 |
Search Options | |||
| Index | Site Map | FAQ | Facility Info | Reading Rm | New | Help | Glossary | Contact Us |  |
|||
Home > Electronic Reading Room > Document Collections > Commission Documents > Commission Meetings, Slides, Transcripts, Meeting SRMs, and Full Written Explanation for Closed Meetings > 1997
>
05/06/1997
1
UNITED STATES OF AMERICA
NUCLEAR REGULATORY COMMISSION
- - -
BRIEFING ON PRA IMPLEMENTATION PLAN
- - -
PUBLIC MEETING
Nuclear Regulatory Commission
One White Flint North
Rockville, Maryland
Tuesday, May 6, 1997
The Commission met in open session, pursuant to
notice, at 2:05 p.m., Shirley A. Jackson, Chairman,
presiding.
COMMISSIONERS PRESENT:
SHIRLEY A. JACKSON, Chairman of the Commission
KENNETH C. ROGERS, Commissioner
GRETA J. DICUS, Commissioner
NILS J. DIAZ, Commissioner
EDWARD McGAFFIGAN, JR., Commissioner
. 2
STAFF AND PRESENTERS SEATED AT THE COMMISSION TABLE:
ANNETTE VIETTI-COOK, Assistant Secretary of the
Commission
KAREN D. CYR, General Counsel
EDWARD JORDAN, Deputy EDO
SAMUEL COLLINS, Director, NRR
GARY HOLAHAN, Director, Division of Systems Safety and
Analysis, NRR
CARL PAPERIELLO, Director, NMSS
ASHOK THADANI, Deputy Director, RES
THOMAS KING, Deputy Director, Division of Systems
Technology, RES
DENWOOD ROSS, Director, AEOD
. 3
P R O C E E D I N G S
[2:05 p.m.]
CHAIRMAN JACKSON: Good afternoon. I'm pleased to
welcome members of the NRC staff to brief the Commission on
the status of the NRC PRA implementation plan. The PRA
implementation plan was first issued in August 1994. The
plan is intended to be a management tool to help ensure the
timely and integrated agency-wide use of PRA methods and
technology in the agency's regulatory activities. The last
written update on the status of activities in the PRA
implementation plan was provided to the Commission in
January of this year. The Commission was last briefed on
the plan in October 1996.
During today's briefing the staff will discuss
recent accomplishments in particular where they have made
risk-informed decisions. They will discuss revisions to the
PRA implementation plan, draft regulatory guidance for
public comment, performance monitoring and pilot
applications, other pilot projects, and plans for future
activities.
The draft regulatory guidance documents and
standard review plan sections provide guidance on acceptable
approaches for making plant-specific risk-informed changes
to the current licensing basis of a nuclear power plant in a
specific area. The staff is recommending that these
. 4
documents be issued for a 90-day public comment period.
I and my fellow Commissioners are looking forward
to your briefing today. I understand that copies of the
viewgraphs are available at the entrances to the room.
If none of my fellow Commissioners have any
opening comments, Mr. Jordan, please proceed.
MR. JORDAN: Thank you, Chairman, Commissioners.
Our briefing this afternoon will focus on the documents that
were forwarded to the Commission by SECY-97-077. We will
also discuss selected achievements described in the
quarterly status update, SECY-97-076, which was issued April
3, and then in SECY-97-095, which is the tech spec program.
With me at the table today are Ashok Thadani and
Tom King from the Office of Nuclear Regulatory Research; Sam
Collins and Gary Holahan from Nuclear Reactor Regulation;
Carl Paperiello from Nuclear Material Safety and Safeguards;
and Denny Ross from the Office for Analysis and Evaluation
of Operational Data.
All of the focus of this presentation is on the
regulatory guide standard review plan. Dr. Ross and
Dr. Paperiello are here representing their offices'
important roles in the PRA program plan and can respond to
questions related to AEOD and NMSS PRA activities.
As you know, Ashok Thadani has recently assumed
the position of Deputy Director of the Office of Research.
. 5
CHAIRMAN JACKSON: Congratulations.
MR. THADANI: Thank you.
MR. JORDAN: In this new capacity he will continue
to be responsible for overall coordination and monitoring of
the agency's PRA program plan and will begin today's
briefing.
MR. THADANI: Thank you.
May I have viewgraph number 1, please.
[Slide.]
MR. THADANI: As you noted, Chairman Jackson, and
Ed Jordan did as well, this is clearly an activity where all
the program offices are involved. The focus of today's
briefing is going to be in three areas: the regulatory
guides, the quantitative measures that we propose be
utilized, the status of the pilots, and the issues related
to performance monitoring.
I will very briefly go through some of the other
issues to indicate that work is going on in other areas as
well, but our focus is going to be on those three areas.
I will cover the background and some of the recent
accomplishments as well as where we are on the
implementation plan. Then Tom King will go through the
draft regulatory guidance, the criteria, and what we are
doing by way of posing a set of questions to get feedback
from the public as well as industry.
. 6
After his presentation is complete on the PRA
portion, he is also going to touch upon the issue that came
up at the last meeting with the Advisory Committee on
Reactor Safeguards, the issue of human data and coordination
of that activity. Chairman Jackson, you had asked that we
address that issue.
Gary Holahan will cover the performance monitoring
and pilot applications and describe our future actions.
May I have the next viewgraph, please.
[Slide.]
MR. THADANI: We have been providing quarterly
reports to the Commission on status of the implementation of
the activities described in the plan as well as semiannual
briefs to the Commission on status of these activities.
At the October briefing we covered some of the
policy issues. These were issues, like should safety goals
be used on plant-specific basis or should small increases in
risk be allowed?
The Commission was also provided in January a
status report on the activities and the plan.
January 22, 1997, the Commission provided guidance
to the staff on those key policy issues. As we had
indicated to the Commission, we were moving in the direction
of using those guidelines and the guides. After we received
the Commission SRM on this issue we finalized our guidance
. 7
documents to make sure that these documents were consistent
with the guidelines described in the safety goal policy
statement, the regulatory analysis guidelines documents, and
other related documents.
We met with the Advisory Committee on Reactor
Safeguards as well as the Committee on Review of Generic
Requirements and have got their endorsement for these guides
and documents to be issued for public comment.
In April, as Mr. Jordan noted, we provided to the
Commission two documents, a status of the implementation
plan, SECY-97-076, as well as SECY-97-077, which is a fairly
thick document. It includes the general regulatory guide,
the standard review plan, and topic-specific guides like
graded QA, in-service testing, and so on.
In that document we also provided a draft Federal
Register notice and highlighted the set of questions we
proposed that we get feedback on from the industry as well
as the public.
May I have the next viewgraph, please.
[Slide.]
MR. THADANI: These are just some examples of some
of the recent accomplishments. Obviously the reg guide and
the SRPs have been provided to the Commission. They provide
framework and guidance for making changes to licensing basis
of individual plants. Tom King is going to say a great deal
. 8
about that.
Another report that we recently sent to the
Commission was the technical specification pilot
application. This is working with the Combustion
Engineering Owners Group wherein they had proposed changes
in allowable outage time in the area of safety injection
tanks. These are basically passive tanks. They wanted to
change allowable outage time from one hour to 24 hours, and
low pressure safety injection train outage time from three
days to seven days.
We have used an approach consistent, as described
in the regulatory guide, and provided a safety evaluation
report approving those allowable outage time extensions.
That information has been given to the Commission for
information. If there are any questions or concerns, of
course we will address them.
The approach we used there was to work with the
lead plant. Arkansas Unit 2 was the lead plant. There are,
I believe, ten plants that would be interested in these
changes. We would expect to issue our evaluation on those
ten plants by the end of July 1997.
CHAIRMAN JACKSON: You expect to issue?
MR. THADANI: Safety evaluation reports, July of
1997.
I will note that there are one or two questions
. 9
that we are going to have to deal with for one or two
plants, because it appears in some cases the calculated mean
core damage frequency is higher than 10 to the minus 4 per
reactor year. That is an element that needs further
discussion. Outside of that, we expect to be able to issue
the safety evaluation reports approving those extensions in
allowable outage time.
In February we issued NUREG-1021, Revision 8,
which is the operator licensing examiner standards. These
standards have now in them a number of the insights that
have been gained through risk assessment studies and they
have become part of the training program as well as
examination portion. They identify, for example, dynamic
testing considerations, pick up the more significant
plant-specific accident sequences to see if they are covered
through simulated training, et cetera.
All of those issues are now captured in this
revision. It was published in February of 1997, after the
Commission approval was received in December 1996.
CHAIRMAN JACKSON: Can you say how the guidance
documents themselves were informed by the pilots or the IPE
reviews, if they were?
MR. THADANI: The guidance documents give a number
of insights and lessons. You will hear some of it.
MR. KING: I was going to cover that as part of
. 10
mine.
CHAIRMAN JACKSON: You are going to cover it in
your presentation?
MR. KING: Yes.
MR. THADANI: Yes.
CHAIRMAN JACKSON: Okay.
MR. THADANI: But we can come back to it again to
make sure.
CHAIRMAN JACKSON: We'll wait.
MR. THADANI: May I have the next viewgraph,
please.
[Slide.]
MR. THADANI: If it appears I am moving quickly, I
am, so that we have an opportunity to go through some of the
issues that I know you are very interested in.
As you know, AEOD staff has been working on
evaluating voluntary approaches to reporting reliability and
availability data and the feasibility and practicality of
that approach, and we expect to have a paper to the
Commission in the next few days and anticipate that there
will likely be a separate briefing as well on this topic.
We have also conducted a workshop on the insights
from the IPE program and we have a briefing tomorrow
afternoon on IPE, and we will cover some of the lessons and
things we have learned tomorrow afternoon during that
. 11
briefing.
CHAIRMAN JACKSON: Let me ask you for the
Commission's edification. What were the objectives of the
IPE workshop and were they met?
MR. THADANI: I would ask Tom King to address
that.
MR. KING: There were several objectives. One was
to give the industry an opportunity to ask questions
regarding what we felt were the important insights from the
IPE, to provide information on things they have done since
the middle of their IPE. Most of those submittals were
years ago. It gave us a chance to talk about our IP
follow-up activities, which you will hear about tomorrow.
Ultimately, we understood the industry had been doing some
IP insights work themselves, and it gave them an opportunity
to present to us what they had been doing on their own
initiatives.
So it was a multipurpose workshop.
CHAIRMAN JACKSON: Let me ask you a question about
your first bullet, your evaluation of the voluntary approach
for reporting reliability and availability data. What would
be the scope of that voluntary approach? How many SSCs,
systems, structures and components, and how does it compare
in terms of the number of risk-significant SSCs in a plant,
and if the scope is different than the scope of the
. 12
maintenance rule, why so?
