ML20210H266
| ML20210H266 | |
| Person / Time | |
|---|---|
| Issue date: | 05/28/1997 |
| From: | Apostolakis G Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-3034, NUDOCS 9708130327 | |
| Download: ML20210H266 (18) | |
Text
CERTIFIED BY:
'd Geo'rgp Ap'octolakte - 5/28/97 kg $3y 1D' Th/9 ADVISORY COMMITTEE ON REACTOR SAFEGUARDS JOINT SUBCOMMITTEE MEETING MINUTES:
PROBABILISTIC RISK ASSESSMENT AND PLANT OPERATIONS OCTOBER 30 - NOVEMBER 1, 1996 ROCKVILLE, MARYLAND The ACRS Subcommittees on Probabilistic Risk Assessment (PRA) and Plant Operations met on October 30 November 1,
1996, at 11545 Rockville Pike, Rockville, MD, in Room T-2B3.
The purpose of this meeting was to continue the joint Subcommittee's review of NRC programs for risk-based analysis of reactor operating experience and to discuss the staff's approach to codify risk-informed, performance-based regulation through the development of Standard Review Plan (SRP) sections and associated Regulatory Guides.
The entire meeting was open to public attendance.
Mr. Michael T.
Markley was the cognizant ACRS staff angineer for this meeting.
The meeting was convened at 8:30 a.m.
each day, recessed at 4:42 p.m.
on October 30, recessed at 5:50 p.m.
on October 31, and adjourned at 3:00 p.m. on November 1,.1996.
Dr< Apostolakis stated that the Subcommittees had received no written comments.
However, Mr. Jerry Eisenberg of the American Society of Mechanical Engineers (ASME) had requested time to make a presentation regarding the proposed ASME Code cases for risk-based inservice testing (IST) and inservice inspection (ISI).
M.TENDEES ACRS G. Apostolakis, Chairman D.
Powers, Me-ber R. Seale, Co-Chairman R.
Savio, ACRS/ACNW Staff I. Catton, Member M. Markley, ACRS Staff M. Fontana, Member R.
Sherry, ACRS Fellow D. Miller, Member J.
Garrick, ACNW Member I[
NRC Staff
- Q(I E. Jordan, AEOD R.
Jones, NRR P. Baranowsky, AEOD M.
Cunningham, RES S. Mays, AEOL T. King, RES D. Rasmuson, AEOD A.
Ramey-Smith, RES P. O'Reilly, AEOD J. Murphy, RES G. Holahan, NRR T. Hiltz, NRR D. Wessman, NRR M.
Rubin, NRR gg*!llfill]?g!gd II,ll p
G. Bagchi, NRR M.
Check, NRR
. l0 V 3
pse-
.w' AEOD - Office for Analysis and Evaluation of Operational Data NRR - Office of Nuclear Reactor Regulation RES - Office of Nuclear Regulatory Research
()f, sif ~ l fhKS h 47 r-6*
ptc1ca.TED m cu n
~
[
fogaggg7 meu cy u u By_
' ' ~ ~
3034 PDR
1 Joint PRA/ Plant Ops. Subete.
2 Minutes 10/30-11/1/96 Industrv/Public K. Balkey, Westinghouse B. Bosnak, ASME*
A. McNeill, Va. Power Co.
J. Pelletier, Yankee Atomic T. Cannon, Palo Verde G. Eisenber J. Perry, ASME*
S. Hartley,g, ASME W. Rowley, The Wesley Corp.
Lockheed-Martin G. Zigler, ASME*
T. Mawson, Northeast Utilities ASME - American Society of Mechanical Engineers A complete list of meeting attendees is in the ACRS Office File and will be made available upon request.
and handouts used during the meeting are attached to theThe presentation slides copy of these minutes.
office OCTOBER 30, 1996 Openina Remarke Dr.
George Apostolakis, Chairman of the Joint Subcommittee, i'
convaned the meeting and noted that Dr. Seale would serve as Chairman for those portions of the meeting for which Dr.
Apostolakia had a conflict of interest.
Dr. Apostolakis stated that the purpose of today's meeting was to continue the Subcommittee's review of risk-based analysis of reactor operating experience.
He noted that review this matter on July 17,the Subcommittee had previously met to 1997.
I He noted that the Subcommittees had received no written comments or requests for time to make oral statements from members of the public for today's meeting.
AEOD Presentation Introduction - Mr. Patrick Baranowsky, AEOD Mr.
