ML19345B564
| ML19345B564 | |
| Person / Time | |
|---|---|
| Issue date: | 09/17/1980 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-1763, NUDOCS 8012020065 | |
| Download: ML19345B564 (7) | |
Text
A84J - /M 3
,. 19.'
DATE ISSUED:
9/17/80 MINUTES OF THE gg ACRS SUBCOMMITTEE MEETING ON RELIABILITY AND PROBABILISTIC ASSESSMENT JULY 1,1980 INGLEWOOD, CA The ACRS Subcommittee on Reliability and Probabilistic Assessment met in Inglewood, California on July 1,1980 to continue its evaluation of develop-ment of quantitative safety goals for nuclear plants, and to review NRC probabilistic programs and budget.
ACRS members present were D. Okrent, W. Kerr, P. Shewmon, C. Siess and J. Mark.
ACRS consulta'nts attending were L. Lave, P. Slovic, R. Wilson, W. Lowrance, S. Saunders and I. Catton.
Presentations were made by Zebroski and Whipple -
EPRI, O'Donnell - AIF, Temme - IEEE, Bernero - NRC/ PAS, Joksimovic - GA, Griesmeyer, Johnson, and Kastenberg - ACRS Fellows.
A list of documents reviewed by the Subcommittee at the meeting is attached.
EPRI Formulation of Risk Criteria Mr. Zebroski, EPRI, indicated his support for safety goals. Zero risk is unattainable in any endeavor.
Misallocation of resources results from not considering risk / benefit and relative risk.
The safety goal should be understandably stated.
It should consider alternative energy sources, as well as risk of deprivation, war, and cost of electricity.
It must be realized that complete public consensus can on the safet'y criteria never be achieved.
Rather, acceptance by rational informed people should be sought.
The safety criteria proposed was 30 years to the next severe core damage accident, which translates to a probability of 2 x 10-4 per reactor-year. Given a core damage accident, there should be a 99.9". probability that the release magnitude is:
8012020066
.,r; Eb N RPA Meeting. July 1, 1980
~(a) < 1 rem at the. site boundary; and (b) < 5000 rem total population exposure.
An additional criteria is -that the total risk ~ from nuclear power be no greater than 10% that of alternative power generation sources.
It was -note'd that if operating experience is properly fed back into' design and operation, then building more plants does not necessarily increase total risk from nuclear power.
Criteria Used for Considering Risk Mr. Whipple, EPRI, described criteria that are used when considering risk.
There are four types-of risk criteria in the literature:
(1) site boundary limit; (2) total population limit; (3) cost effectiveness (cost / benefit) test for measures proposed to reduce residual risk; and (4) lower bound limit on risk, where risks-that fall below this level will not-be reduced further.
AIF Formulation of Risk Criteria Mr. O'Donnell, AIF, summarized a June 2,1980 policy statement on quantitative safety goals. He indicated that:
(a) safety goals chould be applied to all technologies; (b) no individual should bear a disproportionate share of the risk; (c) societal resources should be. optimized in reducing overall risk; and (d) acceptable risk should include consideration of benefits of technology.
The proposed criteria are:
' Maximum individual risk of 10-5 per year for a person at the site boundary, utilizing a best estimate _ calculation of plant risk.
' Societal risk of 10-I per year per 1000 MWe capacity, also on a best estimte basis for plant risk.
'use of cost / benefit for deciding whether to reduce risk further below these levels, with $1 million per life saved as a criterion.
'A limit (currently unspecified) on the incidence of core damage events, to encourage prevention of accidents.
_ d,'
RPA Paeting July 1, 1980 These particular values for risk were chosen partly on the basis that they are small compared to other' risks-(e.g. coal burning, accidents)
IEEE Formulation of Risk Criteria Mr. Temma, IEEE, described progress on the IEEE working group on risk criteria.
Since the group is only a few months old, the work is still in tne early stages.
A draf t standard is planned for mid-1981.
The criteria, as currently planned, will be in terms of dose rather than deaths.
There will also be a criteria for land contamination.
Determining whether a plant meets the criteria is an important matter. Risk from sabotage will be excluded.
It was questioned whether it was proper to exclude sabotage.
NRC/ PAS Efforts on Risk Mr. Bernero, NRC/ PAS, described N'RC programs in risk assessment.
Brookhaven is performing four tasks:
(1) literature survey on use of risk assessment; (2) developing quantitative risk criteria; (3) developing methodology for deter- '
mining if the criteria are met; and (4) rebaselining WASH-1400. The result of these efforts will be four reports, due in October 1980.
There is alsa a program with Mr. Slovic to explore the philosophy of using risk assessment and establishment of safety goals. A draft report was issued in February 1980, with a final report in October 1980.
Some of the current NRC thinking on_ use of risk assessment and safety goals was discussed. PAS is considering setting.a limit on probability of core damage as an interim criteria for decision-making purposes. Dr. Okrent cautioned that not only probability, but also consequence of core damage must be considered.
A summary of ongoing risk assessments of various plants was given. There are a number being sponsored by NRC and by EPRI. Dr. Okrent requested that Mr. Bernero transmit to ACRS any ' reports relating to the Crystal River Risk Assessement program.
