database has a weaker correlation. With
respect to whether or not the expansion of the
database will actually improve the correlation, as
part of getting ready for this presentation, I tried
to do that assessment by looking that as they added
data over the course of several years, and what has
happened in general.
And based on a very simplistic evaluation,
which I did, it looks like the correlation is staying
the same, or maybe getting slightly worse. So even
though they added data, it has not necessarily made
the correlation better.
But the correlation in 95-05 does address
how to handle it if the correlation -- you know, if
there is a correlation or if there isn't any
correlations.
With respect to adding more tubes, the
staff recognized when they issued Generic Letter 95-05
that the limited data then -- and it is still
recognized as it is now, that the results as part of
the methodology that licensees committed to a tube
pull program, either the one that is in the generic
letter, or an industry developed the tube pull
program.
And with this protocol the utilities
periodically pull tubes, and the focus of those pulled
tubes is for seven-eighth inch diameter tubes is the
leakage database. They need more data and the
industry recognizes that.
With respect to -- with the exception of
this commitment, there is really no other regulatory
vehicle and the methodology to require removal of
additional tubes. But the staff will continue to
monitor the effects of additional data as more data is
added as a result of these tube pulls.
The next recommendation that I want to
talk about is flaw growth. The recommendation was
that the staff should establish a program to monitor
the predictions of flaw growth for systematic
deviations from expectations, and that the staff
should develop a database on predictions, and observe
voltage distributions.
As part of Generic Letter 95-05, we asked
the licensees to submit the data to the NRC to permit
putting together -- or to permit the staff to do these
comparisons of predicted and observed voltage.
And I think that the South Texas example
that I just went through is one of those cases where
we do look at that when we do those reviews to
determine whether or not there is something that we
need to follow up on.
So we have and we will continue to review
the 90 day reports with that recommendation in mind.
That was the reason for requesting that information to
be provided to the licensees.
We recognize that it is an empirical
approach and we need to continually assess how well we
are doing with respect to our predictions. The staff
is formalizing the review of inspection summary
reports, which the 90 day reports are a subset of, in
conjunction with the steam generator action plan, Item
1.10.
And there have been instances where the
predictions have been non-conservative, and South
Texas is one of them.
DR. POWERS: As part of this
formalization, you are going to explain how to use a
probability of detection to adjust the numbers that
are sent to you, right? I mean, you have got to deal
with the probability of detection issue don't you?
MR. KARWOSKI: Right.
DR. POWERS: Okay. One of the ways of
dealing with it is to say that I am not going to deal
with it, but I think that would be fairly
unsatisfactory.
MR. KARWOSKI: We can definitely look at
it as -- and whether or not it gets into formal review
or whether or not that is more detailed guidance --
DR. POWERS: Well, how do you handle it?
MR. KARWOSKI: Yes, we need to realize
that there are some indications which you can miss.
DR. POWERS: I think Westinghouse put
together a pretty nice story on what the probability
of detection is for what we needed in this context,
and which strictly is a probability of detection.
CHAIRMAN FORD: There is a question of
probability of detection, but the efforts that you are
doing in this area is combined in our own research,
and is in the 3.6 of the NUREG program. That's in
addition to this one isn't it?
MR. KARWOSKI: I am not -- with respect to
the database, the database is basically a regulatory
issue; whether or not research plans that I am doing
additional testing under these model boiler or
laboratory produced specimens that could supplement
the database, if they develop any of that type of
data, would gladly include in the correlation if it is
applicable.
With respect to the flaw growth, I don't
know if research is going anything on this issue. The
recommendation was more looking at how the
predictions, compared to what we observed in the
field. And so it gets more into how well is my
operational assessment performing.
CHAIRMAN FORD: I am not surprised that
you are not firming up on your correlation, and just
adding more uncontrolled data or bad data is not going
to improve your correlation plan.
MR. KARWOSKI: Right. But that's --
CHAIRMAN FORD: You can have as many bad
data points as you like, but that is not going to help
you.
DR. POWERS: I think they made a case for
the pulling and that it wasn't doing too much to it,
and a case gets made when you say, gee, the three-
quarter inch data gets pulled just the same way, and
it doesn't look all that bad.
What is there so unusual about the seven-
eighths, and it is kind of hard to imagine that there
is something different about pulling one.
CHAIRMAN FORD: So as we go down this
path, and then you realize that you are not going to
improve the correlation factors, what is your fall
back?
MR. KARWOSKI: I don't necessarily want to
say that we won't improve the correlations, but --
CHAIRMAN FORD: I guess at this point that
you had better recognize it that you probably won't.
So what is your fall back?
MR. KARWOSKI: Well, if the statistical
criteria are not met to demonstrate that there is a
correlation, the Generic Letter 95-05 methodology says
that if you can't demonstrate that, then you need to
calculate your leak rates in accordance with the
following procedure, which basically says that the
leakage is independent of the voltage observed.
So there is a methodology that already
accounts for that, because back then when we were
doing the 95-05, some of these databases didn't have
a correlation, and so we had to deal with that back
then. So there is a fall back in the methodology.
CHAIRMAN FORD: And thank you very much
indeed. At this point, is that your presentation?
MR. KARWOSKI: That's it.
CHAIRMAN FORD: Thank you very much.
Mario has to leave at 11:30, and the next talk is by
Joe, and who should be talking about some of the
further DPO issues and the new research program.
Mario, before you go, would you like to make any
comments on what you have heard so far?
One of the issues that we have to address
is what is the next action as far as this subcommittee
is concerned, and we are going to write DPO a letter
for the next ACRS meeting, and we have suggested that
in the November-December time frame that we have a
presentation by NEI/NRR on the 97-06.
Do you have any comments on what you have
heard so far?
DR. BONACA: The only one that I mentioned
before regarding performance, and the issue of
prediction that has already been discussed now. That
is the only point that I think we want to stress is
important.
And also this consideration of what do you
include in the predictions. I mean, what should you
consider a multiplier to that.
CHAIRMAN FORD: Thanks so much. And you
like the idea of having a meeting in the November-
December time frame?
DR. BONACA: Yes. I would like to see if
and when we have a new presentation that there would
also be more focus on the objectives of this
integrated plan.
I mean, one thing that I was left with was
that I think I understood the objectives of the NEI
program, and while clearly stated, for the integrated
plan I heard that the objective was to integrate the
activities.
And still I think it would be nice to have
a statement somewhere of what is the purpose of
reintegrating all these activities. We understand it
generally, but often times if you state what the
objectives are, then it focuses better on the plan
itself.
And I would have liked to have seen that
in a statement at the beginning of the presentation.
DR. KRESS: Our obligation is just to have
a letter on the DOP issues?
DR. BONACA: Yes, for right now.
DR. KRESS: And some of the other things
that he is talking about would be just a briefing?
CHAIRMAN FORD: A briefing to this
subcommittee in November or December.
DR. BONACA: That's right. That is just
a suggestion for the briefing, yes.
DR. DUDLEY: I would like to think that if
we did do a review of the 97-06 letter that the
committee would comment back to the staff on it in the
letter in December.
DR. KRESS: Combine in the same letter as
the one on the DOP issues?
CHAIRMAN FORD: We are going to do that
next week.
DR. KRESS: Oh, you are going to do that
next week?
CHAIRMAN FORD: Yes, if we have enough
information, and if we don't have enough information,
we can't comment.
DR. KRESS: Okay.
CHAIRMAN FORD: Okay. Thanks very much.
MR. MUSCARA: Thank you, Peter. My name
is Joe Muscara, and in June of this year the EDO sent
a letter to the ACRS transmitting the action plan that
included DPO issues.
That plan is updated monthly and is
available to you. So the status is really available
within that plan. So what we thought we would do for
this meeting was to more or less concentrate on the
near term milestones.
So we will try and cover some of the work
that has completed in the past year, and address work
that will be going on for about the next year. In the
presentation, I will start off discussing some of the
issues related to materials, engineering, and
inspection.
