ML19270E961
| ML19270E961 | |
| Person / Time | |
|---|---|
| Issue date: | 12/22/1978 |
| From: | Advisory Committee on Reactor Safeguards |
| To: | Advisory Committee on Reactor Safeguards |
| References | |
| ACRS-1576, NUDOCS 7901100386 | |
| Download: ML19270E961 (15) | |
Text
@li?.[ G '.Tn[ri~
R.'D..
.b h 0
~Pb4 /a/ reg.
d;'Ib:aN s N,L b ACAS- /s 76 (INUTES OF THE RELIABILITY AND PPOBABILISTIC ASSESSMENT SUNDMMITTEE PEETING August 25, 1978, Washington, D. C.
A meeting of the ACRS Subcomittee on Reliability and Probabilistic Assess-ment was held in Washington, D. C. at 1717 B Street, N. W. on August 25, 1978 and continued on October 4, 1978. The purpose of the meeting was to review the plans and programs of the Office of Nuclear Regulatory Research in risk assessment research.
Notice of the meeting appeared in the Federal Register, Volume 43, No.155, August 10, 1978 and Volume 43, No.180, September 15, 1978. ihe schedule for discussim and a list of attendees at the meetirg are attached to the minutes. No written statements frcn members of the public were received and no requests frcm the public to unke oral statemnts were received.
The subcommittee did not approve, issue, or receive any written reports dwing the meetire. A copy of the vu-graphs shown during the meeting is attached to the office copy of the minutes.
AUGUST 25, 1978 MEETING WI'IH NRC STAFF (8:35 a.m. - 4:00 p.m.)
Mr. A. Buhl, NRC Staff, reviewed the goals of the Probabilistic Analysis Staff and three areas where the PAS functions to help focus the NRC decision making process. He discussed six areas in which the PAS is support-ing reactor licensing. Mr. Buhl considers this work to be top priority. He also showed a vu-graph outlining the nine specific areas in which the PAS He said that the more important areas would be discussed in j
does its work.
790110c %,
. a r
I detail during the course of the meeting. During the course of the discussion, Dr. Mark and Dr. Saunders pointed out that methodology i
develognent requires more than combining probabilities.
It involves establishing ways to correctly aM logically handle the interaction of random variables, many of which are highly dependent.
In response to a cocment in the 1977 ACRS report to Congress, Mr. Buhl presented to the subcommittee copies of vu-graphs identifyirg the basis for the research in each of nine subject areas. However, he did not elaborate on these.
RISK ME'IHODOIDGY AND DATA ANALYSIS (tr. pg. 32-66)
Mr. W. Vesely, PAS, discussed the goals and status of the risk rethodology and data analysis programs. Rese programs are essen -ially research and are aimed at developirs methodology for use in licensirg activities and in analy-sis and evaluation of data and associated risks. S e specific programs discussed by Mr. Vesely were:
1.
Flood Risk Analvsis (tr. pc. 33-36) he object of the flood program is to develop a nethodology to predict flood occurrence probabilities, system failure probabilities, ard risk consequences. h is was initiated in 1978 and is scheduled to be completed in October, 1982.
2.
Fire Risk Assessment (tr. pg. 36-40)
His program is designed to construct systems mochls to quantify r
fire risks.
It is scheduled for completion in October 1982 aM was started in March,1978
-3 3.
Cceputer Code Develoment Program. (tr. og. 40-60)
Five computer codes are being nodified and improved.
Rese codes, when ccccleted, will be uW for licensing applications as well as research applications. He SETS and PIIOD codes reduce very large fault trees.
With the FRANTIC code, an analysis of test and nuintenance effec s on component failure rates can be studied. We SAMPLE code is desi 'ed to determine error bounds on failure rate estimates, and the COtO N code will identify all combinations of components in a fault tree that will fail the systems.
Also included with the code develop-ment programs is develoirent of a fault tree hanabook, cxsponent failure data analysis (to be used for statistical reliability evaluations), and reliability data analysis.
S e product of the reliability data analysis work will be a handbook on failure of mechanical ecxt:ponents which will be an adjunct to the IEEE electrical component failure hand-book. Rese programs are scheduled for completion at various dates between October, 1979 ard October, 1983.
4.
Human Error a.d LER Data Analysis (tt ;g. 60-61) l 2e product et this work will be a
.a failure rate handbook.
It is designed to nodel un error rates e
and to evaluate human error data
... n le in LER's.
l R e handbook is scheduled for puo11 cation in May 1979 and the program should be completed in Oc;aber 1982.
5.