Baranowsky summarized previous discussions with the Subcommittee, the mission of AEOD, and the objectives of the briefing.
From previous ACRS meetings, he identified the following issues:
the process for feedback of the results of the AEOD studies to the risk assessments and regulatory decisions,
.~
Joint PRA/ Plant Ops. Subete. 10/30-11/1/96 Minutes the interf ace with ACRS on new methods development and results of ongoing analyses, details of the Accident Sequence Precursor (ASP) Program, and e
details of the Common-Cause Failure (CCF) Program database.
He outlined the Reliability and Risk Assessment Branch (RRAB) activities, plans for risk-based analysis of reactor operating experience, program
- elements, accomplishments, and planned activities.
Mr. Baranowsky provided examples of the AEOD feedback process and products, such as, the results of system reliability studies.
The Members of the Subcommittees and the staff discussed the following issues:
acceptability of models used in the ASP Program, e
the staff approach for rational decomposition of risk, e
definition of a good success criteria, e
role of defense-in-depth in AEOD analyses, and use of risk insights by the Technical Training Center instructors and NRC inspectors.
Risk-Based Performance Indicators Mr. Steven Mays, AEOD Mr. Mays presented the NRC use of performance indicators and the project for increasing the objectivity and consistency with which the NRC uses these indicators.
He explained that the NRC was moving toward developing risk-based indicators by decomposing risk into constituent p, arts, collecting data related to those parts, and analyzing the data to develop indicators and trends.
He illustrated this transition by comparing current and future indicators for reactor scrams, safety system actuations, forced outages, significant events, radiation exposure, and cause codes.
Mr. Mays explained that human performance issues were accounted for by measuring initiating events, system reliability, and common-cause failures (CCFs).
He noted that predicting operator response to transients and accidents, and measuring the impact of organizational factors are issues being studied by RES.
The Members of the Subcommittees and the staff discussed why performance indicators are needed, the completeness of the current set of indicators, and how to design a
set of risk-based 3
m 44 4A durm
^&
o*A W-a
- i Joint.PRA/ Plant Ops. Subete.
-4 10/30-11/1/96
~ Minutes indicators.
The Members suggested that the staff focus on selecting indicators based on - a knowledge of risk and what the
' indicators are telling the analyst.
C---
,n-Cause Failures
_Dr. Dale Rasmuson, AEOD Dr. Rasmuson described the common-cause failures (CCF) database and the associated analysis software.
He presented the number of records, which have been analyzed, and summarized the events in the database.
He described a CCF record, data quality, and the key assumptions-used in preparing reports.
Dr.
Rasmuson provided the following responses to previously identified ACRS issues:
J Issue A key assumption of the model is that the layout of the components is irrelevant.
Response: The statement is incorrect.
The model and rev tw of data are explicitly tailored to account for potential impacts of a specific CCF event, given a particular layout and other arrangements at other plants.
Issue:
Is there sufficient information in the database to allow licensees to exclude events from the analysis?
Responae: Yes, the database includes a listing of each CCF event for an application along with the. source data text and information on how the event was coded for verification or modification.
Issue:
Is it necessary to use CCF events that are not complete failures of redundant equipment?
Response: Barriers and defenses that prevente'd actual failure may not exist at every plant.
Challenges to barriers and
=
defenses are not trivial.
The CCF method accounts for
~
challenges by allowing data ana'ysts to assign a
probability to the degree of coupling based on actual conditions and judgment.
Exclusion of events involving partial failures would underestimate CCF probabilities for predicting future performance and does not represent the " state of knowledge" expectation.
Issue:
Has any licensee ever gone back to the database to estimate CCF parameters?
Response: No, the database has not been released.
Database results have, however, been used in ASP evaluations.
4
.-c t
Joint PRA/ Plant Ops. Subete.
-S-10/30-11/1/96 Minutes Issue:
Another key assumption is that only similar components in the - same system are susceptible to CCFs.
How about similar components in different systems?
Response
Probabilistic Risk Assessments (PRAs) and Individual Plant Examinations (IPEs) treat intrasystem CCFs for similar components. Analysts used the database to search for CCFs between specific systems.
More detailed searches are required.
However, additional searches are resource intensive and are of lower priority.
Issue:
A disturbing f act of the elaborate models and analyses is that several subjective inputs seem to be buried deeply in the analysis and that the average user has no way of knowing that they are there.
Responses-Subjective judgments are made in the data collection effort.
The judgments are well documented and supported by a
well-defined
- process, which includes quality assurance.