And then Charlie Tinkler will give us an
overview of the severe accidents and thermal
hydraulics work; and Steve Bajorek will discuss some
thermal hydraulics calculations for predicting the
loads during a steam line break.
And Chris Boy will provide us some input
on some CFD calculations that have been conducted
recently. Under 3.1 of the action plan, the history
of crack propagation in steam generator tubes under a
steam line break condition, and we have planned some
work in this area to essentially start in the new
calendar year, 2002.
What we will be doing there initially is
to obtain some loads, including cyclic loads, during
the MSLB from thermal-hydraulic calculations, and this
will be covered in a bit more detail later.
At the same time there has been an
analysis conducted, and we have submittals in this
area, and so we will also plan on reviewing those
submittals, and try to obtain some of the loads form
those.
We will put together what we think will be
the bonding loads experienced by the tubes during the
MSLB, and based on that we will calculate the crack
growth, if any, for a range of crack types and sizes
using the loads as determined above.
CHAIRMAN FORD: The crack growth is just
tearing, and not sub-critical crack growth?
MR. MUSCARA: That's right. We will
assume that we have some existing cracks, and then we
have the accident, and then we will determine whether
these cracks propagate or not.
As far as the ranges of crack sizes,
clearly we would like to look at initially at a crack
that is stable under normal operating conditions. But
it would be unstable under the steam line.
So with this largest crack, we can one
that will still not propagate a leak. And then we
will take that crack size and determine whether that
would propagate under the steam line break conditions.
But we will look at a range of crack sizes.
CHAIRMAN FORD: Will we be coming back to
discuss some of the details? For instance, what -- as
I understand it, calculating the delta-Ps by some of
the existing hydraulic codes is not necessarily an
easy thing.
MR. MUSCARA: Right.
CHAIRMAN FORD: So will we be discussing
some of the technical challenges and back up if we
can't meet those challenges?
MR. MUSCARA: Right. The discussion that
follows will address that issue.
CHAIRMAN FORD: Okay. Good.
MR. MUSCARA: Another approach that we
will take is to also estimate the loads that are
required to propagate existing cracks. And based on
that we can determine some margins, and what is the
margin over the MSLB loads.
In fact, if we find that we have large
margins, then we really don't feel that we need to
refine the thermal-hydraulic calculations. If in fact
the margins are not so large, then we have to refine
the calculations again, and that will be discussed
later.
And having conducted these analyses, and
we will be using existing procedure for evaluating the
burst and leakage, and mostly burst in this case, we
will then conduct some tests to validate these
analyses.
So then the tests will then take into
account not only the pressure stress, but also the
bending loads and the cyclic loads, and that work will
be done at the beginning of '03.
CHAIRMAN FORD: Again, the question of the
movement of the plates and things of this nature.
This is again a fairly -- in calculating these loads,
it is not a trivial exercise at all?
MR. MUSCARA: Right, and so again what we
are doing there is we will do some of our own
calculations, and the thermal-hydraulics will be
described, and we will look at what the industry has
provided us.
And we will come up with some upper bound
estimates, and then we will use those loads to
determine what happens to cracks. And if we find that
we have small margins, then we will need to do
additional work to refine the analysis.
And another item that is covered in the
operating plan, and also of course addressed in the
ACRS report was damage progression by jet impingement,
and this is jet impingement both under severe accident
conditions and jet impingement from a steam line
break.
Last year, in October, about this time of
the year, we presented some information on the jet
impingement work under severe accident conditions to
the ACRS, and at that time we were more or less agreed
that jet impingement from severe accidents from the
aerosols are not really a problem. There is very
little erosion that goes on.
And the ACRS suggested that we may want to
look at a somewhat longer term test. Our initial
tests were 10 minutes, and we have conducted some
additional tests based on the recommendation.
DR. BONACA: Let me just ask a simple
question. Going to page four, you have or you
mentioned that starting in 2003 that you will have a
test on the tubes under pressure and axial bending.
Why are you waiting so long?
I mean, wouldn't you want to have results
as you do calculations, and that mostly likely,
especially in doing hydraulic calculations, you raise
a question insofar as the modeling, and whether or not
certain effects are being properly modeled.
MR. MUSCARA: Well, the test that I am
talking about is mechanical tests to validate our
analysis. The analytical methods have been developed
and proven over many years. So we don't believe that
the validation tests are going to give us a different
result.
Our main emphasis is going to be using the
procedures already developed, and in most cases it
will be a flow stress model for essentially the
failure criterion. We will also be using some
fundamental analysis on the structural side.
DR. BONACA: It is only a test, and it
going to be purely --
MR. MUSCARA: It is a validation test just
to confirm that the analysis was proved.
DR. BONACA: And that is dealing with
tubes and some force applied to.
MR. MUSCARA: Right. The tests that we
have conducted so far in the models that we have
developed have been mostly pressure stress. So we
want to add to those pressure stresses some of the
bending loads.
And with the bending loads and axial loads
one might see with the support plates moving what the
tubes are doing in terms of support plates.
DR. BONACA: And you said that this
analytical method or models that you are going to use
already are credible for this kind of test?
MR. MUSCARA: Yes. We conducted back in
the '80s 800 tests with different types and sizes of
flaws to predict failure of these tubes.
DR. BONACA: And so you are talking about
the analysis now, and I am talking about the analysis.
MR. MUSCARA: Yes. Well, based on those
tests, we developed analytical procedures and those
have been validated. And tests have been conducted in
other parts of the world that confirm those methods.
DR. BONACA: And these are analyses as you
mentioned are computer codes that you are going to use
to perform these analyses?
MR. MUSCARA: Most of the analysis will be
under stresses, and the evaluation of MSLB, which is
a parameter that describes the stress on the ligament
of the crack.
DR. BONACA: I guess I am asking because
I am kind of surprised, and I just didn't know that
you already had all this information, and models
available, and they were not being used to address
this issue of main steam line break.
MR. MUSCARA: Frankly, if you consider
axial flaws, for example, and we think that this might
propagate under steam line break conditions, I don't
believe that is credible.
I mean, these tubes have got so much
toughness, and it would need to have so much pull to
propagate those flaws that the tube would fail as if
the flaw wasn't there, and it would take a great load.
Now, the other conditions are when we
would have circumvential cracks, and in those
conditions it would be somewhat a little bit
different. I still believe that based on the work
that we have done that it is going to take a great
load to open up these cracks enough to cause a major
failure.
For example, we find that cracks that are
270 degrees around the tube all the way through still
will not open up and give you a large leakage. So I
guess that part of the reason that we haven't done
these tests is because that we have felt from an
engineering feeling that the steam line break loads
will not propagate these kinds of cracks.
And with respect to cyclic loads, yes, we
have some cyclic loads, but how long are these loads
going to be on there. Again, I don't think we have
enough cycles to affect the growth of existing cracks.
But we will do the work and see where we are.
On the jet impingement work as I
mentioned, we have work that is ongoing on both the
aerosol impingement and from a steam line break. The
work on the aerosols was conducted at the University
of Cincinnati with Professor Tabakoff, and the jet
erosion tests have been conducted at Argonne National
Lab.
And I think I mentioned that the rest of
the items we have conducted tests now of up to 30
minutes for the aerosols. Dr. Ford, if we are
stressed for time, I could skip the view graph here.
DR. POWERS: My feeling is that you can
skip over the erosion results.
MR. MUSCARA: Well, I guess the final
outcome of that is that the 30 minute test did not
provide us any different data. We still have very low
rates, about 2 mils per hour with just nickel, and
about 5 mils per hours with nickel, plus aluminum.
And these are much more severe conditions than the
actual aerosols.
DR. POWERS: And I kind of assumed that
was the results that you were going to get.
MR. MUSCARA: In fact, the data was really
indistinguishable from the prior data. All right.
And some results that we haven't shown are test
results on the jet impingement and steam line break
conditions.