MR. THADANI: Dr. Ross.
MR. ROSS: Of course this will be covered in more
detail in the paper. The description that we got from INPO
shows up very nicely on an embedded diagram, sort of like a
bin diagram, where the safety system performance indicator
is embedded into a larger group of maintenance rule,
safety-related and other equipment. It would be covered by
INPO but not part of the maintenance rule itself.
The voluntary approach would consist of all of the
information under the safety system performance indicators
and other information. As we will explain in the paper, it
does contrast with the scope of the proposed rule that went
out. Our arguments will show where the two are different,
how we intend to make up for the differences.
CHAIRMAN JACKSON: So the answer to the question
is, the scope is different than the scope of the SSCs in the
maintenance rule?
MR. ROSS: The scope of the voluntary approach?
CHAIRMAN JACKSON: Right. That's what I'm talking
about.
MR. ROSS: I believe in detail, yes. In terms of
types of information.
MR. JORDAN: Maybe I could comment. The scope of
the maintenance rule is very large. The scope of the
. 13
reliability data rule was relatively narrow. The scope of
the data that would be obtained and used includes the
principal data elements from the reliability data rule plus
access to additional data for other systems and components.
So we are continuing to structure the scheme of analyzing
the data consistent with the reliability data rule, but
there is not a deficiency in the scope.
CHAIRMAN JACKSON: You are answering questions the
way I answer them. Let me ask it this way. What is the
overlap between the scope of SSCs that are covered in the
maintenance rule and the scope in this voluntary approach?
Not the voluntary approach vice the reliability data rule,
but the voluntary approach vice the maintenance rule.
MR. THADANI: If I may just comment on this, I
don't think the answer is very crisp. However, it is fairly
clear that even within the -- first of all, the proposed
rule scope of systems is fairly narrow.
Let me just now go to the maintenance rule scope,
which is very broad. It includes SSCs, both safety-related
and non-safety-related, covering various aspects. Then the
industry is to convert these SSCs into high
safety-significant and low safety-significant categories.
The focus all along of the agency efforts has been to make
sure we have information on high safety-significant
component.
. 14
Then you go to the voluntary program. The desire
clearly would be to try and get information to cover those
SSCs that have high safety significance. That could be a
plant to plant variable. I think that issue is going to
need some further evaluation, and I think you are going to
see in the paper discussion the need to do some more
evaluations to be able to give a crisp answer.
CHAIRMAN JACKSON: For a given plant, will the
scope of the SSCs covered in the voluntary approach be a
subset of those most safety-significant SSCs in the
maintenance rule, or is it not that crisp?
MR. THADANI: I think it will clearly be a subset.
CHAIRMAN JACKSON: Where does the lack of
crispness lie?
MR. THADANI: The lack of crispness is in that
clearer definition that all of those SSCs are in fact
covered in the voluntary program.
MR. ROSS: Chairman, one of the ways we are going
to break down the answer, matrix or table is looking at the
parameters such as failures -- all of these comments are
under the voluntary approach -- then showing how failures,
for example, would be provided for the small set known as
the safety system performance indicator and then how would
they be covered for everything else of high safety
significance under the maintenance rule. This matrix is
. 15
developed for failures, demands, run times, and so on.
It's a rather complicated answer, but I think we
have covered it all in this table.
MR. JORDAN: I think we owe you that discussion in
a broader presentation.
CHAIRMAN JACKSON: I think you do.
MR. JORDAN: It is not terribly simple.
MR. THADANI: Quite honestly, that is why I
thought it was likely that there will be a need for a
briefing on just that topic.
May I have the next viewgraph, please.
[Slide.]
MR. THADANI: During this three month period we
did not really make any major changes to the plan. However,
I do want to touch upon some of the schedule or issues and
briefly cover the status of the pilots, and then we will
have some additional discussion as we go through.
What has happened basically is the whole process
of developing these documents, making sure that the agency
is involved and supportive of what we are trying to do, as
well as our interactions with various committees. I think
the Advisory Committee on Reactor Safeguards as well as CRGR
has taken a lot of effort and time, more so than I think we
had anticipated. That has had some impact. We have had to
take time away occasionally from pilots to make sure we
. 16
dealt with those issues.
I do want to summarize where we stand on these
pilots. I indicated that the technical specification,
safety evaluation report is complete, and that we would
expect to issue the remaining safety evaluation reports for
other CE plants in July of 1997.
We have a team, as we speak now, at South Texas
working on the graded QA program. Our expectation is that
barring some surprises from this visit we expect to finish
our safety evaluation report by the end of June of 1997.
In-service testing is yet another pilot that we
have been working on. We have recently put together a set
of additional questions to make sure that what we are doing
under IST is in fact completely consistent with what we are
saying in the regulatory guides. We expect to get fairly
quick responses to those questions and complete our
evaluation by the end of June of 1997.
The fourth pilot activity was in-service
inspection area. The in-service inspection, in my view, is
probably more challenging in terms of the issues on
methodology than some of the other pilot applications,
because now you are getting into areas like trying to get an
idea -- incidentally, the scope of ISI is piping, all
classes of piping. You need information on flaws, flaw
distributions, fracture mechanics. These are more
. 17
contemporary approaches in terms of models.
We have been working with the industry on two
approaches. One is the ASME Westinghouse approach, which is
very probabilistic in nature; another approach from Electric
Power Research Institute, which is less dependent on
numerical analysis and more qualitative type of importance
analysis type of an approach.
While have been working on the methodology issues,
we have not received any submittal from any of the pilots.
We expect Surrey to come in in September, using the ASME
Westinghouse owners group methodology. It appears that
perhaps Arkansas, and I think Fitzpatrick, may also come in
using the EPRI approach.
Clearly you will hear through the presentation
that in terms of in-service inspection we cannot complete
our final document until we have actually gone through the
pilot application. However, we do have a draft guide that
we expect to get to the Commission in July. That will go
out for public comment and those will be the ground rules
that we will apply as we go through the pilot evaluation.
CHAIRMAN JACKSON: Do you plan to add any
risk-informed performance-based initiatives to the PRA
implementation plan?
MR. THADANI: I don't know of any specific plans.
The Commission asked us in an SRM to not just be limited to
. 18
performance-based thinking as far as the PRA implementation
plan is concerned and that it may be necessary to develop an
implementation plan for performance-based thinking in other
applications. If I remember correctly, we owe the
Commission that response end of August, and we are working
on that.
CHAIRMAN JACKSON: So you are going to address it
at that time?
MR. THADANI: At that time. As part of that
activity we would be meeting with the industry to solicit
their views in this area.
CHAIRMAN JACKSON: This was asked in the context
of another meeting, but I will ask it again within this
context. It seems that there is some delay. We had
discussions about the development of risk-based indicators,
and the question is, what impact do you think any delays in
developing the risk-based indicators will have on plan
schedules for their use in the senior management meeting
process?
MR. ROSS: Obviously we have taken a good look at
the replacement set for the current PIs with risk-based
indicators. I think it would probably have a moderate
effect. I was looking at one of them in particular. We
have a very deterministic approach to significant events
now, when an event can be called significant. One concept
. 19
to replace it is using a tool like ASP to make a more
quantitative description of what is a significant event. I
don't think it will be perfect. I think there will be some
significant events that it will still quantify low.
I would expect this would have a moderate effect
on the senior management meeting. The admonition is we are
not supposed to be overly influenced by singular events. I
think with that precaution I would expect it to have at
least a moderate effect. Whether it takes some additional
risk-based training to understand this and criteria to
understand what is and what isn't significant, I think it
would probably take some additional training as well. These
are supposed to be phased in, according to the plan, by
1999.
MR. JORDAN: The present set of indicators we felt
have been risk informed, but now this is really a transition
to the risk based.
CHAIRMAN JACKSON: Since we are talking about the
PRA implementation plan, for the record I would like to hear
from Dr. Paperiello on where we stand in terms of the
development of PRA or like methods in your areas, fuel cycle
facilities, industrial devices containing nuclear materials,
et cetera.
MR. PAPERIELLO: Could I have the backup slides
for materials?
. 20
[Slide.]
MR. PAPERIELLO: We have worked in several
different areas. You are going to have a presentation next
week on performance assessment in both high level waste, low
level waste and decommissioning. Performance assessment in
those areas looks much like PRA in the sense that you have
models, you have inputs with, instead of discrete values, a
range of values.
For example, if you look at Yucca Mountain, in a
PRA sense it will be rain or no rain, because that is a
significant factor in the model. For Yucca mountain you
don't have that. You have a range of rainfalls. So that
becomes a distribution that goes into the model rather than
a yes or no or up or down value.
What comes out is identical to what comes out of a
PRA. You have a risk distribution or dose distribution, as
you will, that is characterized by a 95 percent confidence
level and 5 percent confidence level, a mean, a median and a
mode. You can choose how you are going to measure. We use
median values, for example, in reactor space. We have a
tendency to use mean values for what we do in performance
assessment. So there is that group of things.
We have used PRA methods or risk-based methods in
transportation. The modal study done several years ago. We
are looking at that and using it to iterate the existing
. 21
NUREG-0170, the EIS on transportation that was done in the
late 1970s, to update it with the insights in the modal
study.
At our request Research is initiating a plan to
apply PRA to spent fuel storage facilities. We are looking
at methods to look at the risk associated with industrial
gauges containing cesium 137 and cobalt 60, and we have
developed an integrated safety assessment procedure for fuel
facilities to assess the risk from chemical safety, critical
safety and fire safety integrated. So they are the
activities we have undertaken up to now in applying PRA in
the NMSS side of the house.
We have in our budget plans in the future to
actually set up a PRA group in NMSS to see how we can apply
it in all our areas.
CHAIRMAN JACKSON: Can you have the slides shown
again, please?
[Slide.]
CHAIRMAN JACKSON: When do you expect to come to
closure? Let's leave aside the high level waste repository.
For instance, on your next to the last bullet, or the ones
involving transportation, but particularly the ones to
demonstrate methods for PRA of spent fuel storage facilities
or for determining the risk associated with industrial
gauges, when do you expect to come to closure on some of
. 22
these?
MR. PAPERIELLO: On the gauges, I'll have to ask.
MR. COOL: Good afternoon. The contract for that
particular action with Research is scheduled for the summer
of next year, that is, summer of 1998.
CHAIRMAN JACKSON: Thank you.
MR. THADANI: If I may just make a quick comment.
In reactor applications, the Commission's safety goals and
the all the guidance of the Commission has given the staff
is to utilize mean values and not median. I just want to be
sure that there is not a misunderstanding of that. We are
using mean values, and I think it is very important that we
use mean values in these analyses.
Unless you have other questions, I'm going to go
to Tom King.
MR. KING: If I could have slide 6, please.
[Slide.]