Issue:
In general, the mapping-up procedure is subject to larger uncertainties and is not quite as standardized as the mapping-down procedure.
Response: The mapping-up-procedure is nondeterministic in nature.
Analysts need to account for the potential impact of a CCF in a small group on a larger group being modeled.
To compensate analysts use a binomial failure rate model.
Also, analysts can use the evidence contained in the description of the failure event to estimate a mapping-up parameter.
Other mapping-up methods are under develop-ment.
Issue:
The data-averaging approach accounts for the uncertainty in the data in an ad hoc manner.
In the case of the variance and the correlation, the data-averaging approach systematically underestimates the uncertainty.
Response: Uncertainties treated in the current version of the database are associated only with the sample uncertainty about the central estimate.
Other sources of uncer-
- tainty, such as model uncertainty - and plant-to-plant variability, have been recognized and are being addressed by.recently developed analysis software.
Issue:
Not surprisingly, the different prior distributions do lead to different results.
5
+
Joint PRA/ Plant Ops. Subete.
-6 10/30-11/1/96 Minutes
Response
Prior distributions can change and influence results.
The staff obtains data from different plants to in.rease the amount of data available for use in CCF analysis.
The staff uses engineering knowledge and information to increase the quantity of applicable data by tailoring the data to the specific plant and thereby helping to reduce the impact of the prior distribution.
Issue:
There are several issues raised by the ACRS related to the general treatment of uncertainty.
Response: Statistical or sampling uncertainties are treated in the database.
Data classification variability has been minimized by providing analysts guidance in making judgments and by including multiple layers of review.
Uncertainty in impact vector assessment is an area that should receive further attention.
Plant-to-plant variability is handled by appropriate statistical methods by the database software.
Issue:
Since the processing of these data is so full of hidden assumptions and subjective evaluations, considering the project as being data-based can be very misleading.
Response: The assumptions are defined and discussed in INEL-94/0064,
" Common-Cause Failure Data Collection and Analysis System," Volume 3,
" Data Collection and Coding Common-Cause failure Events,"
and Volume 5, " Guidelines on Modeling Common-Cause Failures in Probabilistic Risk Assessments."
Each event is coded according to the guidance provided in the documentation.
Issue:
What is the meaning of the " generic" values of alpha f actors reported in INEL 94/0064, Volume 6, " Common-Cause Failure Parameter Estimations"?
Response: Generic implies that values belong to a group or pooled population set and not to a generic plant.
The generic values are intended to be broad evaluations of industry experience that can be used by regulators and-industry to obtain a general indication of the nature and magnitude of CCF events and to assess industry trends.
Issue:
The presence of significant subjective evaluations in processing the data is confirmed by a random sample of incidents reported in INEL-95/0035,
" Emergency Diesel Generator Power System Reliability 1987-1993."
How are the interpretations of these incidents to be explained?
6
Joint PRA/ Plant Ops. Subete. 10/30-11/1/96 Minutes t
i Are they inadvertent errors?.How can we know that they j
are not representative of a poor job?
4 Response: The random sample of incidents in the INEL report a
describe only some CCF events in the _ database, and include several dependent events related to emergency diesel generators that are not included in the database.
Information from the complete licensee event report is needed to correctly understand the full nature of the events.
e Issue:
The finding during post-maintenance testing at Indian-J Point 3 that service water valves were inoperable due to j
improper maintenance is odd.
Response
The licensee stated in the licensee event report that the cause of the valve failure was " incorrect procedural adherence due to personnel error."
The analysts do not 4
try to interpret licensee root cause determinations. The
{
details of the event contained in the licensee event report support the licensee determination.
4 Issue During an event at Quad Cities, the loss of a 125-VDC Bus was attributed to a contract technician opening a fusible disconnect on battery Bus 1.
A figure contained in the
~
licensee event report states that the 125-VDC control power system is shown for information only and is not part of the emergency diesel generator train.
Isn't it 1
odd that this incident is counted as a CCF?
Response: This is a dependent event that is considered -to be explicitly modeled in PRAs and is explicitly modeled in the common-cause database.
The coding of the event is 4
contained in the database and is not included in the emergency diesel generator reliability calculations.
Issue:
Since using the database is supposed to be what the licensees should do,
it would be useful to give a detailed example in the report. _ Listing all the required steps, detailed calculations, assumptions, and judgments l
would go a long way toward convincing the reader that such an approach'is practical.