Dr. Okrent also requested NRC keep the Committee inforned on progress of the Crytal River risk assessment and its general implications.
g& July 1,--1980 RPA Mecting
~
' GA Formulation of-Risk Criteria Mr. Joksimovic, GA, described.the GA risk philosophy, the principles of which are similar to the other speakers. Use of Design Basis Accidents alone for plant design is inadequate, rather, risk assessment should also be used.
The proposed safety goal is in terms of risk to an individual at the site For accid'ents that occur with a frequency of > 10-4 per reactor-boundary.
year, the maximum individual exposure should be < 5 rem.
ACRS Fellows Formulation of Risk Criteria Mr. Griesmeyer, ACRS Fellow, discussed risk criteria he is developing.
The following preliminary criteria were proposed (because there are large un-certainties in risk assessments, the criteria are presented in terms of 90%
confidence level as well as best estimate):
' Probability of core-damage: 3 x 10-3 per r-y (reactor-year) best estimate; 5 x 10-3 per r-y at 90% confidence level.
' Probability of core melt: 10~4 per r-y best estimate; 5 x 10-4 per r-y at 90% confidence level.
' Probability of large uncontrolled release given the occurrence of a core melt:
10-2 best estimate; 5 x 10-2 at 90% confidence level.
I
' Risk to maximum exposed individual from ace.idents: 2 x 10-3 per r-y best estimate; 10-2 per r-y at 90% confidence level. Societal risk from core melt: 1000 per core' melt.
' Risk to maximum exposed ' individual from releases from normal plant operation: 2 x 10-5 per r-y best estimate; 10-4 per r-y at 90%
confi-dence level.
Societal risk from normal operation:
10 per r-y.
Delayed (cancer) deaths are treated in terms of the fraction of life expectancy lost due to the accident. Total deaths are then the sum of immediate deaths plus adjusted delayed deaths.
1
,.._f,.
RPA' Meeting July 1, 1980
-s It was said that a criteria for property damage is needed, but has not been formulated yet.
Finally, there is a cost / benefit criteria for reducing re-sidual risk. This criteria-is 55 million per early death averted and $1 million per delaycd averted.
It was said that risk assessments have large uncertainties and that the value for level of risk depends on who does the calculation. That is, there are likely to be disagreements over risk assessments. A risk panel consist-ing of scientists was proposed as a way to resolve differences of opinion.
There was considerable discussion on the merits of a risk panel, regarding the fona it should take and whether it would work.
Messers Johnson and Kastenberg, ACRS Fellows, presented results of some calculations based on Mr. Griesmeyer's criteria. The calculations dealt with effect of risk aversion and uncertainties on meeting cost benefit and risk criteria. This was done for the WASH-1400 plant and for a coal plant.
The examples indicate that if a risk aversion factor is used, the amount of money than may be justifiably spent to reduce risk can become enormous.
The same tends to be true when considering calculational uncertainty.
That is, the larger the uncertainty in the risk calculation, the more money that may be justifiably spent to reduce risk, if one chooses to penalize uncertainty.
There was considerable discussion on risk aversion, as to how risk averse society is and whether design of installations reflects risk aversion.
It was said by Dr.
Slovic that people are simultaneous averse and not averse, depending on the parti-cular circumstances.
Comments by ACRS Consultants
~
fu ACRS consultants summarized some of their views on risk criteria and safety goals.
fir. Slovic said that use of a risk aversion factor should be avoided.
He also noted that the effect an accidmt could have in restricting use of ruclear power as an energy source would have massive secondary social, economic, and political. effects that must be considered, along with the direct effects.
7
+.
., ~, w.- 5 RPA deeting; July 1, 1980
.Mr. Lave l indicated that one concern with ' risk criteria is verifiability of very
~
low probabilities, like 10-8 for example. He also said that' risk assessment involves value conflicts and it is the task of NRC to resolve these conflicts, aided by 'public hearings.- Other. federal agencies have not made much progress in use of risk criteria, it was noted.
Mr. Wilson also. indicated a concern over verifiability.
He suggested using, for example, a 2 s-l'evel value rather than a mean value of risk to increase confidence in having bounded the problem.
There was some skepticism or the consultants part on the risk panel idea.
They felt that, although it may have merit,.there are too many political
~
and social obstacles.
Anl alternative way was suggested to attempt to get approv:1 on risk criteria. That is, for the NRC to hold structured public hearings ~ on the matter.
NOTE: For additional details, a complete transcript of the meeting is available in the NRC.Public Document Room,1717 H St., NW, Washington, DC 20555 or from
~
Alderson Reporting Company, Inc., 300 7th Street, S.W., Reporters Building, Washington, D.C. 20024 (202) 554-2345.
e S
9
1.
RPAHee'ing;of.-7/1/80 t
j List of Documents
- 1. -FL0E Code Summary ReportLfor Susquehanna River
~ 2.. EPRI-Formulatioin of-Risk Criteria - 6 Slides
~
~
- 3.
. AIF Formulation of Risk Criteria - 16 Slides
~
i 4.
IEEE Formulation of Risk Criteria - 9 Slides -
- 5.
NRC/ PAS Efforts on Risk - 9 Slides l
- 6. ~ GA Fonnulation of: Risk Criteria - 10 Slides
~
ACRS Fellows Formulation of Risk Criteria - 42 Slides 7.
4 N
a w-n w
r-r
-r-
~,
e<m,,
r r