140 Program (tr. ca. 61-65).
D e purpose of this program is to establish criteria that l
can be used in tech specs for allowed outaae times and sur-i I
veillance test intervals. H e program was started in 1975 and is now beirs expanded into a larger program; scheduled completion is October 1983.
PROBABILISTIC ANALYSIS STAFP INTERNATICNAL ACTIVITIES (tr. 66-71)
Mr. Vesely reviewed the involvement that the PAS has with foreign governments on risk assessment methodology and applications. He also reviewed the CSNI task force and reports it has published. S e task force l'
was established to characterize contributirg factors in nuclear risk anal-ysis.
4-a w
e s
-4 l'
CRAC DEVEIDPMErff (tr. pg. 72-101)
Mr. Roger Blond, PAS, dpscribed the CRAC code work beirg per-formed by the PAS and current uses for the model. % e objective of this work is to perform a realistic probabilistic assessment of the consequences of a release of radioactive material into the environment for a rarge of meteorological sequences. Several in-sights have been provided by the CRAC development.
Tw important ones are: 1) h e RSS model was found to over-estimate concentrations by about a factor of 2, ard 2)
Rain model, wind, and population correlations have little impact on the results. We work is scheduled for completion in October, 1979.
REACIOR APPLICATIOt;S ME'IHODOIDGY APPLICATICN PROGRAM (4 plants) (tr. pg.103-167)
Mr. M. Taylor, PAS, Mr. R. Denning, Battelle Coltebus and Mr. S. Asselin, Sandia discussed the applications program extending the RSS methodology to four different LHR plant designs. he ob-jective of this program is to examine the effects of design differ-ences on overall risk among the plants studied in the RSS and an ice condenser plant, a B&W plant, a CE plant, arri a GE BWR 6/ Mark III plant. he program also includes an elenent called WASH-1400 Base-lining. his element refers to re-analysis of the ckxninant accident sequences using improved models for core meltdown behavior and radio-nuclide release. W e program is broken into two principal tasks:
systems analysis and accident process analysis. We systens analysis
/
task (being performed by Sandia) models the systens that are designed s-
1 5-to niitig3te an accident and determines the &xninant accident sequences.
We accident proc.ess analysis (beirg performed by Battelle Colunbus) i examines release magnitudes for a given accident sequence. Systems analysis and accident analysis for the ice condenser is about 99%
complete. Systems analysis for the B&W and CE plants is about 90%
complete and systems analysis for the GE plant has just started.
Accident analysis for the B&W plant is 25% cmplete and has just started for the CE plant. We WASH-1400 baseline study for the PWR is 95% ccrrplete.
The indications are that releases may be somewhat lower with a consequent lesser risk.
Results from the ice condenser study indicate that overall risk is compar-able to the RSS results, but the dcminant accident sequences are different.
Other significant results reveal that contairrnent fails predminantly by overpressure, sequences related to contaiment leakage are important, and systems unavailabilities are comparable to Surry.
CLASS 3-8 ACCIDENT AIDl,YSIS (tr. pg.167-180)
Mr. J. Murphy, PAS, reviewed the work perforned by the PAS to assess the risk associated with Class 3-8 accidents. h e purpose of this work is to assist the Office of Standards Develogent in its revision of the descrip-tions of class 3 through 8 accidents as they are used in NEPA reviews. De studies as conducted thus far on PWR's indicate that the risks frcan these accidents are well below those estimated by WASH-1400. D e studies have ex-tended the WASH-1400 results for the low release high probability risks.
- N *" N
.yany -
s
.,...s
t
. t W vork on PWR's' is scheduled to te cmpleted in September 1978 and
[
the BWR work in April 1980. W ere is also an effort underway to fit i
statistical distributions to data obtained frm activity releases frm LER's, and to extrapolate to the region in which there is no data. m is has proven difficult because of the way in which releases are reported in the LER's.
i LIQUID PAHMAY STUDY EFFECT ON RSS CONSEQUENCE CAlfULATIONS (tr. pg.181-188)
Mr. J. Murphy described this study which was initiated in response to the Department of Interior comments on WASH 1400. The study performed I
as part of WASH 1400 only estimated liquid pathways up to the point where concentrations in typical drinkirg water were calculated. Since these were small, it was calculated that liquid -born releases would be small when compared with air-born releases. S e current study will describe the transport of radionuclides through the ground and estimate resulting L
health effects to the human population. It will also consider mitigating i
actions which might be taken to reduce risks and will compare liquid pathway consequences with airborn pathway consequences. Se NRC Staff I
expects the results to also be useful in diaracterizing LHR sites for operatiig and proposed reactors. A report on the study is scheduled for the end of 1978.
l I
PUEL CYCLE APPLICATIONS BIGH IEVEL WASTE ISOIATION (tr. pg. 190-214)
Mr. M. Cullirgford, NRC Staff, and Mr. J. Capbell, Sandia Laboratories described the PAS program to study waste isolation in deep geologic media.