Furthermore, repeating i
the analysis for the same system at another plant would be very illustrative, especially with. respect to the changes in the screening process.
Response: A detailed example for the auxiliary feedwater system is
' contained in Volume 5 of the INEL-94/0064.
The example illustrates all steps of the CCF framework contained in 7
Joint PRA/ Plant Ops. Subete. 10/30-11/1/96
- Minutes NUREG/CR-4780,
" Procedures for Treating Common-Cause Failure in Safety and Reliability Studies: Analytical Background and Techniques."
Mr. Rasmuson concluded that the common-cause database is the most complete database of its kind in the world-and that the data identification and collection process is systematic, rational, defensible, reproducible, scrutable, and understandable.
In
- addition, the use of the database will rosult in greater consistency and scrutability among existing PRAs.
The Members of the Subcommittee and the staff discussed how events were determined to be CCFs, the reliability of analysts to properly classify data, the quality of the database, the use of the database for plant specific studies, and the use of impact vectors.
Dr.
Apostolakis requested a copy of the staff evaluation of the South
{
Texas Project screening study of the database.
Several Members of the Subcommittee expressed the view that the CCF i
database was a pioneering and ground-breaking effort that provided i
a rational basis for regulatory decisions.
Dr. Seale cautioned that the information in the tables might be used for purposes other
[
than that intended by the staff, Accident Secuence Precursor Procram -
Dr. Patrick O'Reilly,-AEOD
. Dr. O'Reilly presented the objectives of the Accident Sequence Precursor (ASP) Program, examples of the results of the Program and an overvlaw of the Program plan.
He previded the following responses to previously identified ACRS issues:
Issue:
ASP utilization, feedback, accomplishments concerning I
improved safety.
Response: The ASP program identified the high risk significance of events at Wolf. Creek, Shearon Harris, and Vogtle, which might not have been recognized.
Special ASP analyses of shutdown events supported staff efforts to develop a Shutdown. Operations Rule.
Results of ASP analyses have been used in information notices and as inputs to the Senior Management Meeting (SMM) plant evaluation process.
Issue:-
AEOD uses simplified PRA models, for instance, in its ASP analyses for the plants.
Have these-models been J
reviewed?
Is there a continuing need to keep them?
Why not use state-of-the-art PRAs to evaluate operating experience?
With the wide availability of computerized models, this should not be a problem.
4 8
4
Joint PRA/ Plant Ops. Subete.
-L-10/30-11/1/96 Minutes
Response
AEOD is using Revision 1 simplified plant-specific models in ASP analyses. About one third of the mode).s have been reviewed and comments from licensees have been factored into the model development process.
Revision 2
simplified plant-specific models are undergoing a
systematic quality assurance process.
The staff has experienced dif ficulties in using detailed plant-specific i
models in the NRC SAPHIRE suite of PRA codes.
Currently, 13 detailed PRA models are available for use with SAPHIRE.
Issue:
Are the precursors, listed on pp. 10-13 of D. Rasmuson's ASP presentation to the subcommittee on July 17, 1996, in the plant logic models of the PRA/IPEs for the corresponding plants?
Does AEOD routinely make such comparisons?
Response: Original ASP program efforts to review ASP results for risk insights and comparison to PRA/IPEs was cancelled in April 1995 due to budget constraints.
The staff has j
requested funding for a new insights project.
Issue:
AEOD does not have models for the analysis of containment events or for the identification of systematic deficiencies of licensees' operating and safety culture.
What are AEOD's plans for future work on these topics?
Response: The Office of Nuclear Reactor Research (RES) is developing Level 2/3 PRA capability.
The RES activities associated with organizational factors are underway.
The subcommittee Members and staff discussed whether events are consistent with event trees in the model, the type of validation and verification performed on the SAPHIRE code, and how data uncertainties were treated in the ASP model.
Conclusion - Mr. Patrisk Baranowsky, AEOD Mr. Baranowsky and the Subcommittee Members discussed plans for future interactions.
Dr. Apostolakis requested review material be provide well in advance of future meetings.
Dr. Powers suggested that future meeting discussions include:
use of the CCF database for generic and plant-specific issues, e
availability of proprietary database, and e
9
Joint PRA/ Plant-Ops.-Subete. '0/30-11/1/96 Minutes development of documents to allow non-specialists to use the e
CCF database.
Dr.
Powers stated that the Committee should discuss how the 4
l-database could be used to verify that the present body of
-regulations are suf ficient and adequate, and consistently enforced
]
by the Regions.