MR. KING: Slide 6 through 14 provide a summary of
the reg guides and SRPs that were provided to you in
SECY-97-077.
What I want to do in the briefing is focus on the
overall approach and key issues associated with those
documents. Just a little background information.
As you recall, the specific plans and schedule for
developing these draft documents were put together after a
. 23
November 30, 1995, request from Chairman Jackson. The
purpose of these documents is intended to help implement the
Commission's PRA policy statement by providing guidance on
an acceptable approach for making plant-specific,
risk-informed changes to the current licensing basis of
nuclear power plants.
The documents include general guidance, which
provides an overall approach in guidance applicable to all
risk-informed proposed CLB changes.
Then there is supplemental guidance in specific
areas that are shown on the viewgraph. The supplemental
guidance is not a replacement for the general guidance; as
it says, it supplements the general guidance.
Also included in the package was draft NUREG-1602.
This was prepared as a reference document to aid in making
decisions on the scope and attributes of a PRA that would be
appropriate in proposing a risk-informed change to a CLB.
Chairman Jackson, you had mentioned where the IPE
had really helped influence this package. This is probably
the most prominent area where we took IPE insights in terms
of strengths and weaknesses of PRA methods, databases, and
so forth, that the industry used and folded them into the
guidance that is in that draft NUREG.
CHAIRMAN JACKSON: When using the guidance
documents, will the staff be able to use the documents to
. 24
judge the quality of a PRA-based submittal?
For instance, let me give you some questions.
Will they be able to judge whether the appropriate
models were used, appropriate data used, appropriate common
cause models used, appropriate human performance modeling,
or distributional assumptions? Can you make some comments
in those areas?
MR. KING: The answer is yes to all of those. The
intent of having the draft NUREG and the guidance in the reg
guides and SRPs is to answer yes to all of those questions.
COMMISSIONER ROGERS: Before you leave that, one
question on the CLB. The work that you have done to date
concerns looking at using risk information for changes to
the current licensing basis. How far would that approach
take you, or could you use that to actually restructure the
CLB on a risk-informed basis itself?
MR. KING: Go back through the regulations and see
what would change if you apply risk insights. I think
clearly starting with the safety goals and using metrics
associated with core damage frequency, accident prevention
and mitigation, the containment type requirements, would be
used in any such process. We haven't really thought about
taking this reg guide and are the metrics we developed for
it appropriate for such an analysis, but I think where you
start from would be the same, the overall guidance.
. 25
CHAIRMAN JACKSON: It might be worth considering.
MR. THADANI: We do have other areas, Commissioner
Rogers, as you know, that look at some of the regulations to
see what sort of value there is in some of those
requirements. In making judgments there, it seems to me we
would have to use the same sort of thinking and be
consistent as we go forward.
COMMISSIONER ROGERS: That's a bigger job.
MR. THADANI: A much bigger job, yes.
COMMISSIONER ROGERS: It's always a little easier
to look at incremental effects. But the general approach
that you have had to adopt in analyzing changes, that may
have given you some first steps towards what one might have
to do in restructuring a CLB on a risk-informed basis.
MR. THADANI: Yes.
CHAIRMAN JACKSON: The guidance documents call for
increased management attention when changes approach certain
guidelines. You lay them out: core damage frequencies in a
certain range with deltas of a certain size. Is it clearly
spelled out what increased management attention means in the
guidance documents? Otherwise, can you end up in a case
where an approved pilot becomes the de facto standard
guidance?
MR. KING: There is a set of items that should be
looked at when you are in that increased management
. 26
attention region. It deals with things like recent plant
performance, recent operating events, uncertainty analysis,
sensitivity analysis, scope of the PRA, things like that
that are highlighted. As you get closer to those guideline
values you want to take a closer look at what is going on in
terms of what is the uncertainty range, how do you deal with
it, what other qualitative factors may influence my decision
one way or the other.
MR. THADANI: I would just add to that that the
thought process as you get closer and closer to these
guidelines. The degree of robustness of the analysis would
have to go up, and greater attention has to be paid to
issues of defense in depth or what does that really mean; is
there a great deal of reliance on human actions?
You asked a question in terms of value of IPEs,
human reliability issue. We can give guidance and the best
available techniques. The recognition is still there that
there are very large uncertainties. Those are going to be
difficult to deal with, particularly if we have a plant
whose performance we are very uncomfortable with. Core
damage frequency may be very low; the change in core damage
frequency, while it may be small, we can't lose sight of the
fact that the agency is concerned about performance of that
plant. Those factors have to be integrated, and the
management has to play a significant role in that.
. 27
MR. KING: If I could have slide 7, please.
[Slide.]
MR. KING: As you heard ACRS say last Friday, we
had some extensive interactions with them in developing
these regulatory guides. They felt it was constructive; we
felt it was constructive. We feel the guides are much
better off for that give and take and frank discussion we
had with ACRS.
We also had similar discussions with CRGR. With
the pilot programs we had interaction back and forth. The
pilots provided some real world examples on the types of
changes that the industry will be asking for. The
practicality of the risk metrics and other traditional
engineering type criteria or guidelines that we propose, is
it practical to apply them? Do they cover a wide range of
the types of changes that we believe will be coming in
proposed by the industry?
Also, it had a chance for us to interact with the
industry on expectations in terms of the quality and scope
and depth of their analysis. We felt there was a broad
range of feedback that we got from the pilots in that
respect.
To get back to the IPEs for a minute, they
provided some examples also in terms of the value of the
risk metrics that we proposed. We could see from them what
. 28
their baseline core damage frequencies were, for example,
where they made changes based on their IP, what they
represent in terms of core damage frequency and other risk
metrics. So there was some valuable feedback from that as
well.
COMMISSIONER ROGERS: Could you make any comments
about the nature of the interactions with CRGR and what came
out of those?
MR. ROSS: I can respond. In the first place,
this was not an imposed backfit. So we noted that. So
50.109 really was not triggered. We noted it was what we
call a measured step along the path towards risk-informed
regulation. Small but measured.
At that point you could say our strict CRGR role
was complete. We reviewed the imposition of requirements.
But we also have a value added role, and we noted that we
are really talking about fairly small numbers. In some
cases even smaller increases in these small numbers, and it
might be difficult to characterize this as a change within
the general feeling, especially as you get close to 10 to
the minus 6.
We noted that there had been due consideration of
the safety goal, and by and large we thought it was a good
step. We thought the staff in the period that we dealt with
them over a few months did an incredible amount of work, and
. 29
we so said. We complimented them on the give and take and
the cooperative effort. It was a large job for them.
COMMISSIONER ROGERS: Thank you.
MR. THADANI: Let me note that CRGR views are the
last two pages in this document, summarizing basically what
Denny said.
MR. KING: Finally, before we leave slide 7, as
you noted, Chairman Jackson, the package is at the
Commission for approval. Included in that package is a
Federal Register notice, which has a series of topics from
which we would like feedback. It also indicates our
intention to hold a workshop during the public comment
period. We now have that scheduled for the third week in
July. It will be here at NRC headquarters, in the
auditorium. So we are anxious to get that out on the street
and let people make their plans to attend.
CHAIRMAN JACKSON: See how it floats.
MR. KING: Slide 8, please.
[Slide.]
MR. KING: In developing these documents we had
several fundamental questions which had to be addressed
early in the program so that we could establish and settle
in on an overall approach for these documents.
Specifically, we had questions regarding where do
these documents fit in the overall regulatory process, what
. 30
is the benefit to licensees and the staff of using these
documents, and how do we maintain consistency with
Commission policies and practices.
We have settled in on an approach that basically
puts these documents forth as one acceptable method for
licensees to propose changes to their current licensing
basis where NRC approval is required. These do not affect
50.59 type changes. In effect, they provide an alternative
way to utilize risk insights when licensees propose changes
under 10 CFR 50.90 through 92, which is license amendments.
Since these documents were written basically as a
result of a PRA policy statement, we consider them voluntary
on licensees. However, we have taken the approach or are
taking the approach that using risk insights will be done by
the staff in reviewing proposed changes to a plant CLB. So
even if licensees come in and don't utilize risk insights,
the staff is still free to ask questions regarding risk.
The benefits to the licensees and to the staff I
think we expressed well in the PRA policy statement:
improved decision-making, more efficient use of resources,
and the potential for reduction in unnecessary regulatory
burdens. So we feel there is certainly an incentive for
both licensees and staff to use these documents.
Finally, we spent a lot of time trying to make
sure that these were developed consistent with previous
. 31
Commission guidance and policies. One particular item I
will note in that regard was the definition of the current
licensing basis that we chose to use, which was straight out
of 10 CFR Part 54, our license renewal rule. We feel that's
a good definition. We feel it certainly can fit well into
the context of these documents, and that's what we propose
to use.
Slide 9, please.
[Slide.]
MR. KING: At our last semiannual briefing we put
a slide up that talked about a six-step review process. Our
six-step review process is now a four-step review process.
We haven't eliminated anything, but we have
recognized that what we call engineering analysis, you can't
really separate the traditional engineering from the
probabilistic from the integrated decision-making. It
really has to be done together and it complements one
another.
So structurally we have rewritten the document to
basically be a four-step review process, the steps you see
on the slide here.
We feel that the implementation and monitoring
program is still a very key element in all of this. It's
important to verify the validity of assumptions and analysis
and provide a vehicle for feedback and corrective action if
. 32
we find out from real plant data that things aren't turning
out the way we were expecting. So it's an important part of
this process.
CHAIRMAN JACKSON: Let me ask you a question.
Since you are saying that one would have to do the overall
engineering analysis that has the three pieces you have
outlined, have we ended up adding a layer of analysis net
net? I was going to say for ourselves, but I'll say for
yourselves since you are going to do the analysis. Can you
give me an answer to that?
MR. KING: Clearly you can view it as, well, now
we have to do PRA on top of everything else, but I don't
think that's the right way to look at it. I think the right
way to look at is PRA helps you make judgments on what is
important in the traditional engineering analysis. You may
have been spending a lot of time trying to meet a limit that
turns out isn't very important and maybe you can be relaxed
somewhat. I think it's a way in the long run to be more
efficient and to improve what we are doing.
MR. THADANI: I would like to add to that. We
have been using risk-informed thinking in a number of ways
when some of the license amendments come in and they propose
relaxations. More and more we have tried to obtain insights
from risk assessments, to see before we grant those
relaxations to make sure we are not approving a change that
. 33
could have significant risk implications.
I think it has been done by and large in an ad hoc
manner up to now. What this does is produces the right
infrastructure, a level playing field, so to speak, not only
for the industry but the staff as well to give guidance to
both sides as to what would be a reasonable way to go
forward. Yes, in some cases that does mean additional
analyses would have to be conducted by the licensees.