Here essentially we have run some tests
with the different sized holes, but concentrating on
the 1/32nd inch hole. There is a specimen spot weld
to the leaking tube, with a stand-off distance of
about a quarter-of-an-inch. So the leaking tube
impinges on this group.
We conducted tests as a function of
temperature, and we find that the most degradation is
obtained at about 280 degrees centigrade, which is
about the cold leg temperature, and where you don't
expect to see cracks.
And then the amount of erosion decreases
as the super heat goes up, and so as the temperature
goes up. So we are getting some flashing and not as
much penetration.
The greatest amount of penetration we had
was about 25 percent of the wall over a two hour test
period. And we will move now on to some comments on
the NDE. There was a comment in the ACRS report that
using a constant POD may not be the best thing.
We have been doing work in this area for
a number of years, and last year again I described
work on a mock-up. We have now some results, and I
think I will go into showing some of the results from
the round-robin analysis of the mock-up.
CHAIRMAN FORD: Joe, I asked the question
to Ken Karwoski about the interrelationship between
the work being done by research on this item, and it
being transitioned into use. Can you make a comment
on that?
MR. MUSCARA: Well, let me give a little
bit of background. We issued this work about 5 or 6
years ago, and at that time I was looking for a
physically based model that we could use for doing the
operational assessments.
The big concern was that we were using for
the voltage based criterion, and it is empirically
based, that there is no physical reason why it should
give us good correlations.
Voltage does not relate to crack size.
Therefore, it cannot relate to crack growth, and crack
growth cannot relate to burst pressures. Generally as
the voltage goes up, the crack size goes up, but there
is a general correlation.
What is not true is that for low voltage
that it is not just small cracks. We are going to
have big cracks that have a low voltage. So in my
mind what was needed was something that was more
robust and more physically based.
So at that time we conducted an
operational assessment. We needed to know the
probability detection so that we can take into account
the flaws that were missed during inspection, and we
needed to know something about cracking issues and
what happens during the cycle.
And of course we needed to know crack
growth grade, and not based on voltage, but based on
some physical parameters. And so at that time we set
up work to learn more about these items.
And one of the key areas of work then was
the probability of detection. So by the time the ACRS
had their comment, we already had done a considerable
amount of work trying to develop POD as a function of
different parameters.
And also this data is available. It is
available for us, and it is available for the
industry, and it can be used as people see fit. We
tried to conduct these tests in a realistic way. We
are using procedures that are used in the field, and
we tried to limit the entire inspection processes
conducted in the field.
We have done the degradation assessment,
and we have the right techniques, and qualified
techniques, and qualified people doing the
inspections.
We have a five-person team that has done
the inspections, and so we have tried to reproduce as
much as possible the process that goes on in the
field. With respect to the tubes and the division
itself, the same thing.
We developed a fairly comprehensive mock-
up with different conditions of dents, and corrosion
products, and transitions, and realistic flaws,
developed in the lab with realistic flaws from the
point of view of signal, and so we believe that we
have a reasonable test.
And we do have now some results that may
be POD to some other factors besides the --
CHAIRMAN FORD: Am I missing something?
That although you have this data, it is not being
used?
MR. MUSCARA: Well, this data is just
evolving. In research, the main emphasis is to
develop also a code that can be made available to the
NRC staff so they can do their own independent
operational assessments. POD is one input to this
code, and precision crack code would be another code.
So that code is under development and the
data is becoming available, and our first topical
report will be published before the end of this year
providing these results.
And of course the results have been made
available, and we have reviewed the draft reports, and
so we are aware of the information.
CHAIRMAN FORD: So we are ahead of the
ball game here on this particular result?
MR. MUSCARA: Yes, I think so.
DR. POWERS: The ad hoc committee -- I
think you have to understand that the NRR staff has a
different set of problems. They need to detect and
then they need to predict, and they need to predict
what kinds of things show up in between the two.
What the ad hoc committee was concerned
was about was using a constant POD with respect to
carbon stone was that as the technology for sampling,
for inspecting tubes improved, and as the technical
understanding improved, you wouldn't be able to
correct things, and take into account, and it is a
draconian thing.
So when we moved to something that was
more easily corrected, and that is all that this
research is doing, and it was basically an endorsement
of this research.
MR. MUSCARA: In fact, the Generic Letter
95-05 made some comments at that time, and they in
fact did say that they felt that the voltage raised
criterion is acceptable for now, but we should be
moving towards more physically based criterion.
And the ACRS said that, and so based on
that also we felt a need to develop this kind of data.
The results were that the upper left figure shows the
POD is a function of depth, and for flaws at the tube
support plate, both for the OD and the ID.
Quickly, we noticed that the ID flaws are
more easily detected if the POD is higher, and that is
reasonable because we get in general larger signals
from the ID than from the OD. There is not as much
penetration of the ID currents.
In the next view graph we are showing a
similar plot, but with respect to voltage, and we see
here that the role is reversed. What I need to
mention is that once the voltage gets considerably
high, all the POD get to be about the same.
But for lower voltages, we are getting a
better correlation with the OD flaws. At one point,
for the ID flaws, we also had the dents. So many of
the flaws at the support plate that were originally
from the ID also had a dent.
That means that we had a signal which was
not very clean. Now, because the inspector looks at
the signal rise on the plane to a vertical position
for calling it a crack, and because there is a dent
signal, and because ID flaws only have a small range
of phase angle shift, the signal does not rise very
much, and can also be buried in the noise.
So in this case the ID flaws showed a
lower POD than the ID flaws. But this shows in
general that we can plot that POD is a function of the
depth of the flaw, and POD is a function of the
voltage.
And the bottom graph essentially shows POD
for the tube sheet section, where we have a couple of
tube sheet flaws also with the tube transition, the
role transition being present that complicates the
signal.
Besides looking at the flaw size, flaw
size and voltage by itself, a very useful parameter to
plot the PODs as a function of MLSB, and again MLSB
describes the stress at the ligament of the flaw. It
directly relates to the burst pressure.
So here we can relate POD as a function of
a structural integrity parameter, and we noticed that
the POD gets to be reasonably high if LIDSCC parameter
of greater than 2.3 would correspondence to a flaw
that would fail at 3 delta-P. So the POD for cracks
that are at 3 delta-P can be fairly high.
And just to show it from the view graph
and to make an other point that even though our
results are qualified, what we noticed for certain
conditions, such as the tube sheet, and the top two
graphs, we are plotting the results on a team-by-team
basis. The others were combined results.
And we noticed that the teams more or less
cluster fairly close together for those two examples,
but in other cases -- for example, the free span,
where the teams are not use to looking at the flaws of
the free span, they find lots of flaws on top of the
tube sheet and support plates, and not so much at the
free span.
And also for the support plate for the
LIDSCCs, there is quite a bit of scatter in the team
performance. The good team is quite good, and the
number of teams right there is sort of an average.
But there is always a team that does not
perform as well, and again I would like to stress that
these are teams that are commercial teams, and they
are qualified, and they are conducted in inspections
in a manner that is similar to what they do in the
field.
And if anything of course they know that
they are under test conditions, and so this is under
best performance.
CHAIRMAN FORD: And the lines on these
grants -- I'm sorry, but what are they?
MR. MUSCARA: They are just a different
team. The assembles are a team and also the line is
also a team. So we had 11 teams participating in this
round robin.
CHAIRMAN FORD: Oh, I see.
DR. POWERS: The best team and the worst
team had to change lines, and everybody else --
MR. MUSCARA: And it is just a logistic
thing. So we are showing you essentially the variance
between the best and the worst team. I mean, this is
very useful data when we are doing probablistic
analysis.
So I think more or less we have addressed
the issue for ACRS as to other methods may be useful,
and we already have data in this area. There is one
item that I would like to cover --
CHAIRMAN FORD: I'm sorry, but I am
violating my own principle of not asking questions,
but if you would go back to the bottom right-hand
slide, the IDSCC tube support plate and the biggest
scatter. Is that purely because the cracks are on the
ID and the eddy can't pick those up for some reason or
other?