Dr.
Fontana suggested a discussion related to organizational i
factors.
Dr.
Apostolakis ' requested that the AEOD staff summarize its presentation at the November 7-9, 1996 ACRS meeting.
He suggested that the staff present the philosophy behind its approach to l
performance measures and the point at which models become so unreliable that the results are analyses instead of indicators.
OCTOBER 31. 1996 Openina Remarks Dr.
George Apostolakis, Chairman of the Joint Subcommittee, convened the meeting and stated that this was the second day of the joint meeting of the ACRS Subcommittees on Probabilistic Risk Assessment and Plant Operations.
He stated that the purpose of today's meeting was to discuss the staff's approach to codify risk-informed, performance-based regulation through the development of Standard Review Plan (SRP) sections and associated Regulatory Guides.
Dr.
Apostolakis stated that the Subcommittee had received no written comments.
However, Mr. Jerry Eisenberg of the American
-Society.of Mechanical Engineers (ASME) had requested time to make a presentation regarding proposed ASME Code cases for risk-based inservice. testing (IST) and inservice inspection (ISI).
ASME Presentation
==
Introduction:==
Mr. Jerry Eisenberg, ASME Mr. Eisenberg, Director of Nuclear Codes and Standards, provided an overview of the American Society of Mechanical Engineers (ASME),-
its organizational structure, and programs.
He explained the consensus process by which ASME Codes and Standards are developed and summarized the organizational relationships between the different ASME risk-based projects.
The Subcommittee Members and Mr. Eisenberg discussed the consensus process and the use of expert opinion.
10
Joint PRA/ Plant Ops. Subete.
11 -
10/30-11/1/96 Minutes Risk-Informed Inservice Testino Initiatives Mr.
Gilbert L.
- Zigler, ASME Operations and Maintenance Committee, Innovative Technology Solutions Corporation Mr. Zigler presented the risk-informed ASME Code case initiatives for component
,cortance ranking and strategies for testing pumps, check valves, nd motor-operated valves.
He stated that the proposed Code case would provide a framework for categorizing components in an inservice test pramram using risk insights for the purpose of developing dif ferent testing strategies. He stated that the methodology is based on current technology and feasibility demonstrated in pilot plants.
Mr. Zigler explainad that the Code case blends probabilistic and deterministic considerations through the use of a plant expert panel.
The Subcommittee Members and ASME representatives discussed the following issues:
use of the consensus process and expert opinion, selection process for the plant expert panel, how ASME standards are updated, use of performance-based requirements, e
requirements for Level 2, 3,
and shutdown PRAs, and possible addition of risk-related requirements without any benefits (i.e.,
regulatory burden).
Risk-Informed Inservice Insoection Initiatives - Mr. Alex McNeill, ASME Section XI, Virginia Power Company Mr. McNeill provided an overview of the development and current requirements of ASME Section XI, Inservice Inspection.
He presented industry experience and Probabilistic Safety Assessment (PSA) insights concerning piping f ailures. He explained that three proposed Code cases involve methodologies based on qualitative insights and analyses derived from PSAs.
Mr. McNeill concluded that ASME Section XI.will provide standards implementing risk-informed processes for pressure boundary piping and that risk insights can enhance the current ASME Section XI requirements by concentrating non-destructive examinations on the high safety significant components.
The Subcommittee Members and ASME representatives discussed the completeness of safety evaluations, what unnecessary burdens will be removed by the proposed code cases, and how NRC staf f members on the ASME Code Committees can maintain their independence.
11 i
Joint PRA/ Plant Ops. Subete.
12 -
10/30-11/1/96 Minutes NRC Staff Prp,pentation Standard Review Plan Sections and Recrulatory Guides - Mr. Gary Holahan, NRR, and Dr. Thomas King, RES Dr.
King requested feedback from the Subcommittee Members concerning the draft Standard Review Plan (SRP) sections and associated Regulatory Guides developed to support risk-informed, performance-based regulation. He presented the status and schedule for developing of these documents and associated open items.
Dr.
King provided the following staf f recommendatione concerning policy issues:
The staf f should implement performance-based regulation in the context of the current PRA Implementation Plan through the current process.
The staff should development guidelines for plant-specific decisions that are derived from the Commission's current Safety Goals and subsidiary objectives.
The staff should allow small increases in risk under certain conditions.
l The staff should approve requested changes to risk-informed l
ISI and IST requirements as authorized alternatives under i
50.55a (a) (3) (i).
