ML19331E439
| ML19331E439 | |
| Person / Time | |
|---|---|
| Issue date: | 07/09/1980 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-1743, NUDOCS 8009100183 | |
| Download: ML19331E439 (7) | |
Text
/ke<.s - / ry's
'e
- e.
j DATE ISSUED:
7/9/80 11solto MINUTES OF THE ACRS SUBCOMMITTEE ON RELIABILITY AND PROBABILISTIC ASSESSMENT WASHINGTON, D.C.
APRIL 30,1980 The Subcommittee on Reliability and Probabilistic Assessment met with representatives of the NRC Staff and its consultants on April 30, 1980, to discus the PAS budget, research on the flood program, Interim Reli-ability Evaluation Program, and the development of quantitative safety criteria.
A notice of the meeting appeared in the Federal Register on April 15, 1980. There were no written or oral statements from members of the public.
The entire meeting was open to members of the public.
Principal participants at the meeting were the following:
ACRS NRC Staff and Others D. Okrent, Chairman W. Vesely, NRC J. Mark, Member F. Rowsome, NRC
~
M. Ber. der, Member R. Budnitz, NRC J. Ebersole, Member R. Bernero, NRC W. Mathis, Member W. Bivins, NRC S. Saunders, Consultant M. Temme, IEEE R. Wilson, Consultant A. Garcia, SAI W. Lowrance, Consultant W. Kastenberg, Fellou J. M. Griesmeyer, Fellow G. Quittschreiber, Staff *
- Designated Federal Employee MEETING WITH THE NRC STAFF (OPEN SESSION) 1.0 Subcommittee Chairman's Opening Remarks Dr. Okrent, Subcommittee Chairman, introduced the members of the subcommittee and noted the purpose of the meeting. He pointed out that the meeting was
.8'009100/g3
=
April 30,1980
.. Reliaiblity & Prob. Assmt Mtg.
f the Federal Advisory being conducted in accordance with the provisions oA t and that Mr. Gary
~
Committee Act and the Government in the Sunshine c eting. He Quittschreiber was the Designated Federal Employee for the it stated no requests for oral statements nor wr t en the meeting.
of the pubTic had been received' with' regard to ac:
Probabilistic Analyses Staff Research Budget. on the NRC research budget R. Budnitz, NRC/RES provided some general commentsexists and as it ha 2.0 l
noting that the research program as it present yIt is presently being rd risk analysis in the future.
in the past is LOCA oriented.
oriented and will be significantly reoriented towad certain that the FY 81 b Budnitz informed the Subcommittee that it appeareSince the research bud for the NRC will be cut significantly.
the non research budget is mostly for about half of -the total NRC budget and, since from the research area. The salaries, most of the NRC budget cut will come in the FY 82 research following will be given the highest funding priority d):
budget'(FY 81 cannot be significantly redirecte Risk Assessment Phenomena of severe accidents and mitigation
'l.
2.
Plant operational safety 3.
1 Siting 4.
Seismic and structural engineering 5.
the ACRS comments on risk R. Bernero NRC/RES provided a brief discussion onthe FY 81 budget a ittee assessment in its latest report to Congress onNo significa llowing were mentioned:
activities of PAS.
were noted; however, suggestions for the fo thod to-
' PAS should investigate the possible development of a me uncover design errors.
h d of
' PAS should investigate
..e possible development of a met o i
quality assurance for reliability.analys s.
~
- Reliaiblity & Prob. Assmt Mtg April 30,1980
' PAS should provide ACRS with FY 82 PAS budget and details of action taken on ACRS recommendations in written form by the end of May.
R. Bernero ' told the Subcommittee that reliability analysis or systematic anal,ysis techniques should be applied to safety review and inspection analysis techniques to evaluate effectiveness. This should be done to determine whether the safety yield is proportional to the resources going into these programs.
3.0 Flood Risk Study Members of the NRC Staff and JBF Associates discussed the presently funded research programs to predict flood occurrence (probabilities versus flood level). A code, FL0E, has been developed and is operational to predict flows.
PAS has performed a preliminary analysis of flooding on the Susquehanna River and has calculated 3 x 10-4 for-the occurrence of the prot 61e maximum flood (PMF).
It was noted that there are two orders of magnitude uncertainty
^ t the 5 and 95% confidence levels.
Also, it was noted that occurrence a
of the PMF does not necessarily mean core damage but may only require certain actions, such as installing protection barriers.
Dr. Okrent expressed his dissatifaction that the NRC had not brought this matter to the attention of the ACRS and the public at a much earlier date.
I Mr. Wagner, NRC/ PAS discussed the formal basis for assessing impact of floods on nuclear plants. The importance of the flood event was stated to result from common cause failure of multiple system components submerged in j
water due to a single event. Okrent suggested that the computer code was interesting but that possibly some knowledgeable individual who knews the plant might arrive at the same answers simply by looking at elevations of critical systems.
l 4.0 -Interim Reliability Evaluation Program Status and Findings The Subcommittee was given a briefing on the IREP evaluation at Crystal River by Mr. A. Garcia, Science Application, Inc..
Mr. Garcia indicated that the V
~ April 30,1980 Reliaiblity & Prob. Assmt Mtg t-evaluation was much more difficult than had been initially anticipated due to the many interdependencies.
PAS does not presently plan to use Crystal River as a model for the IREP for other platns.
It is expected that industry will be asked to perform a large portion of the IREP analysis for the remainder of the plants.