CHAIRMAN JACKSON: I noted in what you sent to the
Commission you say that these documents apply for
risk-informed applications, but there is this performance
monitoring program associated with each application. So
what has to happen to make the applications both risk
informed and performance based?
MR. THADANI: We are going to cover that. That is
part of what Gary Holahan was going to cover. You had
specifically asked, I think in the last SRM, that we should
discuss that issue.
CHAIRMAN JACKSON: We will wait.
MR. KING: If I could have slide 10, please.
[Slide.]
MR. KING: Slide 10 starts with the top level or
general guidance that is in the draft general reg guide, and
it's also applicable to the application-specific reg guides.
Basically, the top level guidance is stated in
. 34
terms of five fundamental safety principles that are
intended to preserve the essential element of NRC's
regulatory philosophy, policies and practices, and to
accomplish the integration of the traditional engineering
along with the risk insights.
The five items are shown on the viewgraph. Let me
just say a few words about them.
First, licensees are expected to meet the
regulations or propose a change or an exemption if their
proposal needs such a change or exemption. We don't view
these regulatory guides and SRPs as a process to circumvent
the regulations. I want to make that clear.
Second, defense-in-depth has certainly been a
philosophy to assure safety and reliability in plant systems
and features, and certainly a way that has been used to
account for uncertainties in the past. Therefore we think
maintaining the defense-in-depth philosophy is important.
We believe that PRA can provide a useful role in
looking at the extent of defense-in-depth. We have provided
some additional guidance on what we mean by
defense-in-depth. For example, defense-in-depth is thought
of in some respects as a balance between prevention and
mitigation. Clearly PRA can play a role in trying to
quantify and illustrate is that achieved or isn't that
achieved. We don't view defense-in-depth as strictly
. 35
engineering judgment; we think PRA can provide a useful role
in assessing the extent and usefulness of defense-in-depth.
CHAIRMAN JACKSON: Is there a difference between
the staff's perspective and ACRS perspective? They speak of
maintaining the defense-in-depth philosophy.
MR. KING: I caught that on Friday. No. What
Dr. Apostolakis has said was, gee, I thought the principle
was going to say maintain the defense-in-depth philosophy.
The explanation of the principle says that, but we have
tried to keep the statement of the principle itself short,
and in the explanation you will find the word "philosophy"
in there several times. So I don't think there is a
difference.
CHAIRMAN JACKSON: Okay.
MR. KING: Safety margins have also been a
traditional part of our safety analysis. Safety margins can
be in terms of conservative methods, conservative acceptance
criteria, use of codes and standards, and so forth.
We think it's important to maintain safety
margins, although we believe that in this process of using
risk insights it's reasonable to take a look at the extent
of the safety margin: Is it above and beyond what is needed?
Is it focusing on an item that really has some risk
significance? Again, we believe risk can provide some
useful insights into adjusting safety margins to focus in on
. 36
the right things.
COMMISSIONER DIAZ: I hope that we are narrowing
down what is sufficient means.
COMMISSIONER ROGERS: That was exactly the same
question I was going to ask. Provide you a way of defining
sufficient.
MR. KING: What we say in the guide, in FSAR
analysis, for example, there are criteria that have to be
met. Part 100 dose guidelines, for example. We are not
proposing that you throw those away. Maybe a plant that
meets them with lots of margin could now meet them with a
little less margin. If it would allow some relaxation on
valve timing or something that would improve the reliability
an operability of the valve, clearly that kind of thing is
what we had in mind.
The fourth item is where we bring in the risk
insights. We are going to talk more about the proposed
metrics and guidelines that go along with that, but the idea
is to use the safety goals to try and define what level of
risk we believe is acceptable for the plants.
Finally, the fifth item emphasizes the usefulness
and importance of performance-based implementation and
monitoring strategies to assess whether the analysis and
assumptions are really coming out as you would hope they
would and there aren't any surprises.
. 37
CHAIRMAN JACKSON: How do you get at the
cumulative effect of changes?
MR. KING: We would expect licensees that come in
and propose a change and it's approved, that that would now
be factored back into their baseline PRA. So if they come
in again, their core damage frequency, their containment
performance reflects the fact that they have made this
previous change. They keep track of these things.
CHAIRMAN JACKSON: So they will have had to have
continually updated the PRA in order to get you to consider
the next proposed change based on this?
MR. KING: Basically, yes.
MR. THADANI: Yes. The guidance document says
that when they come in with the submittal, that submittal
should reflect design and operation of the plant, and if it
has undergone a change, they have to make sure that the
analysis is now consistent with whatever the design and
operation track records are.
CHAIRMAN JACKSON: That's interesting. I have
visited some plants. Admittedly what they may have in the
plant may be different than what is in the resident's
office. What you have in the resident offices many times
may be five or six years old in terms of the data on the PRA
that they have sitting in the offices. So it intersects
with what you are talking about.
. 38
MR. THADANI: Yes. In fact I am sure you are
correct. In many cases some of the studies are old and they
don't really reflect plant design and operation today, and
if they want to utilize these techniques, they have to make
sure that they update that study so that it is in fact
consistent with what is out there today. Otherwise we would
just not know where we are.
CHAIRMAN JACKSON: What does it mean, that there
has to be some relevant PRA submittal or update of the PRA
submitted?
MR. THADANI: Every time a licensee comes in and
requests a change to the licensing basis, for that
application they would have to show that the analyses in
fact do reflect the plant. That has to be done.
CHAIRMAN JACKSON: Mr. Holahan, you were going to
make a comment?
MR. HOLAHAN: I was just going to mention that I
think the staff has a previous commitment to the Commission
to keep a database of cumulative changes made in this
context.
CHAIRMAN JACKSON: Are there current plans that
exceed the Commission's safety goals today?
MR. KING: You are going to hear more about this
tomorrow.
CHAIRMAN JACKSON: The answer is yes, right?
. 39
MR. KING: I don't know if the answer is yes or
not. The answer is maybe.
MR. THADANI: Maybe.
MR. KING: The answer is maybe.
MR. THADANI: But we will be discussing it further
tomorrow.
CHAIRMAN JACKSON: You are going to be doing an
awful lot tomorrow, and the day after tomorrow. You know I
will come back on this.
MR. KING: If I could have slide 11.
[Slide.]
MR. KING: Slide 11 provides some additional
information. We recognize that in the five fundamental
safety principles there are rather important things that
needed to be factored into the guidance. We have put
another section in the reg guide that we call expectations.
Basically it's some more general guidance on implementation.
The key items from that guidance are shown on this
slide. Just a few words about those.
The licensee can do PRA and he may find out that
there are things that need to have some safety improvements
made and not just burden reduction. So we would expect an
integrated assessment by the licensees of the safety impacts
of their analyses and expect not just burden reductions to
be proposed, but, if warranted, some safety improvements
. 40
made as well.
Because these are plant-specific changes, it's
very important that the analyses reflect the as-built,
as-operated plant using plant-specific data. We express
that expectation in the reg guide as well.
The quality of the analyses in terms of is it
appropriate for the nature and scope of the proposed change,
are the appropriate models being used, appropriate data
being used?
There is some guidance in there on that as well as
the traditional quality assurance type activities: Are
qualified people doing the analysis? Are records being
kept? Is there independent verification and checks on the
analysis? That kind of thing. So there is guidance that
covers that as well in the regulatory guide.
We have defined the risk metrics of core damage
frequency and large early release frequency. I will talk
more about those. Basically they are intended to cover both
accident prevention and mitigation in terms of looking at
the risk impacts of the proposed change.
Then consideration of uncertainties is very
important. We have a fairly long section on uncertainties
in the general reg guide. It is written not as a
prescriptive cook book type guidance, but it really will
require some thinking to apply.
. 41
It talks about where does the uncertainty come
from, what are ways to assess it in terms of qualitative
factors. It does express the intent to start off using mean
values, but then you need to take a look at what the
sensitivity analyses tell you, what is in scope, what is out
of scope, and make some judgments on how you treat
uncertainty and how does it affect your decision.
CHAIRMAN JACKSON: Is the guidance clear enough?
Is everyone who uses this guidance to review licensee
submittals going to need to go through a training program to
be able to know enough to give meaning to all of these
guidelines?
MR. HOLAHAN: Yes, and we have some training
plans. The other thing I would say is I'm not sure there is
any individual who is going review these sort of complicated
issues.
CHAIRMAN JACKSON: You are going to do it as a
team approach.
MR. HOLAHAN: I think we are still thinking that a
team approach is probably most appropriate.
MR. THADANI: I think it is important that we
maintain that concept of team particularly for what I would
call the more difficult and challenging submittals. We want
to be sure that the right level of attention is given
through a team process.
. 42
CHAIRMAN JACKSON: So you will pull the teams
together as appropriate for the particular review on hand?
MR. THADANI: Yes.
MR. KING: This is an area we did highlight in the
Federal Register notice for feedback and we do intend to
continue some work on looking at the treatment of
uncertainties and possibly enhance what we have in the
regulatory guide.
CHAIRMAN JACKSON: That is an important area.
Commissioner McGaffigan.
COMMISSIONER McGAFFIGAN: I would like to ask a
question that follows up on a question I asked ACRS last
week. Are we essentially saying in so many words that in
order to take advantage of this approach you are going to
need a living PRA, and at what level, at level 2 or level 3?
CHAIRMAN JACKSON: Or scope level 1.
COMMISSIONER McGAFFIGAN: What sort of
documentation is really going to be required to work in this
area, and is it a very small number of licensees, the South
Texases, the Palo Verdes, who are going to be able to go
down this path?
MR. THADANI: I think it's going to be application
driven. If the applications are very broad scope, covering
much of the plant, then clearly one would have to have a
robust risk assessment with the right scope.
. 43
On the other hand, you can get into some simpler
applications where one could in fact rely on a risk
assessment which doesn't necessarily have a very broad scope
of information in it.
I would expect that licensees who have conducted
IPEs, essentially all of them can use some parts of it to
some level in addressing some issues. I don't know the
numbers, but probably a good number of them, if they want to
go to a very broad-based application, in-service testing or
--
CHAIRMAN JACKSON: In fact, what I was going to
suggest, if you take the pilots that you are talking about
bringing to some closure, the tech specs, the graded QA, and
ISI, IST, in-service inspection, in-service testing, how
roughly would the IPE submittals that we have fall out
relative to the criteria in terms of the potential for their
use in each of these areas? You can pick one or two.
MR. THADANI: I think with some small changes most
of the licensees should be able to utilize these studies for
changes to technical specifications. Again, it depends on
range and scope of those changes. That is one end, so to
speak. Yet, in some cases, depending on the scope of
technical specifications, we would want to make sure that
the analysis is very robust. So it would depend on what
pieces they pick.