The staff presented the general approach used in the SRP sections and Regulatory Guides, and the fundamental safety principles in the risk-informed, plant-specific regulatory process.
He outlined the key technical and process issues, and the proposed approaches to resolve the issues.
The Subcommittee Members and the staff discussed the following issues:
resources required for licensees to prepare and the staff to review proposed change requests, generic effects of the regulatory requirements, e
basis for a deterministic engineering evaluation during preparation of a change request, and reason for licensees to submit a change request.
e 12 p
.J
- ~.
Joint PRA/ Plant Ops. Subete.
13 10/30-11/1/96 Minutes Dr. Apostolakis expressed the view that the staff should not wait for the ASME Standard Committees to approve proposed Code cases before issuing the draf t SRP sections and Regulatory Guides for public comment.
Dr. Apostolakis explained that risk evaluation consisted of two parts; a qualitative evaluation of event trees and a quantitative evaluation of the core damage frequency.
Dr. Catton noted that risk neutrality is extremely difficult to demonstrate.
PRA Acceptance Criteria - Mr. Gary Holahan, NRR, Dr. Thomas King, RES, and Dr. Mark Cunningham, RES Dr.
King presented the core damage frequency and containment performance guidelines, which vary depending on the type of PRA available and the calculated core damage frequency (CDF) or large, early release frequency (LERF).
He also explained the staff position on key issues such as the extent that PRAs should be used, what determines the quality of a risk analysis, and how uncertainty is addressed.
The staff compared the performance guidelines to the Nuclear Energy Institute / Electric Power Research Institute's PSA Application Guide (PSAAG) criteria for permanent changes.
He also compared the guidelines to the number of plants within ranges of IPE calculated values.
The Subcommittee extensively discussed the staf f guidelines for the use of CDF and LERF for making decisions regarding licensee submittals for risk-informed changes to the current licensing basis.
The Subcommittee debated the technical merits of the staff's proposed charts for CDF and LERF for guiding reviewers in considering the acceptability of proposed changes.
They compared the staf f's charts to the three-region approach provided in the industry Probabilistic Safety Assessment (PSA) Applications Guide (PSAAG).
The staf f stated that the use of decision charts requires judgment and asserted that the boundaries are somewhat " fuzzy" in terms of their use as guidance for decisionmaking.
Dr. Apostolakis expressed concern that these charts could become de facto requirements or acceptance criteria.
He suggested that these charts be deleted and that the narrative guidance be enhanced.
Other Members of the Subcommittee expressed support for keeping the charts in that they provide a good overall perspective of the decisionmaking process.
, Overview qL NUREG 1602 - Mr. Adel El-Bassioni, NRR Mr. El-Bassioni explained how the guidance contained in NUREG-1602,
" Standards for PRA to Support Risk-Informed Decisionmaking," was developed.
He presented the objectives, key considerations, and 13
Joint PRA/ Plant Ops. Subete. 10/30-11/1/96 Minutes structure of NUREG-1602.
Mr. El-Bassioni presented the character-istics of a full-scope PRA and the standards for modeling-full power, low-power, and shutdown operations.
He also covered the information presented in Appendix A,
"Use and Limitations of Importance Measures and Sensitivity Studies," and Appendix B, " Peer Review,"-of NUREG-1602.
The Subcommittee and the staff discussed the following issues:
what source term to use during shutdown operations, quantification of uncertainties, e
penalty for not having a full scope PRA, e
use of scenarios as importance measures, and e
difficulty in reading draft NUREG-1602.
e Dr. Apostolakis stated that allowing use of sensitivity analyses l
provided licensees too much flexibility and that he-would prefer l
that uncertainty analyses be required.
General Discussion Dr. Garrick expressed the opinion that the staff should take what its knows about reactor safety, safety analysis, and PRAs, to develop a new theory of regulations.
He stated that deterministic and probabilistic engineering analyses are not separate but are part of each other.
He suggested the staff review the ef fectiveness of each defense-in-depth barrier. Dr. Garrick stated a concern regarding the documentation required by the draft Regulatory Guide for change requests.
He recommended using calculated risk, such as offsite dose, instead of surrogate numbers for risk.
He also recommended that the safety goals not be used as absolute limits.
j Dr. Miller stated that the staff should establish standards and should consider industry standards and guidance, which already exist.- He expressed the opinion that the staff should wait until proposed ASME Code cases are approved.