MR. MUSCARA: No, because one thing is
they are doing quite well if you look at the green
light.
CHAIRMAN FORD: Yes, but the scatter.
MR. MUSCARA: Well, yes, the scatter, but
what is the complicating factor of course with these
flaws is that there is a role transition, and that
role transition provides a fairly large signal.
CHAIRMAN FORD: Oh, so you have a float
between the --
MR. MUSCARA: It is a complicated signal,
although --
DR. SHACK: But this is the tube support
plate there though?
MR. MUSCARA: I'm sorry? Oh, yes, this is
the ID with the dent. So you do have considerable
noise, and some things do better than others. I think
here again that it is a matter of -- there may be a
signal there as a matter of calling it a crack.
And because the signal is more and doesn't
have a large shift-in phase, and there is a
complicated noise signal, it still is difficult for
the inspector to notice it to call it a crack. They
may confuse it as being part of the noise signal. But
the good inspectors do quite well.
And this next view graph is not really at
all to do with materials. I see that Jack Hays is in
the back of the room and he can answer any questions
on this.
This is the item on the item spiking. We
have conducted an assessment of the ADAMS and Atwood,
and Adams and Sattison spiking data this summer, and
I understand that this review has been completed.
And the plant having a response to the
ACRS comments by December, and our evaluation of this
will be published for public comment around February,
and then based on the public comment, there is a final
position that will be put together.
I understand that after we evaluate our
position on this issue that we could be willing an
able to provide a presentation to the ACRS on that
position before it goes out for public comment.
So I think this is something that is up to
you if you want to hear about this or not after we
have assembled a position on it.
DR. POWERS: Comments are always the same.
That is more work than it would take to solve the
problem completely. Do it the way that you want to,
but that is an awful lot of work for a problem that I
think is susceptible to a technical resolution.
MR. MUSCARA: Jack, do you want to
respond? No? Okay. Well, I am almost finished,
because the next view graph is milestones and is
fairly far into the future, but there will be work
going on next year in this area, and I know that Peter
will be interested in this.
So I decided to discuss this a little bit.
Now, we are planning on conducting some tests to
better understand the crack initiation and crack
growth. And we are taking the comments from the ACRS
to heart. We want to conduct tests under realistic
conditions of stresses, temperatures, and environment.
That means that we need to evaluate better
what goes on in crevices. As far as the tests
themselves, they are not defined yet, but we may be
using model boilers so that we can reproduce the
thermal hydraulic conditions and the crevice
conditions, and therefore, have the appropriate
crevice chemistry.
We may have to measure the crevice
chemistry, and we may just run tests and evaluate the
cracking behavior, and then measure the crevice
chemistry at the end when we are not at operating
conditions anymore. But it is very difficult to
instrument these crevices.
So there are a number of ideas that we are
considering. The work is not defined, but we will be
looking at crack initiation, and crack growth, and
using tubular specimens, along with other types of
specimens.
And hopefully under realistic fuel
conditions, and the idea here again is not necessarily
to develop the mechanisms, but to develop data that
will be useful for our code for doing the assessments,
the operational assessments. And we need crack
initiation data and crack code data.
DR. POWERS: A couple of questions, Joe.
As people move to 690 are you going to be testing 690?
MR. MUSCARA: Yes, thank you. We will be
testing 690, along with the 600. The idea here is
that we have a great deal of information on the
behavior of 600 in the field.
So we will be conducting tests with 600
mill anneal, and 690 thermally treated, so that at
least we know the behavior in the laboratory; and then
knowing the behavior of 600 in the field, hopefully we
can extrapolate the behavior of 690.
It may be well that on 690 to just make a
couple of comments. Now, 690 is susceptible to
cracking in different environments. It has cracked in
the laboratory, and cracks in environments that are
not overly aggressive. It cracks in neutral solutions
and sulfates, and in copper, and in lead.
So what we want to do is with respect to
690 to evaluate the range of conditions under which
this material is susceptible so that we can get a
better idea about its behavior in the field.
In addition to this, we have had Professor
Staley working on crack initiation. This work was
just started about a year ago, and he is modeling
this. But we have also been looking at some of the
field data.
When we look at the data for 600 mill
anneal, and we consider the cracking that we are
experiencing these days, and not necessarily the
caustic cracking that we got in the early days.
We will consider cracking at the support
plate and crevices. Well, 600 mill annealed has taken
10 years before it experiences this kind of cracking.
So the fact that 690 has gone 10 years doesn't give me
that much more comfort yet.
We know that in the laboratory that it
behaves better, and I do believe that it will behave
better, but I don't know whether it will last 40
years. But through this work hopefully we will get a
better feeling for the behavior of 690, as compared to
600.
DR. POWERS: Another thing that I noticed
-- and as you say, trying to instrument to understand
what is going on in crevice corrosion -- and probably
because it is small, and things just don't fit in
there -- I noticed that within the corrosion community
there are people -- I mean, crevice corrosion is not
peculiar to nuclear plants. It is a lot of places.
But there are people who are trying to
develop what they call scaling laws for crevice
corrosion. In other words, to do experiments that are
scaled where you can instrument, and then you try to
find out how does that scale down to the real
crevices. Are you paying any attention to that kind
of work?
MR. MUSCARA: Well, actually there is work
also going on related to steam generators. Jesse
Lumpson at Rockwell Science Center is doing some work
for EPRI, and he has been doing work for a number of
years having a typical crevice.
And he has done quite a good number of
studies himself, but also this crevice model has been
taken to a plant in Japan, where they are conducting
tests using the coolant from the plant.
So they are developing good model data,
and we will take advantage of that. My feeling is
that we will still need to run some model boiler
tests, where we reproduce the crevice under thermal-
hydraulic conditions, and see how the materials
behave.
We will try to research it as much as
possible. Some of the things that we can certainly
get are temperature, and maybe potential, and maybe
MPH. It would be interesting to be able to get
chemical species, and that is a harder problem.
EPRI is working on it and they may in fact
by the time we are ready to do something have some
solutions on how to do that experimentally. But one
thing that we can fall back on is what is in the
crevice after we have shut down the system. That will
give us a clue as to what was there in the operating
conditions.
CHAIRMAN FORD: I have a couple of
questions, Joe. On Task 3.8, that relates to the
whole question of how can you correlate a bonding, a
linear correlation of voltage of this type, with non-
linear performance, time dependent performance, of the
cracking phenomena?
That latter part would come out at 3.10,
and how are you going to from a management point of
view compelled in this information in 3.8?
MR. MUSCARA: From 3.10 and also from the
inspection work. My belief truly is that the voltage
does not track crack size or crack code. The linear
literature is not with crack code, but with voltage
code, which is meaningless.
So we happen to have a linear correlation.
We didn't try to make a scatter code really. There is
quite a bit of scatter, and so I don't know whether it
is linear or what it is.
But I think my point is that there should
not be a correlation there with crack growth, but we
will find a correlation with actual flaw sizes.
CHAIRMAN FORD: So as we look down the
time, and if what you say is correct, which I think it
is, should we not be looking for another spectrum
methodology which is more related to the physics?
MR. MUSCARA: Yes, and I think in general
that we are doing that in our program, and we have
come up with some fairly good techniques for sizing
flaws. I presented the slides here and some reports
are being published on this.
But we have come up with a very good
technique for characterizing flaws, and particularly
the flaw profile. And from that we can get directly
MSLB, and we have been able to predict the bursts of
these tubes from the flaw profile and from the MSLB
correlations.
EPRI is also working on different
techniques for better characterization flaws, and the
industry has moved towards other plugging criteria.
For example, at the tube support plate crack and the
idea with dents. This is an area where they are using
the profile of the flaws.
They are getting away from voltage and
using the actual profile to determine the burst
pressures. And I believe that is a direction to go
into, and I think we are moving in that direction.