. 44
Let's use in-service inspection as an example.
One can use these studies in a very limited way for
in-service inspection, because by and large the risk
assessments make assumptions about frequency of small breaks
and large breaks. They generally do not really discriminate
which sections, which pipes, et cetera, may be more
susceptible, which ones may be less susceptible, and thus
where should one's inspection focus be, because there is a
lot of dose commitment involved as well through these
inspections.
That means a new methodology has to be applied to
be able to discriminate among these pipes, so to speak,
various categories of pipes, and that methodology has not
been used. I may be wrong, but I don't think that has been
done in probably any of the PRAs. I hope I am right on that
one.
Westinghouse owners group and ASME are now
developing that methodology and the staff has been working
with them, so that pretty much on a real time basis we know
what is going on within the industry.
I might also note that, based on my understanding,
the monetary value is probably highest in areas of
in-service inspection and graded quality assurance.
Technical specifications could lead to substantial monetary
savings. We have some examples in South Texas that
. 45
indicate, depending on the scope, one could end up having a
fair amount of savings.
In-service testing of the pilots that we have been
talking about, if I were to rank them, are probably the
lowest dollar return, monetary return.
On the other hand, this approach we are on has a
different type of value. As we go to in-service testing, we
are not only talking about frequency of testing, but we are
also looking at the scope of testing. The two together are
important, because it could be that the testing required
today may not cover some of the more important failure
modes, which means the scope of testing has to be revised to
make sure that those important failure modes are covered
through testing.
In the end this approach may end up leading to
improved safety even if the frequency goes down, because
it's more focused and it's focused on the right failure
modes.
CHAIRMAN JACKSON: Mr. Holahan had a comment.
MR. HOLAHAN: I would like to add something to it.
This is a very important issue. I know it has gotten a lot
of attention between the staff and the industry recently. I
would say with the exception of those PRAs which the staff
sort of sent back on the IPE program, saying they needed
more work even to address the vulnerability issues, I think
. 46
all of the PRAs that have been developed can be used to a
certain extent.
If you remember back to the framework document
that the staff developed as a prelude to these guidance
documents, we talked about there being categories, as Mr.
Thadani mentioned, of some of the simpler to more complex
range of issues. I think there are numerous day-to-day type
issues that licensees can use their existing PRAs for. For
prioritizing their own work, for example. I think virtually
all the PRAs help and give licensees good insights for
making those kind of decisions.
I think all of the pilot activities we are
envisioning now can be addressed with the existing PRAs to a
certain extent. I think even those which have limitations
don't mean that they couldn't be used at all. I think the
guidance documents will allow the industry to understand and
the staff to understand some of those limitations so that
some benefit, some improvements could be made even with
limited PRA.
I think that is one of the reasons that we wrote
what I think is a rather flexible document, that invites a
range of qualitative insights to very detailed quantitative
analysis and didn't provide just a cook book that says, if
you do it this way, you pass, and if you don't do it this
way, you fail.
. 47
CHAIRMAN JACKSON: I think Commissioner Dicus has
a comment.
MR. HOLAHAN: Can I just follow up on one thing we
didn't mention, on Commissioner McGaffigan's issue?
CHAIRMAN JACKSON: Fine.
MR. HOLAHAN: You asked if a living, continuous
PRA was necessary. I don't think it's implied by this
process. I think the word that Chairman Jackson used was
"continual" updating is more appropriate in the sense that
it is updated when it's used for a license amendment and not
necessarily continuously in between. So it's sort of a once
in a while update to be appropriate to the decision that is
being made.
COMMISSIONER McGAFFIGAN: It depends what the
words "as-built" and "as-operated" mean. If it's continual
and if they are coming in for repeated amendments, then it's
going to be pretty living. If they make a change every five
years, maybe it's only every five years they have to. Is
that right?
MR. HOLAHAN: Yes, I think that's right.
COMMISSIONER DICUS: My question comes out of a
couple of things that I think you commented on. I pick up
on or hear, and I think this is what you were at least in
part addressing, that the nuclear power plant industry at
some point bought into the whole concept of PRA, and
. 48
obviously has put resources into this, as we have as well.
I am picking up and hearing now that the industry may be
less enchanted with PRA than previously, in part because
benefits that they perceived would be available at some
point in time are not being realized.
Is that accurate, and if it is, what might we do
about it? Because it's labor-intensive to us as well. If
it's not really accurate or not as close to what is really
the case as it should be, then where is this perception
coming from? I think your views on that would be useful to
me.
MR. JORDAN: Certainly it's a perception and we
have all heard it at various meetings and in discussions
with industry people. I think this guidance is now an
articulation by the staff of how the industry and the NRC
may use PRA in a wise fashion for beneficial purposes for
both industry and the regulator. I believe this is the
right answer, and now it's a matter of getting the industry
comments on this set of material and seeing how this now
fits their perception.
MR. HOLAHAN: It's clear that there is industry
frustration at the timing. I hope they are not disenchanted
with PRA as a tool. They might be somewhat disenchanted
with the staff, at the pace of our progress, but I think
that is easier to deal with than reinvigorating their
. 49
interest in the technology.
Hopefully the pilot applications that we are going
to try to get out in this month and next month and putting
the guidance document on the street may bring them back.
I think this is the right thing to do, and I think
the industry will be receptive when they see that the staff
is receptive.
CHAIRMAN JACKSON: Can you do the tie-in for us
between these documents and the pilots? Are the pilots
being evaluated relative to the criteria in the documents so
that in fact in interacting with the industry on the pilots
you are de facto getting feedback on these guidance
documents?
MR. THADANI: Yes. That is in fact what we are
doing. I would again add the industry has been very anxious
to get the documents out in the public arena for further
discussion. They have been concerned with the time that the
staff has taken in getting these documents completed.
As I noted earlier, there are some products we can
get out now. We do not have to wait much longer. For
example, technical specifications change. For example, if
we get the graded QA work completed on South Texas by the
end of June, the understanding we have based on Commission
SRM is that we will provide that information to the
Commission. Should there be some objections, of course we
. 50
will not issue these evaluation reports. Barring that
concern from the Commission, we would be able to issue the
safety evaluation reports.
I would like to think that that would be a good
signal to the industry once we get these documents out.
CHAIRMAN JACKSON: Let me make sure I understand.
Were the safety evaluation reports that either have been
done or you are saying will be done or should be done by
July done relative to the guidance that is in the guidance
documents that the Commission is considering for release to
the public?
MR. THADANI: Yes. The Commission indicated to us
that they would not review and approve issuance of those
safety evaluation reports but that the Commission would like
to see them for information.
CHAIRMAN JACKSON: Right. The point I'm asking
is, were the safety evaluation reports themselves done
referencing the guidelines in these guidance documents?
MR. THADANI: Yes, indeed.
MR. HOLAHAN: Indeed that is to a certain extent
what has taken more time on the pilots, because they started
out with a certain format and content and we have in fact
imposed on them the approach that we have in the guidance
documents here.
The one exception is the staff did approve the
. 51
boiling water reactor owners group testing program more than
a year ago. I think that was done in line with our thinking
at the time and is not quite the same scope and content as
we have here.
CHAIRMAN JACKSON: But the others are aligned?
MR. HOLAHAN: Yes.
CHAIRMAN JACKSON: Okay.
MR. KING: If I could have slide 12, please.
[Slide.]
MR. KING: Slide 12 and 13 show our proposed risk
guidelines. Slide 12 is the risk guideline for accident
prevention, which we are proposing to use core damage
frequency as the metric, and slide 13, the risk guideline
for accident mitigation where we are proposing to use large
early release frequency.
Basically these risk guidelines define the
conditions under which changes in risk would be permitted
both on an absolute scale and on a relative scale. The
absolute scale is derived from the Commission's safety goals
and their subsidiary objectives, and the relative scale from
the regulatory analysis guidelines.
In effect, what we are proposing defines the terms
"small" and "under certain conditions" which were discussed
in the Commission's January 22nd SRM.
Core damage frequency. What we are proposing is
. 52
to use on the absolute scale 10 to the minus 4th per reactor
year as the value above which further increases in risk
would not be permitted. This is the same value the
Commission endorsed for use back in 1990 as a benchmark for
accident prevention.
For the relative change we are proposing a delta
CDF or change in CDF of 10 to the minus 5th per reactor
year. That guideline is consistent with the guideline in
the regulatory analysis guidelines document. It essentially
limits changes in risk to small steps.
We think from the regulatory analysis guideline
standpoint it doesn't make sense to allow big changes,
increases in risk that would essentially be candidates for
backfit. Therefore we feel using the regulatory analysis
guidelines value is appropriate.
CHAIRMAN JACKSON: I think Commissioner Diaz wants
to ask something.
COMMISSIONER DIAZ: A relative change or each time
change?
MR. KING: This is each time change.
COMMISSIONER DIAZ: Then the total cannot approach
absolute.
MR. KING: Yes.
The other reason we think limiting increases in
risk to small steps makes sense is it provides time for the
. 53
monitoring and feedback and corrective action process to be
put in place and utilized.
MR. THADANI: Commissioner Diaz, if I may add to
what Tom King was saying. The real thrust is if we allow
one-time changes which are in this area of 10 to the minus 4
to 10 to the minus 5 frequency, then if you go to regulatory
analysis guidelines, that can become a candidate for backfit
because that is a definition of substantial improvement in
safety. We are trying to be careful that we are not
marching in a direction and then stepping back and saying,
wait a minute, we can now backfit. That is really the key
point.
CHAIRMAN JACKSON: Let me ask you a couple
questions. Essentially your discussion of large early
release frequency parallels that of core damage frequency.
So let's talk about CDFs for the moment. Does this mean
that plants with IPEs with core damage frequencies greater
than 10 to the minus 4 need not apply for any relaxations?
MR. KING: No. We think relaxations can accompany
risk decreases.
CHAIRMAN JACKSON: I understand your point.
Do any of the pilots have CDFs or LERFs such that
"increased management attention" is required?
MR. HOLAHAN: Oh, yes.
MR. KING: Yes.
. 54
MR. HOLAHAN: Most, I would say.
CHAIRMAN JACKSON: Do you anticipate that for
those specific applications the guidance documents would
incorporate what that increased management attention
process would be?
MR. HOLAHAN: The guidance documents treat the
topics in general, and they are listed in the guidance
document. It's not exactly a cook book. It's guidance as
to what issues ought to be looked at more deeply.
CHAIRMAN JACKSON: These are mean values that are
compared to the core damage frequency and to the LERF,
right?
MR. KING: Yes, mean values.
CHAIRMAN JACKSON: Let me go through here. This
is where a little bit of knowledge makes you dangerous, or
dangerous to yourself if nothing else.