He noted the apparent lack of. industry involvement in the SRP and Regulatory Guide develop-ment.
Dr. Miller suggested that the development of risk-informed regulation was driven too heavily by schedule concerns.
Dr. Seale concurred-with the use of expert panels.
Dr. Powers questioned the necessity and the sufficiency for each of the implied requirements, and some demonstration that they are necessary and sufficient.
He stated that the purpose of the requested documentation was not clear.
14
Joint PRA/ Plant Ops. Subete. 10/30-11/1/96 Minutes Dr. Apostolakis stated that both the staff and industry were treating PRA as a new methodology to be used in addition to the current regulatory framework.
He expressed the opinion that the schedule for issuing the Regulatory Guide and SRP had not allowed open-thinking about the regulatory ap?toach.
He suggested that a basic point of view was missing and t.Lat the staff should rethink its approach.
He also suggested that the staf f should define a set of principles for risk-informed, performance-based regulation.
NOVEMBER 1, 1997 OPENING REMARKS f
l Dr. Apostolakis, Chairman of the joint Subcommittees, convened the i
meeting and stated that the ptrpose of today's meeting was to continue the discussion of the NE staff's approach to codify risk-
- informed, performance-based regulation through development of Standard Review Plan sections and associated regulatory guides.
NRC Staff Presentation-i Deterministic Evaluation and Acceptance Guidelines Mr. Gary Holahan, NRR Mr. Holahan presented the basic considerations that should be addressed in deterministic engineering evaluations for changes to current licensing basis (CLB). He explained the considerations for preserving defense-in-depth, maintaining adequate safety margins, and meeting current regulations.
The Subcommittee Members and the staff discussed the following issues:
difference between meeting the 10 CFR Part 50, Appendix A, a
" General Design Criteria,"
and meeting safety margin requirements, overlap between probabilistic and deterministic analyses, e
use of the deterministic evaluation to impose requirements on e
licensees, and statement of the purpose, objective, and acceptance criteria for the required change request documentation.
15
~
Joint PRA/ Plant Ops. Subete. 10/30-11/1/96 Minutes Intecrated Decision Makina - Messrs. Gary Holahan, NRR, and Mark Rubin, NRR Mr. Holahan presented the key elements in the integrated decision making process.
Mr. Rubin explained how uncertainty should be treated in the decision making process.
Mr. Holahan explained how the integration process should be carried out by various NRC organizational groups.
The Subcommittee Members and the staff discussed the precision, completeness, and rigor of analyses; the imprecision of uncertainty distributions; use of risk-informed analyses in the general regulatory processi and weaknesses of expert panels.
Imolementation and Monitorino Strateales - Mr. Mark Rubin, NRR Mr. Rubin explained that implementation and monitoring strategies should be based on the findings of deterministic and probabilistic engineering evaluations, and on licensee performance.
He stated that the monitoring strategies should be periodically evaluated.
The subcommittee Members and the staff discussed the use of performance measures and monitoring to address uncertainty, Documentation and Licensee Submittals - Mr. Thomas King, RES Dr. King presented the dif ferent documents identified in the draf t Regulatory Guide DG-1061 (General Guidance) that licensees would need to support submittals for changes to the CLB, He explained the composition of
- archival, PRA peer review, and submittal documentation.
The Subcommittee Members and Staff extensively discussed the peer review process for proposed changes.
Kev Issues for Discussion -
Mr. Robert Jones, NRR Mr. Jones presented revised staf f positions to address Subconmittee concerns that were raised earlier in the meeting.
He addressed issues regarding the process elements, level of detail in licensee submittals, quality of PSAs, and PSA decision criteria.
Mr. Jones explained a reformatted diagram of the process elements that combined PSA and traditional analyses into a single diagram.
He presented the major principles associated with risk-informed decisionmaking.
The Subcommittee Members and the staff discussed waiting for the approval of ASME Code cases before issuing the draft Regulatory
- Guide, feedback loops in the process elements, the meaning of 4
" safety margins," the need for defense-in-depth in risk-informed regulation, and the rationale for accepting only small changes in risk.
16
__ _.~
I Joint PRA/ Plant Ops. Subcte. 10/30-11/1/96 Minutes
',:e-ents of subce=4 tree Ma=hers The following are summary comments provided by individual Subcommittee Members.
Dr. Powers stated that the new process diagram prepared by the staff provided cne picture that communicated several thousand words.
Dr. Miller stated that the new diagram was better than the old six block approach.