Rowsome said they were trying to find procedures and approaches that can be readily implemented by utilities instead of the NRC.
Rowsome noted that industry is starting to use risk assessment techniques to justify changes on backfitting. The NRC is welcoming the proposed risk assessment but will not give blanket exemption from ratchets. Rowsome said that a number of companies are volunteering personnel for work in IREP.
5.0 NRC Staff's Program to Develop Quantitative Safety Goals W. Vesely, NRC/ PAS, discussed the schedule for the NRC/RES sponsored program for numerical risk criteria. Vesely said that by late August 1980 they will have some criteria formulated for review. The final report for Senate', Commission and ACRS review will be ready by January 1981.
Vesely suggested a one year review of the final report before acceptance.
He added that in addition to criteria they will look at the following:
'modeling techniques
'modeling systems
' kinds of equations to use
'what accident sequences to cover
'what quantitative approaches to use
'how to propagate uncertainties
'what data bases to use W. Vesely suggested that an unacceptability level approach could be used in conjunction with IREP to define what is intolerable to the NRC.
This would not replace existing criteria; but, as experience is gained it may replace some of the deterministic criteria such as the single failure criteria. -Yesely mentioned three criteria that are being considered.
1.
Pr'obability versus Core damage (10-3/ reactor year unacceptability level) t
Reliaiblity & Prob. Assmt Mtg April 30,1980 2.
Probability versus release category 3.
Probability versus health effects (10-5 individual fatality).
4 6.0 IEEE P'rogram to Develop Quantitative Safety Goals Mark Temme, representing IEEE, discussed the IEEE Working Group 5.4 organization, objectives, and schedule for development of risk criteria for nuclear power plants.' Temme indicated he would like to have the quantitative safety criteria serve as the bases for nuclear safety such that rules, regulations, and regulatory guides would be related to these criteria.
Temme hoped to have a draft standard on safety goals in about one year.
Part of the IEEE task will be to demonstrate how the criteria are met.
7.0 Risk Management Framework for LWRs J. M. Greismeyer, ACRS Fellow, discussed a proposed framework on risk management decision rules for LWRs. He proposed the following as a set of possible acceptable quantitative decision rules:
Hazard State Limits 1.
Significant core damage < 1 in 200 reactor years 2.
Large scale core melt < 1 in 1000 reactor years 3.
Large scale uncontrolled release < 1 in 5000 reactor years Individual Risk Limits 1 10-5/ year risk of early death 16 x 10-5/ year risk of delayed death Societal Risk Limit 110 equivalent deaths / year due to reactor accidents Societal Risk Reduction AL' ARA g
..o l
Reliaiblity & Prob. Assmt Mtg April 30, 1980 s'.
Risk Fee Pay risk fee to the NRC at a rate equal to the societal risk times the marginal cost limit.
Special sci'ence court would hear arguments and make final legal determination of estimated levels of risk.
Subcommittee member comments were favorable toward the approach that Griesmeyer suggested.
W. Kastenberg discussed the application of some of the proposed criteria with regard to specific plants. Application of the criteria looked feasible for the plants discussed.
8.0 Comments on Task Force Report, NUREG-0667 J. Ebersole discussed an item in Task Force Report, NUREG-0667.
The draft NRC Staff proposal recommends that license applicants be given the option of selecting either primary system cooling (feed & bleed) or secondary system' cooling (emergency feedwater) as the acceptable qualified method of cooling the core following a seismically induced loss of main feedwater.
It was the general feeling of J. Ebersole that the emergency feedwater system should be seismically qualified.
D. Okrent suggested PAS quantify its decision by stating the risk it was accepting.
9.0 Consultants Comments on Proposed Criteria and Concluding Remarks W. Lowrance summarized the general feelings of the Subcanmittee with regard to risk criteria. He concluded:
1.
quantitative risk ceilings for high technology are needed and are being used.
2.
arriving at the numbers to use for the criteria will be very ll difficult since they cannot be derived empirically.
Lowrance suggested that the hazard state limits be tied together and based on biological health effects. He suggested that the system should be checked to determine the sensitivity of the numbers and the implications
~
for design.
l L-
l} = ~.3 e
i Reliaiblity & Prob. Assmt Mtg April 30, 1980 R. Wilson agreed with Griesemeyers approach for separate and societal risk.
He felt that -it was not the individual that cared about additional risk aversion
'for large risks but society in general or an individual thinking about society.
Okrent and Wilson discussed the problem of having uncertaintiet for large risks.
It was suggested that uncertainties were not so important for indiviaual risk, Since for the case of cancer it is not relevant to an individual whether the uncertainty is due to'a basic randomness of the cancer process or due to lack of knowledge of-the parameters.
Uncertainty for societal ' risk can involve a very large number of people; therefore, more confidence may be needed fe large accidents and this should be factored into the risk aversion criteria.
Wilson suggested that any proposed criteria for reactors be tested on other industries, such as chemical industry, so they can be balanced.
A meeting has been scheduled for-July 1,1980 to continue the discussion of quantitative safety goals.
Industry will be asked to participate in that meeting with discussion of their specific proposals for safety criteria.
Note:
For additional details, a complete transcript of the meeting is available in the NRC Public Document Room,1717 H St., N.W.,
Washington, D.C. 20555 or from International Verbatim Reporters, Inc., 499 South Capitol Street, S.W., Suite 107, Washington, D.C.
20002,(202)484-3550.
Viewgraphs shown at the meeting are provided inthe meeting transcript and in the ACRS office file for this meeting.
O e
O N