CHAIRMAN FORD: And industry is responsive
to these?
MR. MUSCARA: Well, that is what industry
is proposing, and utilities have come in with an
ultimate criterion.
CHAIRMAN FORD: Now what sort of time
scale are we talking about for this more physically
realistic inspection?
MR. MUSCARA: Well, I think the
characterization methods that we have now -- in fact,
EPRI is a member of our IC program, international
cooperation. And they are aware of this process that
we have developed for sizing flaws, and we are
exchanging information, even to the point where we are
going to turn over the algorithms.
CHAIRMAN FORD: Are we talking about six
months, a year?
MR. MUSCARA: Again, right how this is a
laboratory tool, and so in order to develop for the
industry more work needs to be done to make it more
user friendly.
And once it is in the hands of someone who
wants to turn it into a field system, we are talking
over a year or so. But again besides their own work,
there are other things that are coming up. For
example, this probe for doing better detection and
probably better characterization of flaws.
We are evaluating that, and that is
something that is almost industry ready. They have
done a lot of work getting data from plants, and we
are also incorporating them into our round robin
exercises. So we are evaluating that advance in
technology.
So technology is advancing, and I think to
the point where we can start making use of the actual
parameters of the flaw. They should be profiled and
length in depth, and then we can more accurately
predict failure.
CHAIRMAN FORD: I have one more technical
question, and then we should discuss the ACRS type
actions that we have to take. On this one here, Joe,
how do you take into account that we just don't know
what is a good heat and what is a bad heat?
MR. MUSCARA: That's true, but what we
will probably do is catch bad heats, and work on the
bad heats so that at least we will be conservative on
what we find. If we get a good heat, we will be
testing forever and get no data.
CHAIRMAN FORD: Yes, I understand that,
and so your strategy on this is that we will go for
the worst case scenario and just happens to have by
chance some good heats?
MR. MUSCARA: Frankly, I have not thought
too much about doing heat variability in this work.
We will probably wind up doing several heats, but
probably not a tremendous amount of heats.
And again the idea generally would be to
find some susceptible heats, where we can do our work
to evaluate different parameters on cracking.
CHAIRMAN FORD: Okay. Joe, thanks very
much.
MR. MUSCARA: So I guess now we will have
the discussion on thermal hydraulics.
MR. SULLIVAN: This is Ted Sullivan from
the staff. I would like to make one additional
comment. I think you started to touch on it when Joe
was mentioning that this is a laboratory tool, and it
is being made available to the industry.
But in terms of making a transition to
applying that to ODSCC as a substitute for the
voltage, first of all, you have got to get industry --
I don't know who the you is, but industry has to be
interested in basically making another proposal to the
staff, and developing it to the point where it is a
suitable substitute for the staff.
And it has to happen -- if something like
that were going to happen one of two ways, either the
industry has to take it up and make a proposal in the
room, or the staff would have to make a safety case
that this sort of thing needs to be done.
And I don't think it is our view that it
would be easy to make any sort of safety case, but
that sort of transition needs to be conducted.
CHAIRMAN FORD: Okay.
MR. TINKLER: Joe described for you some
of the work being done by the Division of Engineering
and Technology in the Office of Research. I am going
to summarize the work that is being done in the
Division of Systems Analysis Regulatory Effectiveness
in the Office of Research that primarily addresses the
issues related to severe accident and design basis
thermal-hydraulic conditions that create at least in
part some of the loading conditions on the steam
generator tube.
Be advised that all three divisions in the
Office of Research actually are contributing to this
initiative. The Division of Risk Analysis and
Applications is also heavily involved with NRR in
integrating this analysis into our understanding of
risk that are posed by steam generator tubes, both
from the standpoint of initiating events on the design
basis, as well as the risk from severe accidents.
Oh, and I am Charlie Tinkler, and I will
be followed by Steve Bajorek, who will talk to you
about our current thinking on the thermal hydraulics
questions related to support and steam generator tube
loads.
Chris Boyd will also describe in more
detail some recent analysis that he has completed on
the staff to address the details of mixing in the
steam generator and the steam generator tube --
CHAIRMAN FORD: If I could just give you
some guidance. WE have another meeting beginning at
one o'clock, and I guess the members would really like
some lunch. So if we can try and finish the whole
thing by say, 20 by 12:00 at the latest, and bearing
in mind that the information that we want to get a
feeling for right now is whether the recommendation in
NUREG 17-40 are being incorporated into this joint
proposal.
MR. TINKLER: Okay. This is a list of the
major recommendations of the ACRS Ad Hoc Subcommittee
on the DPO. They are going to be addressed in this
presentation and that are covered by the work in our
division.
We want to develop a better understanding
of the behavior of the steam generator tubes under
severe accident conditions specifically addressed by
Steam Generator Task 3.4.
The evaluation of the -- and ACRS also
recognizes that we evaluate the potential for
progressions of damage to steam generator tubes during
the rapid depressurization caused by a main steam line
rupture. That is the more traditional thermal-
hydraulic issue, and that is specifically addressed in
the action plan under Item 3.1.
To address the severe accident response of
steam generator tubes, and general hydraulic boundary
conditions in the reactor coolant system, and
corresponding component behavior in the steam
generator tubes, we have four basic parts to this
research.
We have the system level code analysis,
and the system using SCDAP/RELAP. That is where we
model the core, the RCS, the steam generator tubes,
and all the other related components.
We are relying in part now on
computational fluid dynamics code analysis and
modeling, principally the FLUENT code, to model the
single phase natural circulation and mixing in the
steam generator tube bundle.
It gives us a much better portrayal of the
spacial dependencies and resolutions of temperatures
within the system. We are assessing again the 1/7th
scale test data.
These are the tests that were sponsored
originally by EPRI in the 1980s, and later co-
sponsored with the NRC as a mock-up of a steam
generator -- of two steam generators and a reactor
vessel.
The tests were designed and conducted
primarily by Westinghouse personnel, and so
occasionally you will hear them referred to as the
Westinghouse 1/7th scale test.
We are also contemplating conducting some
new experiments to investigate conditions that weren't
addressed in those original 1/7th scale tests that
have been raised in the DPO and raised by the ACRS,
and I will talk about those briefly.
Under 3.4, we have a multitude of subtasks
that address a lot of the technical issues related to
severe accidents. These are some of those technical
issues. Some of these have their own separate
subtasks in the action plan.
Plant design differences. We have done
the bulks of our calculations for the SERE (phonetic)
design, which was the original basis for our tube
integrity analysis for NRR.
We started looking at -- and we have done
calculations for other plants, and we are now focusing
our attention on the Zion-like geometry, and that has
a number of advantages.
It is representative of a bigger group of
plants, and it also allows for a little better
comparison with some of the industry analysis, because
the industry analysis more often is done for a Zion-
like geometry.
And we have plant sequence variations, and
we typically focus on station blackout type sequences,
where one steam generator is also depressurized. The
steam generators have all boil dried, and the core has
become uncovered, and now we have super-heated steam
circulating through the loops.
Now we have a counter-current flow that we
are primarily concerned about because for most of our
calculations we predict the loop seal for the red
coolant pumps is filled.
So we get counter-current flow out through
the hot leg, and through the steam generator, and to
one-third to one-half of the steam generator tubes,
and returning through the remaining portion of the
tube bundle, and back along the bottom of the hot leg
to the reactor vessel.
This task is to look at variations on that
sequence, and to look at the effects of reactor
coolant pump seal leaking, and to look at leakage from
PRVs or safety valves to see if there are variations
on the sequence that pose some unique challenge.
In response to past ACRS comments, we are
conducting a more rigorous uncertainty analysis to
look at the influence of mixing parameters and other
phenomenological issues in this calculation as part of
the system analysis.
The ACRS raised in its ad hoc subcommittee
report, and we recognize the importance of loop seal
clearing in this analysis. The effect of clearing the
loop seals is to have unit-directional flow through
the steam generator tube bundle,and not get the
benefit of return mixing through the coolant portion
of that flow.