L
.m--
e The objective of this program is to learn information and develop methollogy about important processes and mechanisms which govern l.
transfer of radionuclides frcm a waste repository to humans. Se results are expected to furnish guidance for licensirg decisions.
The first phase of the program, which is nearly ccrrplete, is to de-fine a reference repository and to develop various models to study interactions and transport processes to the human food chain.
Subsequent phases will extend the programs to cover media other than bedded salt and to identify data deficiencies, uncertainties, and pro-perties important to long-term safety. n e schedule c _ls for program completion in December 1979.
A variety of scenarios will be.ssumed to statistically assess the un-certainties in the models. Rese scenarios will investigate the impact of a variety of release, transport, and pathway nechanisms on the health effects to humans.
It is hoped that a wide range of future geologic, population, and weather conditions can be covered by this technique.
SUMMARY
OF 'nIE AMERICAN PHYSICAL SOCIETY S'IUDY CN WASTE MANAGmENT (tr. pg. 216-229)
Mr. F. Donath, University of Illinois, discussed the APS study on high level wate disposal and how confidence in the transport nodeling can be developed. In his remarks,'he said that there are three aspects to the problem of dissolving and transporting radionuclides in ground-water. An independent solution to any of the three could solve the
' ),
1 problem of isolation of high level waste. Because of the lack of knowledge in two of the three aspects, primary enphasis hs been placed on generic modeling of gromd water transport. Dr.
Donath described the principles on which gmundwater modeling is based and the areas in modeling where work should be concentrated.
Generic mdeling focuses on the major effects on transport, and not on a specific environment as the NRC program does. He indicated pmblem areas that need wk are: characterizing the actual geologic medium, and evaluating disruptive features in the waste storage environ-ment.
OCIDBER 4,1978 MEETING WITH NRC STAFF (1:05p.m.-4:45p.m.)
Non-Radiological Transportation Accident Risk Study (Tr. pg. 7-31)
Mr. Bruce Varnado, Sandia laboratories, discussed the work being con-ducted for the PAS. The study is aimed at developing a method I
for estimating risks to re r.: tor plants from :n-radiological hazardous
!a materials being transported near plants.
T;.< procedure for the study will be to identify materials which may enate a hazard, determine the e
environment which might be created by accie r.ts involving these ha:ards, j
and develop models for risk assessment. The result of the study will be g
a model t. e NRC Staff can use to estimate risk to a variety of reactor b:
sites. These hazards are being studied in an ef{ ort to provide a more systematic cupmach to what has been performed on a case-by-case basis f
in the past.
IMERGENCY RESPONSE PIANING (tr. pg. 32-58) d Roger Blond, PAS, discussed this program which was conducted in response 1
to requests from the Office of State Programs and the Office of Nuclear
)$
Reactor Regulation. The objective of the pmgram was to develop a real-
@u istic description of accident sequences, including release concentrations p
and health e'ffects, and to study the effectiveness of various energency re-ii sponse actions. The product of the study will be a doctanent describing a g'.
I spectrun of accidents which the Office of State Pmgrams and others can use g
.g.
lt to evaluate emergency planning in a consistent manner. Dr. Okrent, who has E
read a draft of this document, expressed some skepticism that it would pmvide enough guidance for emergency planners. H e redels developed g
in the Reactor Safety study were used to rodel the accident descriptions.
A set of transportation accidents involving radioactive materials also was described as part of the study. Eksever, there were only a few ex-i 7 treme situations where there was a substantial inpact on the public.
'I i
i INSIGUS INID D! PROVING 'IEE EFFEC. IVENESS OF 'IHE ISE PROGRAM T
(tr.pg.59-89) i Mr. Richard Denning, Battelle Colunbus, described this pmgram which u g
coupleted six tonths ago. The results were published as NUREG CR 0153.
The objective of the study was to reco::raend ways in which risk analysis might be used to make plant inspections rore effective. Three reccr:raend-ations were made and they are being addres.ned bv the Office of Inspection and Enforcerent. Th three recomendations are:
11 I6E reallocate their inspection effort bas _l on an assessment of risk, 2) I65 make a more detailed analysis of IIR's to try to better identi_'y causal factors, 3) a better identification of test and maintenance deficiencies that might allow LER reportable events to occur be made.