As I understand the PRA process, mean values can
only be calculated if distributions are propagated through
the fault trees. That's the way I learned it.
MR. HOLAHAN: Yes.
CHAIRMAN JACKSON: So how many of the IPEs
actually propagated distributions through the fault trees?
MR. KING: Let me ask Mary Drouin, who you will be
hearing from tomorrow. Maybe she can answer that one.
MS. DROUIN: What I first say is that they were
. 55
not asked in the generic letter to do a formal uncertainty
analysis as part of their IPE. We did see that some of the
licensees did do it. My suspicion is that most of the
licensees probably did it but did not report it.
CHAIRMAN JACKSON: The real question is not so
much whether in the IPEs as done in response to the generic
letter were the distributions propagated through the fault
trees, but that in making assessments relative to these risk
guidelines will we be expecting that in those PRAs that the
distributions are propagated through the fault trees in
order to arrive at these judgments?
MR. THADANI: For those applications, yes.
CHAIRMAN JACKSON: That's what I'm saying.
MR. THADANI: For those applications, yes. They
have to come back with mean values.
CHAIRMAN JACKSON: The appropriate mean values.
MR. THADANI: Yes.
MR. HOLAHAN: However, we had said that there may
be some simple cases where the changes are so small and the
risk is relatively low.
CHAIRMAN JACKSON: I'm talking about when you are
talking about satisfying things according to what you have
on these two sheets here, that you actually have to do the
full distribution propagation.
MR. THADANI: Yes, for those.
. 56
MR. HOLAHAN: We did say there may be some cases
in which even comparison with these numbers, if they are far
enough away, if they are more than a factor of 10 away from
these numbers, that point estimates could be --
CHAIRMAN JACKSON: But you have already specified
by virtue of what you are saying what the delta CDF is and
the delta LERF is. You have already said that, right, that
that is the factor of 10?
MR. HOLAHAN: Once you are within the factor of
10, yes.
CHAIRMAN JACKSON: That's all I'm really asking.
MR. HOLAHAN: Yes, that's true.
CHAIRMAN JACKSON: In a certain sense that is what
undergirds all of this. These are probabilistic quantities.
So we can never have 100 percent certainty.
MR. THADANI: That's right.
CHAIRMAN JACKSON: Should we be saying or are you
saying that these criteria should be met with some kind of
assurance or confidence level?
MR. KING: There are some general words in the
uncertainty section of the reg guide that talk about
confidence level.
CHAIRMAN JACKSON: But you haven't really fleshed
that out?
MR. KING: We did not specify a confidence level.
. 57
You will see a comment in the Federal Register notice. We
are soliciting comment on confidence level, what are
people's views on the confidence level that these things
should be met at.
CHAIRMAN JACKSON: We talked about it coming back
as a policy issue.
MR. THADANI: Yes. That's what I was going to
say. That would be a policy issue.
CHAIRMAN JACKSON: The ACRS has proposed that the
lower tier risk acceptance criteria, the CDFs and the LERFs
again, be derived directly from the prompt fatality QHOs and
be of such value as to bound all the current sites. Does
the staff have a view on this?
MR. KING: Where we derived our LERF value was
from starting with the early fatality QHO and using
NUREG-1150 analysis and looking at if you were just to meet
the early fatality QHO, which is the most controlling QHO,
what kind of LERF would you need to have. The 1150 plants
were below the QHO; they met it with some margin. We looked
at what would it take for them to just meet it.
There was some adjustment and conservatism for the
fact that 1150 didn't cover low power and shutdown, for
example, and not all the plants included external events,
but providing some adjustment factors for that, we arrived
at the 10 to the minus 5th.
. 58
We are looking at the ACRS proposal, which I think
maybe goes into a little more detail in that. Certainly we
may want to adjust our number, but at this point we think we
are pretty close to ACRS in terms of the numbers they
proposed using their methods. So I think it's a good,
reasonable ballpark number to work with.
CHAIRMAN JACKSON: I think it's important that you
try to work to resolve this during this period that you are
also resolving other public comments.
MR. THADANI: If I may make a comment. The
Commission in an SRM -- I think it was in June of 1990 --
CHAIRMAN JACKSON: That was before our time.
MR. THADANI: -- recognized that the frequency of
large early release of 10 to the minus 5 was probably more
appropriately representative of meeting the quantitative
health objective, the prompt fatality criterion.
The Commission also recognized that there are
uncertainties in these calculations, recognized that the
selection of 10 to the minus 6 guideline value for
implementation purposes was a reasonable way to go, with
full recognition that there was probably some conservatism
in that guideline and that that level of conservatism was
appropriate.
What we are talking about now is that -- I think
that's the large early release frequency discussion that Tom
. 59
is going to go through -- we would like to hold at 10 to the
minus 6 also. However, there may be some cases where the
frequency could exceed 10 to the minus 6, and then what kind
of attention would we give to that element.
I just wanted to make sure and bring up the issue
of the 1990 SRM.
CHAIRMAN JACKSON: Let me ask you one last
question. When full scope PRAs are not available, are you
going to use something like bounding analyses to address
things like external events, fire, earthquakes, and
shutdown?
MR. KING: The approach we have taken now is if
the proposed change, for example, doesn't affect low power
and shutdown, then just a full power analysis would be fine,
but if it does, the licensee is going to have to show either
quantitatively or with some good qualitative arguments how
the risk is impacted in those conditions that aren't
explicitly modeled in the PRA.
The other thing we have done is provide in the
general reg guide an appendix that if someone has just a
level 1 PRA there is a way to estimate the level 2 results
and estimate a LERF based upon the level 1 analysis and the
previous work we have done, particularly with the 1150 and
Lasalle PRAs.
CHAIRMAN JACKSON: So you would use the level 1
. 60
analysis with some kind of a bounding analysis to get some
sense of the effect of external events on the overall risk?
MR. KING: This is for internal events only.
Where just the level 1 analysis has been done and someone
wants to estimate their LERF, there is a method proposed in
the appendix to the general reg guide that allows them to do
that.
CHAIRMAN JACKSON: I guess I am really asking you
is, how do you intend to take account of external events
within this context?
MR. HOLAHAN: What I would add is that the first
thing is we would like to have licensees submit an analysis.
That is always the easiest. I think if they wish to put a
bounding analysis, that is certainly acceptable to the
staff. In the absence of those, we will ask the licensees
to make a judgment about how that would affect their
proposal.
We haven't taken a position that it necessarily
has to be bounding. In this arena we would like to keep the
judgments closer to best estimate. Otherwise there is some
biasing about what is important and what's not.
I think we will press the licensees to address
full spectrum of issues and the staff also in a judgmental
way when there is lack of analysis.
CHAIRMAN JACKSON: That's why you need a team.
. 61
MR. HOLAHAN: I think it helps.
MR. THADANI: Let me also emphasize that that is a
very important issue, because if one were to truly go with
bounding assessment of assumptions, then clearly external
events are going to be the key. As you well know, in the
hazard functions there is large uncertainty, and if one goes
for bounding values, then those will be controlling.
CHAIRMAN JACKSON: And they have to be very
conservative.
MR. THADANI: Yes, very conservative.
CHAIRMAN JACKSON: Commissioner McGaffigan.
COMMISSIONER McGAFFIGAN: Again I am going to
return to questions I asked ACRS last week. I'm looking at
a paper ACRS gave us about shutdown operations that you are
familiar with. If you look at a BWR with a core damage
frequency of 4.1 times 10 to the minus 6, how much of that
number should I believe? Is it 4.1 times 10 to the minus 6,
somewhere between 4.1 times 10 to the minus 10? Is it a
factor of 10 or a factor of 100?
MR. THADANI: I will give you just a personal
view. Every time I see numbers like 4.1 times 10 to the
minus 6 my immediate conclusion is that there is much
greater perceived precision than there really is in these
analyses and calculations, particularly when you go to
shutdown conditions where the majority of the contribution
. 62
is coming from human errors.
CHAIRMAN JACKSON: You have to be careful, though,
because if you are talking about starting with a core damage
frequency of one 10 to the minus 4 and you talk about delta
core damage frequency that is 10 to the minus 5, you are
talking about going from 1.0, 10 to the minus 4, to 1.1, 10
to the minus 4, right?
MR. THADANI: Absolutely, yes. I think we can
come back to this issue for confidence in delta.
COMMISSIONER McGAFFIGAN: I'm going to get to
that. As I told him last week, it's a matter of arithmetic
why you have greater confidence in deltas than you do in the
total, and I understand that, but that gets to the delta
question. If I believe that this plant, whatever it is, a
Mark 3 BWR, is at 4.1 times 10 to the minus 6, and now I say
in the risk guidelines you can make changes of up to 10 to
the minus 5 in core damage frequency, then I'm making a
factor of a 2-1/2 change in that, if I believe any of this.
Is that a small change in risk? When you guys chose 10 to
the minus 5 as the delta, did you think about 10 to the
minus 6 as an alternative?
MR. THADANI: Yes. In fact, what we are saying is
by and large many of the changes actually are going to be
below delta of 10 to the minus 6. If you look at general
license amendments, most of them are not that significant.
. 63
We are saying 10 to the minus 6 delta is a fairly small
change.
CHAIRMAN JACKSON: I think he's saying something
else. If you start low, are you going to allow a factor of
10 to 100 increase?
MR. THADANI: Yes. I'm saying now you go up to 10
to the minus 5, which is an appreciable change. We are
saying we are going to have to look at a number of factors
before we say, yes, indeed, go ahead.
One issue we said we would take a very hard look
at is going to be the issue of uncertainties. The other
issue that we are going to take a very hard look at is, does
it really potentially bypass two barriers? During shutdown
condition, if it's a boiler, very likely the containment is
open. So we have got to be very careful, because now we are
are talking about delta CDF as well as potential for perhaps
a significant release.
So one has to integrate all those issues as one
goes to deltas, which are now appreciable. Ten to the minus
5 delta CDF, in my view at least, for a change through an
amendment process is a very significant change.
COMMISSIONER McGAFFIGAN: That is my impression as
well. Why not a number 2 times to the minus 6 rather than
10 to the minus? You are saying 10 to the minus 5 is where
there will be more analysis.
. 64
MR. KING: Actually it starts at 10 to the minus
6. We are now within a factor of 10 of the value shown on
the viewgraphs. You go into the more analysis tension
region. I think that region is intended to address the
concern you are expressing.
MR. HOLAHAN: There has been a lot of discussion
on this issue. In the industry guidance to themselves as to
how to use the PSA applications guide they chose to give
their guidance in terms of percentage of the current value.
In other words, if you were at 10 to the minus 6, 10 percent
of 10 to the minus 6, not 10 percent of the safety goal
subsidiary objective.