So it typically predicts higher
temperatures. It is normally associated with slightly
depressurized sequences, and so we looking at those
two effects combined. The effect of tube leakage on
inlet plenum mixing --
DR. POWERS: Are you going to be able to
resolve the issue of loop seal clearing just with
analysis?
MR. TINKLER: We think so. We know that
we have to present more analyses and our rationale to
the committee on this matter, but we believe that is
the case, and we understand the comments that have
been raised, and we understand the concerns about
small delta-P clearing loop seals.
We understand that, and we have work to do
on that, but right now we expect to address that
analytically. The effect of tube leakage on other
plenum orientation, and this is the notion that if you
have tube leakage up in the bundle that it will
disrupt the mixing in the inlet plenum that was
observed in the 1/7th scale test.
So you won't get quite as an efficient
mixing and you get perhaps channel flow or tunnel flow
up through the inlet plenum, and that can create a
locally hotter plenum.
And hot leg/inlet plenum orientation. The
1/7th scale test looked at a proto-typic Westinghouse
steam generator, where the hot lay comes in low on the
inlet plenum. The CE designs have a hot leg
orientation that comes in a little higher on the inlet
plenum.
And so it is a little closer to the tube
sheet, and so the argument there is that the mixing
path lends a shorter -- you might not get effective
mixing in the inlet plenum and the tubes will be
exposed to higher temperatures.
These are areas that we expect -- that are
well-suited to CFD calculation, but they also would
benefit from additional testing, and we are
considering that.
The things that we have to be mindful of
are the scaling issues associated with these kinds of
tests, and the need to run them with a denser fluid,
like SF6, and that poses a problem in some facilities.
There are a host of instrumentation
issues, as well as costs. Tube to tube variations.
When we do air calculations with control volume codes,
we have relative coarse nodalization of these volumes.
And inlet plenum is basically three control volumes.
Now, that's okay if you are using the
empirical models developed by the experimenters, but
if you want to hope to model the response of tubes or
clusters of tubes in a 3,000 tube bundle, you need
finer resolution.
So we were looking to see if the analysis,
as well as perhaps additional testing, to get more
insights on that. And fissure pipe deposition. This
relates to the risk impacts.
The ACRS has commented in the past that we
might not be taking full credit for those severe
accidents where tube leakage or tube rupture occurs.
The fact that that tube bundle and the upper internals
of steam generator will serve as a mechanism for
deposition of aerosols. These radioactive aerosols
wouldn't be transported off-site
Now, there is testing that is planned in
the Artis facility in Switzerland, the Paul Shearer
Institute is conducting tests where they have a mock-
up of steam generator tube bundle, and they are
looking at the deposition of aerosols under their
severe accident conditions or a range of conditions.
Here is 3.4, the near items. We are doing
system level analysis to look at sequence variations
to look at the effect of reactor coolant pump seal
leakage,and to look at issues associated with safety
valve leakage, and to look at the effect of tube
bundle leakage.
And we are looking at the effect of tube
bundle leakage from a systems standpoint, and not a
local CFD level. We are also looking at alternate
steam generator depressurization. Typically, we do
these calculations with the pressurizer loop steam
generator being the one that is blown down and
depressurized.
And we have calculations being done
looking at the other three loops to see if it makes a
difference, and we are also looking to see to the
extent that we clear loop seals in some of these
calculations.
We have done the calculations where we are
halfway between a draft report and a final report, and
so we are not quite ready to talk to you about these
results, but we will in upcoming subcommittee open
meetings.
Our next task is to reevaluate some of the
SCDAP/RELAP modeling and simplifications of
assumptions, things like radiation heat transfer and
the hot leg; and some of the loop seal clearing issues
we hope to address in this.
It might also give us a vehicle for
looking at some of the comparative items between
industry calculations and our calculations.
Subtask 3.4e.1, benchmark of the CFD
methods. That is the FLUENT against the 1/7th test
data, and this work was just recently completed on
schedule in August. Chris Boyd will talk to you about
it in more detail.
Lastly, design basis and thermal
hydraulics. This was to address the issues in the DPO
that were raised by the depressurization by blowing
off a relief valve, or a main steam line break. That
is just a cryptic summary of the kinds of loads.
Steve Bajorek will just describe to you a
little more of our thinking at this point on how we
are going to tackle that issue, and he is next.
MR. BAJOREK: Good morning, or good
afternoon, I guess now. My name is Steve Bajorek, and
I am also a member of the SMSA branch, and relatively
new to that branch.
What I am going to talk about are some of
the issues pertaining to generating the hydraulic
loads that we are going to need to evaluate the blow
down forces on the steam generator. The work that we
are doing initiates from two different contentions.
I have listed them both here, and both
arise due to the uncertainty in what are the hydraulic
loads and forces that result across the tube sheet,
and across the tubes during the break, and the rupture
of the main steam line break, or potentially another
relatively large pipe connected to the secondary side
of the system.
By way of background, I think it is useful
to think of the high pressure depressurization of a
system into two overall segments. We can think of the
first phase; that while this fluid is primarily
subcooled, and while the depressurization waves
propagate through the system at a sonic velocity,and
then another phase of that depressurization once those
waves have dissipated, and the system depressurizes
primarily dependent upon the break flow and the size
of the break.
This is an issue that is actually of
fairly well-studied in the initial design of a reactor
system from the point of view of the primary; whether
you have rod drop or not, or whether you will have
grid crushing within the core, is dependent on your
design and how you evaluate the breaks to the primary
system to take a look and track the depressurization
waves as they move through the loops, and through the
core, and potentially move the core barrel from one
side of the downcomer to the other.
A good analysis of that type of event
tracks the waves at sonic velocity, and incorporates
a fluid structure interaction between the core barrel,
which is the primary component of interest in that
type of an analysis, and generates the delta-Ps from
one side of the downcomer to the other side that we
give to the structural analysis so that they can
perform a structural analysis and tell us whether the
rods are dropped, or whether the core barrel deflects.
We have a similar situation now that we
need to address on the secondary side. Now, I think
the reason why that has not received as much attention
as the hydraulic forces that develop on the primary
side has to do with the rate at which those waves move
through the primary or through the various systems.
For the primary system Tcold -- and C
stands here for the sonic velocity, and this is at
about 550 degrees fahrenheit, at typical Tcold at
pressure, moves through the system at a little bit
greater than 1,000 meters per second.
If you think of the primary system full of
sub-cooled liquid early in the transient, this wave is
certainly capable of moving through the loops in the
core on the order of a couple of dozen times, and
interacting with waves which move throughout other
parts of the primary system, generating fairly complex
loading across the core barrel or the steam generator
divider plate, and other things that need to be looked
at.
And causing some of those components to
move. And we need to start thinking about what that
type of analysis or evaluation does now over on the
second side. But it is important to keep in mind that
the most important physical parameter which determines
the velocity of that wave is its density.
And in the primary system, typical
conditions are that we are seeing velocities a little
bit later than a thousand meters per second. On the
secondary side, the velocity that we might find in
saturated liquid at about 900 psi, just a little bit
less than what we would see on the primary system, the
difference being the difference in the density.
However, in the vapor space, that velocity
drops significantly to roughly half of its value. Now
when you do a thermal-hydraulic evaluation of the
primary system, that analysis to take a look at the
interaction of the waves goes for on the order of
milliseconds, because what happens is that as soon as
we start to form some voids within the system, those
waves are dissipated very rapidly.
And the interaction of the waves becomes
a no, never mind, in the analysis. It is something
that will probably help out the structural evaluation
here on the steam generators secondary side. That is
not to say that those loads are going to necessarily
be small, because there will still be a fairly
significant shock to the tube sheet and resulting
motion.
Now, because the steam generator either
has significant voids through the bundle region at a
steady state, or has an interface at no load
condition, the most significant pressure wave that is
going to cause motion of the tube sheet and transient
stresses on other components within the steam
generator is going to be this initial wave that moves
through the steam generator.