Dr. Garrick stated that one remaining major problem was how to interpret, present, and implement risk-informed performance-based decisionmaking conforms with the PRA policy guide.
He suggest the staff prepare a white paper that explains how the risk-informed approach will result in less work.
He noted his concerns related to the use of limit lines and uncertainties.
Dr. Garrick concluded that the staff is generally headed in the right direction.
Dr. Apostolakis stated that he liked the principles approach and that the staff was moving in the right direction.
Dr. Powers stated.that the guidance should not provide instruction without considering the approach provided in the principles.
Followuo Action I
The staff committed to provide the Subcommittee a revised version of the Standard Review Plans and the associated Regulatory Guide before the November Subcommittee meeting.
SUBCOMMITTEE RECOMMENDATIONS The Subcommittee decided to continue discussions on codifying risk-informed, performance-based regulation at the November 21-22, 1996 Subcommittee meeting.
BACKGROUND MATERIAL PROVIDED TO SUBCOMMITTEE MEMBERS 1.
Staf f Requirements Memorandum dated June 11, 1996, from J.
Hoyle, SECY, to John T. Larkins, ACRS,
Subject:
Requested ACRS i
actions regarding meeting with the Commission'on May 24, 1996 2.
Staff Requirements Memorandum dated May 15, 1996, from J.
Hoyle, SECY, to J.
Taylor, EDO,
Subject:
Requested actions regarding Commission briefing on PRA Implementation Plan on April 4, 1996 3.
Letter dated September 6,
1996, from S.
Jackson, Chairman, NRC, to T.
Kress, Chairman, ACRS,
Subject:
" Risk-Informed, Performance-Based Regulation and Related Matters" 17
,(
-Jdint PRA/ Plant Ops. Subete. 10/30-11/1/96 Minutes -
4.
Letter dated August 15, 1996, from T. Kress, Chairman, ACRS, to
'S,
- Jackson, Chairman,
- NRC,
Subject:
" Risk-Informed, Performance-Based Regulation and Related Matters" 5.
L Letter dated July 18, 1996, from S.- Jackson, Chairman, NRC, to
-- T. Kress, Chairman, ACRS, Subject : " Potential Use of IPE/IPEEE Results to Compare the Risk of the Current Population of Plants with the Safety Goals" 6.
Letter dated June 6, 1996, from T. Kress, Chairman, ACRS, to S.
- Jackson, Chairman,
- NRC,
Subject:
" Potential Use of IPE/IPEEE Results to Compare the Risk of the Current Population of Plants with the Safety Goals" l
7.
Letter dated June 3, 1996, f rom S. Jackson, Chairman, NRC, to -
l T.
- Kress, Chairman,
- ACRS,
Subject:
Probabilistic Risk
-Assessment Framework, Pilot Applications, and Next Steps to Expand the Use of PRA in the Regulatory Decision-Making Process 8.
Letter dated April 23, 19'1, from T. Kress Chairman, ACRS, to S.
- Jackson, Chairman, MC,
Subject:
"Probabilistic Risk Assessment-Framework, Pilot Applications,- and Next Steps to-Expand the Use of Pra in the Regulatory Decision-Making Process" 9.
Memorandum from dated September 9,1996, from M. Markley, ACRS
}
- Staff, to P.
Baranowsky,
- AEOD,
Subject:
"Ccmments -and Questions by Individual ACRS Members for Use by AEOD in Preparing for Next PRA Subcommittee Meeting October 30, 1996."
10.
Memorandum dated October 11, 1996, from W. Hodges, RES, to J.
Larkins,
Subject:
" Transmittal of Pre-Decisional Draft NUREG-1602, ' Standards for Probabilistic Risk Assessment - (PRA) to Support Risk-Informed Decisionmaking'"
11.
Memorandum. dated October 22,
- 1996, from M.
Markley, ACRS Staff, to G.- - Apostolakis, ACRS PRA Subcommittee Chairman,
Subject:
Meeting Summary:
" Commission Briefing on Status Update of Probabilistic Risk Assessment Implementation Plan,"
and attachments NOTE:
Additional details of this meeting can be obtained from a transcript of this meeting available-in the NRC Public Dccument. Room, 2120 L Street, N.W.
Washington, D.C.
20006, (202) 634-3274, or can be purchased from Neal R.
Gross & Co.,-Inc. Court Reporters and Transcribers, 1323-Rhode-Island Avenue, N.W. Washington, D.C.
20005, (202) 234-4433.
18
_ A