So we had considerable discussions among the staff
and with the ACRS as to should changes be measured with
respect to where you think the plant is or with respect to
your safety guidance values. We came around to saying that
it's more important to believe your speed limits than to
just deal with the changes. I think that means you are
treating the 10 to the minus 4's and 5's and 6's as though
they matter more than in effect penalizing a plant that is
very safe and saying it can make almost no changes.
MR. KING: I think we probably talked about most
of the material on slide 13. So let me propose to go on.
[Slide.]
MR. KING: In slide 14, all I wanted to do there
. 65
was illustrate the areas where the application specific
guidance supplements the general guidance. I don't intend
to go into those in any detail, but I did want to point out
that these are unique areas, that you will find discussion
in the application specific guidance that you won't find in
the general guidance.
CHAIRMAN JACKSON: So all of these are part of
what we already have?
MR. KING: That is all part of the package of what
you have.
COMMISSIONER ROGERS: Before you leave that I do
have a concern, and that is that whether any of these really
represent something that's in a rule or the equivalent of a
rule and therefore using a PRA analysis strictly speaking
might take one out of compliance with that rule. For
instance, where any ASME codes are involved and referenced
by rulemaking such as test intervals or something of that
sort. How do you propose to deal with that?
MR. THADANI: There is under 50.55(a) an
alternative approach option that the Director of NRR can
approve. That is indeed what we indicated as one of the
policy issues and indicated that is the path we go on in
terms of in-service testing. I believe the Commission
approved that path.
You are quite correct. Otherwise it could mean
. 66
change the regulation.
COMMISSIONER ROGERS: In every case is there some
disclaimer of that sort?
MR. THADANI: Yes.
MR. HOLAHAN: In fact it's our first principle,
that you meet the regulations or you get an exemption or we
have a rule change.
CHAIRMAN JACKSON: Therefore any of the guidance
that comes out of here is not going to conflict.
MR. THADANI: That's correct.
MR. HOLAHAN: In fact it ought to contribute to
convergence between compliance and safety issues.
MR. KING: The final thing is not a viewgraph, but
Ashok had mentioned in the beginning that we owed you a
short update on the human performance and reliability
assessment plan that you asked for last Friday.
As you recall, ACRS suggested we need such a plan.
So did our Nuclear Safety Research Review Committee. We
agree. We have responded to ACRS that we plan to have such
a plan ready for review by the end of June.
That plan is going to cover human performance and
human reliability aspects for both reactors and materials
facilities. It is going to be based on an integrated model
of human performance; it is going to deal with activities
related to events assessment, inspection, design; it's going
. 67
to cover the database question; and it's going to talk about
where do we get the data, both domestically and
internationally, both nuclear industry and applicable data
from outside the nuclear industry.
Our schedule is to have that plan available to be
given to ACRS the end of June. We are having a subcommittee
meeting with them June 3 where we will give them a status
report and discuss it in viewgraph form.
We also plan to meet with ACRS later in the summer
and eventually request a letter from them. We also plan to
meet with the Nuclear Safety Research Review Committee on
this.
Ultimately we hope to have it finalized and we'll
provide it to the Commission by the end of September.
What we are not waiting for is to move out on the
agency database question. We recognize that across the
agency we have several databases. It probably would be more
efficient to get together and have a common database. We've
had a kickoff meeting among the offices to start that
activity, to identify what are our data needs, what data do
we want to put in there, what's the quality of the data we
need, and we hope to begin implementation of that by the end
of September.
In a very short fashion, that is what we plan to
do in that area.
. 68
MR. THADANI: Gary.
[Slide.]
MR. HOLAHAN: On slide 15 there is a discussion of
performance monitoring. Back in the January 22, 1997, SRM
the Commission asked for a summary discussion of performance
monitoring in the context of both the pilot applications and
the guidance documents.
The guidance documents do have sections addressing
performance monitoring. It is one of the four key steps
that Tom King mentioned. In fact it's the third step. It's
covered by discussion in section 2.5 of the regulatory
guide, and there are corresponding sections in each of the
other reg guides and standard review plans.
There have been discussions between the staff and
the pilot applicants on the issue of performance monitoring.
Those are along the same lines as we have presented in the
guidance documents.
In the staff's report and even up until today the
only document that we have really taken a final position on
this issue is the CE owners group lead plant, the Arkansas
tech specs.
In effect this issue is still in the review
process for the graded QA and the IST pilot applicants. We
have asked them questions and we are pursuing the issue
consistent with the guidelines in the reg guides and the
. 69
SRPs.
CHAIRMAN JACKSON: How do the guidelines for
performance monitoring here compare with the guidelines
under the maintenance rule for performance monitoring?
MR. HOLAHAN: In our guidance documents and in the
pilot applications, the Arkansas one as an example, we say
that that the maintenance rule is the expected starting
point for the licensee in their performance monitoring
activities.
There are a few differences between what is
monitored under the maintenance rule and what would apply to
a specific application. One is that the maintenance rule
calls for monitoring in the context of maintenance
activities. So what they count, for example, is
maintenance-preventable failures. That may or may not be
sufficient for a given application. We may be interested in
other type of failure mechanisms.
In practice many licensees are keeping a broader
set of data even under the maintenance rule than just
maintenance-preventable failures. As the data rule or
voluntary approach to reliability data moves ahead, we are
seeing that the industry will be developing a database one
way or another for addressing these issues.
The other thing that the maintenance rule differs
from some applications is that for low safety-significant
. 70
systems the monitoring in the maintenance rule is usually
done on a plant basis and not on a component reliability
basis. Since many of the applications we are talking about
are making changes, reducing requirements for the low
safety-significant systems, the monitoring we are talking
about is making sure that those systems with reduced
requirements don't become significantly less reliable than
was expected.
The maintenance rule as it's currently written
doesn't necessarily provide component or even system level
information. So when we come to a specific pilot,
performance monitoring on that application would either
require reliability or availability information, depending
upon what sort of pilot application it is.
Some of that might be available through the
maintenance rule. But we see in most cases is you probably
have to stretch the amount of data that is kept from the
maintenance rule. It's done similarly, but I think probably
a little more data has to be kept.
For example, in graded QA the concern is, with
less quality assurance, is it possible that the equipment is
becoming less reliable? So some sort of reliability data is
the check to see whether that's happening or not.
In contrast to that, under the technical
specification amendment in which longer outage times are
. 71
being allowed, what we are interested in checking is seeing
whether those longer outage times are contributing to
increased inappropriately large unavailabilities. So the
monitoring approach is tailored to the individual issue.
If we go to slide 16, it discusses the specific
example of the technical specifications in ANO 2. In that
evaluation there is a specific section in the safety
evaluation report parallel to what is in the guidance
documents addressing performance monitoring.
As I mentioned, equipment availability is the
concern with increased allowable outage times. That is
written into the safety evaluation report and it's tied to
the maintenance rule for the specific equipment that has
changes.
The safety evaluation report has basically not
only performance monitoring, but a corrective action
section, in which case corrective action through the
maintenance rule would look at whether the technical
specification is contributing to an inappropriate amount of
unavailability, in which case it would be addressed in the
context of the maintenance rule.
This is done on a two-year basis in looking at
reliability and unavailability, and if those numbers are
exceeding the balance or the goals that licensees have
established, then we would either consider rewriting the
. 72
technical specifications to pull that back or look at what
other actions ought to be taken to address that issue.
This is the only case in which we have actually
written when the Commission approves it, which would be an
approved example. I think it establishes a general format
that will be used in other cases, but since graded QA may be
the next example to come by, I think we will see emphasis on
equipment reliability data as opposed to availability in
that case.
CHAIRMAN JACKSON: Commissioner McGaffigan.
COMMISSIONER McGAFFIGAN: In the paper you sent us
on ANO 2's proposed change you say at some point here that
in approving the proposed tech spec changes the staff is
relying on a commitment made by the licensee with respect to
utilization of a risk-informed configuration control
technique to assess the risk associated with removal of
equipment.
Are we essentially changing the "should" to
"shall" in the maintenance rule with regard to configuration
control by having this license condition or administrative
control and tech specs put into this license?
I don't know how broad the configuration control
is going to be, but if they have a risk-informed
configuration control system, that is the "should" versus
"shall" issue in the maintenance rule.
. 73
MR. HOLAHAN: I think what it says is this
licensee has made a commitment to have a program.
COMMISSIONER McGAFFIGAN: Since this is a
precedent, we will expect similar commitments from other
licensees as they come in, and for that category of
licensees the maintenance rule "should" has converted to a
"shall."
MR. HOLAHAN: I think it's not quite converted.
In terms of enforcement against the rule versus enforcement
against this particular license amendment, I think there are
a little different implications. But I think it does move
it into a regulatory requirement of some sort.
MR. THADANI: Our focus as we are conducting these
reviews is to make sure that prior to allowing relaxation
that we have taken an integral look at safety implications
of the change. We believe configuration control is very
important because of the way risk analyses are traditionally
done. We have indicated as a condition of approval that
that control has to be maintained if this relaxation is to
be granted.
You are exactly right. I have had calls from the
industry, very unhappy with the staff at taking that
approach, and why is this not covered under Part A3 of the
maintenance rule, which industry, as I was told, considers
is a requirement.
. 74
I know you have asked us to take a look to see if
we should revise Part A3 of the rule, to change "should" to
"shall," and I hope we will come back to you very quickly
with a recommendation. Quite frankly, the interaction that
I have had with NEI, they have indicated to me that they
would support changing Part A3 of the rule from "should" to
"shall" if that resolves this issue.
CHAIRMAN JACKSON: Do you want to vote it this
afternoon?
COMMISSIONER McGAFFIGAN: I think you may have at
least one Commissioner who is receptive. Having seen this
paper, it struck me in a machiavellian sense that one reason
you answered the question in a more ambiguous way when first
asked is that you have these other methods to catch
licensees' attention and you end up converting the "should"
to "shall" anyway, so we might as well just do it up front.
MR. THADANI: We think it's an important safety
issue and it ought to have some enforcement capability
behind it.
CHAIRMAN JACKSON: And this is how you are doing
it for now until you come back to us with a specific
recommendation fast.
MR. THADANI: Yes.
CHAIRMAN JACKSON: Okay.
MR. HOLAHAN: The last thing I would like to cover
. 75
on performance monitoring is to go back to Chairman
Jackson's question about what does it take to be fully risk
based as opposed to being what we have called risk-informed
with performance elements built into it. I think there are
really two differences between what we have done here and
what would be a fully risk-based approach.
The first is, frankly I'm not sure what a
risk-based approach is. If you ask 100 people, you might
get 100 different answers. So I think there is some
development work to be done. The staff has an assignment to
get back later this year to address that issue more fully.