We won't have much in the way of
reflection or interaction, with the exception of the
fact that we have more voids on the interior of the
steam generator, and sub-cooled fluid in the
downcomer, and so conceivably we could see a wave
moving down the steam generator downcomer, and
reaching that portion of the tube sheet earlier than
we would in the interior of the bundle.
So our initial approach -- and we have got
to admit that we are in the very initial stages of
developing a plan of attack at this point -- is to try
to develop relatively conservative hydraulic loads
that we can give to delta-P(t) for them to apply to
their finite element model, and to determine the
bending stresses and other stresses that they get out
of that type of an analysis.
Our approach is first going to try to use
what I will call glorified hand calculations to
determine, one, what is the initial time at which that
depressurization wave reaches the tube sheet and
various parts of the steam generator base, and augment
that with track 3-D calculations to look at the later
stages of the blow down of the steam generator
secondary side.
Now, during that phase of the accident
something like a TRAC or RELAP should give us a
reasonable depressurization. I would not expect it to
do a credible job during this very initial part, where
you have to TRAC the sonic wave and the interactions
that it has with the various components.
That's why our initial plans are to try to
get something that is conservative with the hand
calculation, and augment it with the TRAC-M
calculations, and give that to the finite element.
And if you can come back and tell us
whether we have lots of margin, or there is a little
bit of margin on that. If the answer comes back that
we have just a very small amount of margin, the next
part of our evaluation would be to replace the hand
calculation with something better.
That would not necessarily be TRAC-M. I
think we have to look at that closer and make up our
minds whether it could or could not do that. The
tools that might be available to us to analyze this
are the things like the multiplex code that is used by
Westinghouse to evaluate the subcool blow down on the
primary side.
The staff a number of years ago to my
recollection did have access to a code, and I think it
might have been called SLAM, to take a look at that
type of a scenario on the primary side.
That might be a better starting point than
trying to force the TRAC-M to give us that type of
sonic wave depressurization. But we would go along
that path if we were to find that we wouldn't have
enough margin and structural analysis, and then make
a decision on what would be a more appropriate tool.
If necessary, then look at some
experimental testing to try to augment our code
validation at that point. At that point, if we had
such limited margin, that might also be a good time to
go back to the vendors and use perhaps their tools to
try to evaluate the same type.
CHAIRMAN FORD: Thanks very much indeed.
MR. TINKLER: Thank you.
MR. BAJOREK: I have more slides than
eight minutes, but I am just going to go through them
quickly. Charlie covered a lot. This Charlie
described, and I am just going to show this as the
thermal-hydraulics of interest that we are going to
focus on in this small subtask that I am carrying on.
I want to make this point before I start,
and that is that the SCDAP/RELAP code is what we are
relying on to get our thermal-hydraulic results to
pass on to the materials people.
The tube temperature predictions that the
tubes are subjected to come out of that code, and they
are influenced directly by mixing parameters. So I am
making the point that we are going to use SCDAP/RELAP
and that gives temperatures that are affected by
mixing parameters, and these mixing parameters are
fixed in the code, and they are determined from the
1/7th scale testing principally and other tests if
possible.
So these mixing parameters are what I am
going to focus on. The advantages of CFD. I just
give this slide to show an example that we are about
four orders of magnitude more cells, on the order of
hundreds of thousands, to a million, versus 10 to a
hundred.
Less expensive experiments as you pointed
out, and we are going to have a direct resolution of
mixing. We are not tuning the code. We are using the
most appropriate turbulence models from an academic
point of view, and then just letting the code go.
So again no fixed mixing parameters. We
are extending the data with CDF, or we will to full-
scale, full-pressure, full-temperature steam, and then
we can look at this inlet geometry effects and tube
leakage effects that Charlie mentioned.
DR. KRESS: Do we have options in the
fluid code or for what turbulence parameter, different
options for turbulence parameters?
MR. BOYD: We have different turbulence
models, and several to choose from, and then within a
turbulence model, you can then tune that to the data.
We are not doing that type of tuning. We are kind of
using industry standard coefficients.
We don't really have data to do that kind
of tuning, and we are not tuning to get the answer we
have from the 1/7th scale test. We are just letting
it fly.
DR. KRESS: Are you choosing one option,
or are you --
MR. BOYD: We chose several options just
to look at the differences. In the end, they did not
make a lot of difference. The one that we chose was
the second order of Reynolds Stress Turbulence Model,
which is for this type of flow, it is -- I guess on
paper it would be the appropriate model, as opposed to
a two equation K-epsilon model.
So in this type of flow field, I guess we
chose the academically appropriate, and in all the
selections that we made there wasn't a large
difference. It did not affect these types of
parameters.
This is a quick slide to show the CFD
approach versus a lumped parameter. The top picture
shows the hot leg, and I guess that is not really
showing up, but what you see is a full counter-current
flow profile, with velocity vectors and temperature
profiles.
And on the right in a lump parameter code,
SCDAP-RELAP, there is just two pipes with a single
temperature, and you have mass flow and temperature.
In the inlet plenum, this is the SCDAP-RELAP
nodalization in the middle on the right, and you will
see the three mixing volumes.
Flow comes in and based on the mixing
fraction, it either goes to a mixing volume, or it
passes up through to the tubes, to again a fixed
number of tubes.
With the CFD predictions, we are going to
calculate the mixing implicitly with the code, and
then as far as the tubes go, this is something that we
will add a benefit to our predictions.
In the SCDAP-RELAP predictions, you are
going to get one temperature and a number of tubes and
up-flow that is predetermined. And in the CFD
predictions, we will get the number of tubes
calculated implicitly, and then we will also get tube
to tube variations.
So we will know not just the average
temperature going into the tubes, but what the peak
average ratio is.
DR. KRESS: On your counter-current flow,
what do you do at the reactor end?
MR. BOYD: At the reactor end, initially
I put the core in there, and I just had a heat source
and let it go, and it picked up that counter-current
flow. I had a lot of uncertainty in my core model
obviously.
I was using a lot of core options, and I
cut that off, and at this point I just put on the end
of the hot leg a mass flow in.
DR. KRESS: You just put it at one end?
MR. BOYD: That's right.
DR. KRESS: And that stuff going out just
disappeared?
MR. BOYD: It is called a fixed pressure
boundary there.
DR. KRESS: A fixed pressure boundary?
MR. BOYD: Yes. And I did a lot of
variations with different velocity profiles, and all
sorts of things to match the mass flow given in the
test results.
DR. KRESS: And you had to specify the
profile specification?
MR. BOYD: That's right, and I found that
my profile specification wasn't all that significant.
By the time that it got to the steam generator end of
the hot leg, it had dissipated anything that I had put
in.
So, CFD is going to provide an improved
understanding of the 1/7th scale data. We have got
these tests, and obviously what went on in the tests
was fine, but we have a limited view of the tests from
the limited instrumentation.
So we can fill in some of the gaps with
CFD, and then we can extend to full-scale. One of the
big questions is does scale affect the mixing
parameters, and that is something that we are looking
to address right now.
At that point, when we have gone to full-scale,
we have answered that question among others, and then
we can start looking at the effect of tube leakage and
how that affects these inlet plenum flows, and mixing
parameters, and the effect of the inlet geometry
variations, like the CE plants with the hot leg
entrance closer to the tube sheet.
And again we will get implicitly out of
this tube to tube variations that then would give some
understanding of what the hottest tube really is.
The schedule. Validate the technique by
looking at the 1/7th scale. That is our best
available data set. That has been done and in general
the answer is that the code picks up all the relevant
physics and does a pretty good job.
At this point, we are sensitivity studies,
and extending the predictions to full-scale, using a
kind of best estimate conditions out of a SCDAP-RELAP
analysis.
Again, what is the effect of scale, and
then we are going to complete additional studies on
tube leaking and inlet geometry variations, as well as
other sensitivity studies.
And just to give a quick view. This is
the mesh that we that was used for the 1/7th scale.
It's a symmetry model, half of the hot leg in the
plenum and tubes. All the tubes in that test, 216,
were modeled individually.
We won't do that at full-scale, and we
will come up with a model for the tubes. But that
gives an idea of the resolution. There are several
hundred-thousand cells just in the inlet plenum alone.
There are some qualitative results. This
is the first thing that hits you when you -- well, all
of the qualitative flows predictions are correct. In
other words, a sloping interface in the hot leg, and
a plume that rises and dissipates fairly quickly into
the inlet plenum, and about a third to a half of the
tubes in up-flow.
The temperature of the tubes reaching the
given values in the test, and by the time it reached
the top of the tubes. All these kinds of qualitative
features were matched by the CFD predictions. This is
quantitative data, but I'm just talking qualitatively
there.
When we go to the actual mixing parameters
of interest, this table shows some of the results.
These are the tests of most interest. In general,
what you saw was about a 10 percent deviation.
If you look at the Westinghouse data
carefully, you will determine that the uncertainty in
that data is around 10 percent or more. The one big
variation was the number of hot tubes.
We were 23 tubes over, which is about 10
percent of the tube sheet, and we are currently doing
some sensitivity studies to determine what boundary
conditions or condition in our model might affect that
to see if we may have a problem.
And all the hot average temperatures, and
mass flows, and things like that, were all pretty
close, and in this particular run we had a 15 percent
difference in the recirculation ratio, which again I
believe is in the uncertainty of the data.
So as a quick look, what I get out of this
is that the code can do this type of analysis, and
that the results are pretty close. This is the tube
sheet flow, and this is the number that I mentioned,
10 percent over-predicted.
The dark region on the tube is from the
data. There is two lines there because the data had
an uncertainty, and not every tube was instrumented.
So somewhere in that range is where the boundary
between where up-flow and down-flow in the tube sheet
occurred.
And then the outer dashed line represents
the FLUENT predictions. On the right, I give the peak
temperatures. The peak thermal-couple in the data
read 59 degrees celsius in this case. These again are
cold tests done with SF6.
The maximum predicted value from FLUENT
was 61.5 degrees, and that was on the center line.
The data did not have any center line thermal-couples.
If you look off-center line, it would be more
consistent with the data. I had a max prediction of
58.5, which was pretty close to the measured value.
So as a summary, the CFD predictions are
generally within 10 percent of its 1/7th scale data,
and that is generally within the experimental
uncertainty. There was a fair amount of uncertainty.
There was no mass flows directly measured in the
tests.
They had to infer that from energy
balances, and some of these energy balances were
inferred from small delta-Ts. So this added to the
uncertainty.
The phenomena observed during the tests
were all predicted by the CFD code in a qualitative
sense, and so the general flow features are there, and
work on full-scale predictions is proceeding now, and
I think we have a high degree of confidence in our
technique, and so when we go to full pressure, full
temperature steam, there is not going to be as much
uncertainty.
So this benchmarking exercise has been
very valuable, and I think this is just a restatement
of that. The CFE technique has been demonstrated to
be applicable, especially for predicting these mixing
parameters, which are kind of average values.
And this work provides this high degree of
confidence, and we are going to go to full-scale
analysis, and at full-scale, then we will spend our
time doing the tube leakage and geometry variations,
and our sensitivity studies. I am just a few minutes
over.
CHAIRMAN FORD: Thank you very much
indeed. I would ask for any comments from the members
here. We have on our schedule for the next ACRS
meeting next week -- we are charged with a letter
relating to the DPO.
And essentially hopefully saying that the
recommendations that were in 17.40 are being followed
in the new NRR research plan. That is hopefully what
the letter would say. Is that correct?
DR. KRESS: The intent is to address that,
yes.
CHAIRMAN FORD: Okay. Could we have some
comments to help the staff and research as to how they
would appropriate their time for the 30 minute
presentation that they would have in that one hour?
DR. KRESS: I would like the approach
where they are listing what the ad hoc committees'
recommendations were, and then to say how we are
addressing them in the plan. That would work very
well. I certainly would want to have the full
committee see this CFD stuff, and that addresses some
of the --
DR. SHACK: But we will never get through
it in 8 minutes.
DR. KRESS: But that addresses some of the
real issues that the staff may have.
DR. POWERS: The plans are sufficiently in
the works, and I don't see why the subcommittee
chairman can't just summarize it.
DR. KRESS: I think that is probably right
there.
DR. POWERS: Well, all you are going to do
is say the staff has plans to address this issue, this
issue, this issue, and this issue.
DR. KRESS: And they look like good plans.
DR. POWERS: And in 9 out of 10 cases,
they have great plans, and in one case, I haven't got
a clue.
CHAIRMAN FORD: The one question I have
got, Dana, because I know nothing at all about it, is
the thermal-hydraulics codes. Are you all feeling
that these are the right approaches?
DR. POWERS: The one thing I know is that
if you put two thermal-hydraulicists in a room, the
one thing they cannot arrive at is a conclusion. What
I would say is why don't we have the subcommittee
chairman draft a summary, and put it up for the rest
of the committee, and say we are addressing the issues
that have been raised, because there is no more
content than really that that they are addressing.
I mean, most of these things are in the
works, and they are working on it, and then allow the
speakers on this CFD stuff and the counter-current
flow, because that implies so many things other than
the steam generator tube --
DR. KRESS: And Dr. Wallis hasn't heard
that.
DR. POWERS: Well, more in the context of
here is some research that is going on now, and here
is where we stand, and more as an update of general
interest than just a DPO issue.
CHAIRMAN FORD: And that you think will be
enough sufficient information to allow George to sign
his name to a letter saying essentially that the
recommendations from the ad hoc committee, and
therefore the ACRS, are being followed?
DR. POWERS: Are being addressed, yes.
They are taking them into account. That is what we
were asked, and they know them better than I do.
CHAIRMAN FORD: So the answer could be
yes?
DR. POWERS: Yes.
CHAIRMAN FORD: Just one word.
DR. POWERS: Yes.
CHAIRMAN FORD: So you are asking me to
stand up in front of the ACRS committee and summarize
what we have heard today, and then for general
information to have the thermal-hydraulic guys
specifically get up and talk?
DR. KRESS: As an alternative, if that is
uncomfortable to you, you could ask one of these guys
to summarize.
DR. DUDLEY: Just from a public holding,
and a presentation in a public meeting, at a full
committee meeting to write a letter from, I think it
would be more appropriate if the staff presented a
summary, and then it would also save the subcommittee
chairman the effort of pulling that together.
DR. POWERS: But a summary presentation.
DR. KRESS: Yes, a summary presentation.
DR. POWERS: I think the committee as a
whole is going to be very interested in what they are
doing with this counter-current flow issue because it
has been around since the dawn of time, and there has
been lots of concern about it for a variety of things.
And let that talk go on at length.
DR. SIEBER: And also the tube sheet --
DR. POWERS: Well, that one is
interesting, but I think that we are fixing to work on
this. I think we can hold that one off until they
have got some more results.
CHAIRMAN FORD: Could I suggest the
following? Who is going to stand up and say I am the
project leader for this and this is a problem, and
where you are going, and this action plan, the joint
NRR/research plan, is feeding into that overall
thrust.
Just one draft, and one slide saying this
is where we are going in general, and I am quite ready
to stand up and say this is in line to go alone with
your line. Here is the action plan, and here are the
actions in the NRR/research program, and these are the
ones that we specifically recommended, et cetera.
Does that sound fair?
DR. POWERS: Yes.
CHAIRMAN FORD: Is that clear?
DR. POWERS: All right.
CHAIRMAN FORD: All right. We are
adjourned.
(Whereupon, at 12:30 p.m., the meeting was
concluded.)
Page Last Reviewed/Updated Tuesday, August 16, 2016