But there is another issue involved in it, and
that is we are making license amendment in the context of
the existing regulatory framework. To be fully performance
based, I think you would have to break out of part of the
approach. We are still using Appendix B and the
programmatic elements of that; we are still using technical
specifications; we are still using a staff review and
approval process; and I think all of those things might be
changed to some extent in a fully performance-based program.
Within the context of these sort of measured
steps, as Dr. Ross has mentioned, I don't think we can
become fully performance based without changing some of the
other paradigms.
The last thing I would like to cover is future
. 76
activities on slide 17.
[Slide.]
MR. HOLAHAN: I think much of this has already
been mentioned. The staff would hope to issue the guidance
documents in May, based on Commission guidance.
You will note that the package does not include an
ISI program. We are looking towards getting the reg guide
and SRP on ISI in July of this year.
I think Tom King already mentioned the workshop,
and I think Mr. Thadani did a pretty detailed job of going
through the status of the pilots.
One thing that I would mention. On the slide
where it says graded QA, 12/97, that really applies to the
three pilots, South Texas, Palo Verde, and Grand Gulf. We
really are hoping to get the South Texas piece of that done
end of June, early July, something in that time frame, in a
much faster time frame than December.
[Slide.]
MR. HOLAHAN: On slide 18, this is just to remind
the Commission that there are a few more IOUs from the
January SRM. A number of these subjects are covered in our
guidance documents.
In part, we will get public comment on those
before we come back to the Commission. In addition, there
are questions for OGC about the legal implications of some
. 77
of these that are also needed in responding to that
Commission guidance.
I think we mentioned earlier that in fact there is
some training planned for the staff on the reg guides and
the standard review plans to help that process along.
Following the public comment period there will be
a resolution process; there will be a series of meetings
with the ACRS; there will be a second round through the
CRGR; and we are still hoping to and are committed to
getting the general guidance, the tech specs, IST and graded
QA completed by the end of the year.
We are hoping to get in-service inspection done by
February, but I think that date is somewhat dependent upon
there being a pilot application by September. So I think
that date is less certain than the others. We will have a
number of opportunities to discuss that with the Commission
well before that date and we will know more about how that
is going with respect to a pilot application and progress on
the guidance documents.
I think that's all we have for our presentation.
CHAIRMAN JACKSON: Thank you.
Commissioner Rogers, any follow-on questions?
COMMISSIONER ROGERS: Just two. One involves
in-service testing. Do the failure rates that are being
used for some pieces of equipment that are subject to
. 78
in-service testing depend on the rate of testing?
MR. THADANI: I was looking around to see if the
specific staff member is here or not. The intention is to
look at that issue specifically as part of our evaluation
process. There are two key elements. I only touched on
one. The second one is the one you mentioned. If you
change frequency of testing from, let's say, every month to
every year, you may introduce some new failure modes that
one may not have.
COMMISSIONER ROGERS: Or you may reduce the
failure.
MR. THADANI: Absolutely correct.
MR. KING: That is one of the items in the IST reg
guide that has to be specifically addressed. That is one of
those supplemental items you won't find in the general reg
guide, but your specific question is in there.
COMMISSIONER ROGERS: I think for some equipment
it is really very important. The value of a reduced
testing, if out of a PRA the conclusion comes that a testing
rate could be reduced, then you may even get a double
benefit there. Not only an economic benefit. You may
actually a real safety benefit from that.
MR. HOLAHAN: And I think these guidance documents
provide a road map for the licensees to take those issues
and present them to the staff in a way that we would be
. 79
receptive to change.
COMMISSIONER ROGERS: The other question involves
the quality of the PRAs. We know they are of uneven
quality, and yet the approach so far that we have heard
about here is one that doesn't seem to specifically take
that into account. I wondered to what extent you are
thinking of somehow or other imposing something that
provides a uniform standard here if one is going to apply
these constraints on deltas and LERFs.
MR. KING: We have a long-range goal to look at
standardization. At this point we think maybe that draft
NUREG-1602 is a good start toward a standard for PRA
quality. In fact, a couple of the items in the Federal
Register notice soliciting feedback has to do with the use
of that as a standard or any other suggestions for what
could be a standard.
You are right. At this point we haven't required
certain attributes or certain scope and depth of a PRA. It
is sort of up to the licensee to come in and justify. But
our long-range goal is to head in that direction.
MR. THADANI: There is a very strong
recommendation in the guide for independent peer review,
which I think is an important element in addressing quality
as well. It is strongly encouraged throughout the guide as
well as when you go to quality assurance section that the
. 80
independent peer review can go a long way towards satisfying
the intent behind Appendix B of quality analysis.
COMMISSIONER ROGERS: When we started in on this
IPE process we didn't really think it would ever take us as
far as we are today. So now we have to look at what the
quality is, it seems to me, if we want to use them.
CHAIRMAN JACKSON: Are we tracking regulatory uses
of IPEs?
MR. THADANI: I have to make sure that this is
correct, and I will need help. As part of the
implementation plan, every time we make use of individual
plan examination and regulatory decision we are supposed to
keep track of it. I will confirm that in fact we are doing
that.
CHAIRMAN JACKSON: Please do.
Commissioner Dicus.
COMMISSIONER DICUS: One question regarding the
concept of current licensing basis and the application in
this program that we are in now. I don't want to go back
into what is current licensing basis. I recall from
previous briefings and meetings there has been lengthy
debate and discussion over how you use something that is
undefined. In the applicable regulations it is only defined
in Part 54 with license renewal. We have been through that.
In these applications for probabilistic risk
. 81
assessment and in submittals that licensees might be making
do you have any plans to actually use the definition of
current licensing basis in Part 54 for this?
MR. THADANI: What we are saying is what is within
the scope. We are not suggesting with this that one needs
to compile all this information. However, if there is an
issue that impacts those elements that are within the
current licensing basis, the licensee's proposal to make
changes in that element has to cover both aspects,
deterministic and probabilistic.
We are not suggesting in this guidance that one
needs to compile current licensing basis information. I
think that was the more difficult issue, who is going to
compile this information. The scope of the risk assessment,
we are not suggesting that changes as a result of this.
COMMISSIONER DICUS: I think the answer to my
question was maybe. I'm not sure I heard yes or no, but I
think that point needs to be made very clear, particularly
to licensees.
MR. HOLAHAN: In most of the applications we have
seen this hasn't turned out to be a problem. When you are
writing a general guidance document to try to cover all
future type applications, we needed some way of describing
sort of the scope of all possibilities, and current
licensing basis is kind of shorthand for doing that. If you
. 82
look at the actual examples, ISI and IST and graded QA,
these are areas where the licensees understand what their
licensing basis is and their need for a license amendment.
COMMISSIONER DICUS: I just don't want the
regulatory guide to begin to confuse the issue. We should
clarify the issue.
CHAIRMAN JACKSON: Commissioner Diaz.
COMMISSIONER DIAZ: The main question I have
Commissioner Rogers, using seniority, already asked.
CHAIRMAN JACKSON: We will go in reverse. You
will move up the queue.
COMMISSIONER DIAZ: I don't have any questions. I
just want to say that I am very pleased that we have gotten
to this point. I think it's a very, very great step, and I
certainly commend you.
CHAIRMAN JACKSON: Commissioner McGaffigan.
COMMISSIONER McGAFFIGAN: On the time line on the
last chart, 12/97 you hope to have final reg guides out.
You have a 90-day comment period. That will take you into
August. Do you expect there to be significant comments and
policy issues that will then have to be resolved? Is that
period between 8/97 and 12/97 optimistic?
CHAIRMAN JACKSON: A drop dead date.
MR. THADANI: It's a drop dead date that we have
been working towards. If you look at the set of questions
. 83
in the Federal Register notice, they are very tough issues
and a number of them are really policy issues. I would
expect that we would end up having probably at least two
separate meetings with the Advisory Committee on Reactor
Safeguards and extensive discussion with CRGR, and very
likely we may have to come August-September time frame to
the Commission to seek guidance on some of these issues.
Example. What confidence level one must ascribe. Is it 80
percent? 95 percent? Whatever it is, we will come back to
Commission.
CHAIRMAN JACKSON: You could be doing some
parallel processing.
MR. THADANI: Yes, and in fact we are going to be
doing that. But I think it is a very, very tight schedule.
CHAIRMAN JACKSON: It's ambitious, but at the same
time we have waited too long to get to this point.
COMMISSIONER McGAFFIGAN: I'm anxious to get to
the concluding point too. There will be a lot of, as we are
coming to call it, parallel processing going on if you are
actually going to get to that point.
MR. HOLAHAN: We hope that the workshop we have in
July will provide us early public feedback that we can start
working on. That should be helpful.
CHAIRMAN JACKSON: I'd like to thank the staff for
a very informative briefing. As you can tell by how much
. 84
time we have taken on the agency's PRA activities and as you
have heard, we do commend you for the progress you have made
to date and for being responsive on developing these
documents and working on the pilots.
I know it has been sometimes a difficult area, but
at the same time we encourage you to continue to improve the
process and to provide appropriate review mechanisms, both
internal and in terms of external reviews to ensure that we
appropriately use PRA. It is becoming an important tool in
support of the regulatory process. So we need to enhance
the process where necessary, but, as you've heard, to ensure
its consistent use where appropriate. I will just call out
one or two of those.
For instance, we discussed that relative to the
use of the reg guides and standard review plans in the
pilots.
We talked about performance monitoring in the
pilots compared with performance monitoring in the
maintenance rule.
We talked about the implications of all of this
for risk-informed configuration management in plants.
As you heard, relative to the definition of
current licensing basis as defined in Part 54 and what that
suggests relative to what we need to do in Part 50.
I want to especially commend you for your work in
. 85
producing these documents. I had asked you to do them
within a certain time frame. The schedule slipped a little
bit. We understand that. As Commissioner McGaffigan said,
it's still ambitious, but we are still aiming for 12/97. So
you should continue your efforts to complete in a timely
manner the pilot applications of risk-informed regulation
and to complete these draft guidance documents, particularly
the ones for in-service inspection, on the time line that
you have mentioned.
You should also evaluate the proposed decision
criteria. You spoke to this yourself, Mr. Thadani. And the
rationale for assuring conformance to those criteria. You
need to develop additional guidance on acceptable approaches
for confirming the assumptions and the analyses that are
conducted to justify current license-basis changes. As we
have discussed, this would include consideration of the role
of uncertainty.
We look forward to getting some recommendations in
the policy areas relative to the appropriate confidence or
assurance levels in the use of PRA for decision-making as
well as the development through the pilots of any additional
guidance that is needed on this increased management
attention process.
Unless my fellow Commissioners have any additional
comments, we are adjourned.
. 86
[Whereupon, at 4:15 p.m., the briefing was
adjourned.]
|
Privacy Policy |
Site Disclaimer |