FUEL CYCLE RISK ASSES 9ENT PROGRAM (tr. pg.91-102)
Mr. M. Cullingford, PAS, reviewed the fuel cycle risk assessment pro:n ms that were not reviewed at the August 25, 1978 meeting. The fuel re-processing risk program was designed to develop information that will d
assist in formulating licensing criteria. The program is divided in e
/
three phases but is not being pursued at the present because of national policy considerations. Phase A, dealing with defining a reference prue.% eg i
facility and a methodology, has been corpleted. 1he c :er areas in the fuel reprocessing program have not begun yet.
A progran on waste management of f
radioactive gases associated with fuel cycle facilities will begin in FY 79.
I i
l
10 -
Programs in ocean bed isolation, away from reactor storage, front end of the fuel cycle, decommissioning, and shallow land burial are slated to begin sometime in the future.
ACCEPTABLE RISK (tr. pg. 103-127)
Mr. Buhl discussed the NRC Staff's plans in the area of defining I
acceptable risk. We Staff has selected consultants to assess (within j
the next 8 or 9 months) the state of the art in definirg acceptable i
risk. Mr. Buhl said that the NRC Staff does not know, at this time, what
{
will be accomplished with the results of the study. Se Chairman remarked
?
that he thought there should be some focus for the work beirq planned.
j i
He also said that he thinks the research should be performed by sancone not 1
associated with nuclear power.
?
WrE: he transcript of the meeting is an uncorrected unofficial stenogra-graphic record of the meetings. Copies of the transcript can be ob-tained frcrn ACE Federal Reporters, Inc., 415 Second Street, N. E.
Washington, D. C.
(202-547-6222) o O
O
/
O 9
SCIEDULE FOR SUBCGI41TTEE EETING CN RELIABILITY /A'D PROBABILISTIC ASSES 94ENT WASHINCTION, D.C.
AUGUST 25, 1978 The Staco=dttee will rrect in P.com 1046 at 1717 H Street, N. W. to review programs and plans for risk assessnent Inscarch.
l The tentative schedule for the reeting is:
8:30 a.m. EXECUTIVE SESSION 8:45 a.m. EETING WITH NRC STAFF 4:00 p.m. ADJOURNENT e essem
i.
s 1
ATENDANG LIST SUB00NITEE CN RELIABILITY AND PROBABILISTIC ASSESSME?if AUGUST 25, 1978 ACRS W. Kerr, Olairman H. Isbin J. C. Mark R. L. Wright, DFE S. 22unders, consultant J. Arnold, consultant NRC SANDIA R. Di Solvo J. E. Carpbell J. Murphf S. V. Asselin B. Hall J. Johnson -
J. L. Speung R. L. Inan M. Taylor G. J. Kalb F. Goldberg S. R. Sturges R. M. Bland BATIELUI F. M. Mcaning R. S. Denning D. Rubenstein J. Curry AG P. Riehm R. Gilman W. Bloom L. Abramson G. Edisen m
em..
+ e m-wo-
g
'IOPICS IVR DISCUSSION I
1.
Review of PAB activities and accomplishments in IY 78.
2.
Actions taken an recomendations in Chapter 9 of ACPS 1977 Report on Research.
3.
Proposed risk assessnent program _and plans for FY 79.
4.
Work being perforned outside NRC in risk assesstent research.
5.
Efforts being made to incorporate risk assessnent into licensing.
- e G
g
ATIINDANCE LIST SUBCONITIIE ON RELIABILITY AND PROBABILISTIC ASSESSMENT OCI0BER 4,1978 ACRS W. Kerr, Chaiman D. Okmnt J. C. Mark J. Amold, consultant M. Bender M. Plesset J. C. McKinley R. L. Wright, DI'E J. Austin, ACRS Fellow NRC R. DiSalvo J. Hickey R. Blond J. Curry M. Cullingford G. Edison P. O'Peilly A. Buhl OIIIFAS B. Varnado, Sandia Laboratories J. Pen' ad, Science Applications, Inc.
H. Filacchione, Bechtel R. S. Ibnning, Battelle-Columbus t
R. A. Touchton, OPS l
5
/.
5 e..
^*
e SCIEDULE FOR SUB00'MTII.E LEETING Q1 RELIABILITY AND PROBABILISTIC ASSESSbTNT WASilING10.'i, D. C.
.. 3 OCIDBER 4,1978 he Subcomittee will ncet in Room 1167 at 1717 H Struct, N. W.
to continue its review of NRC programs and plans for risk assessnent research.
De tentative schedule for the meeting is:
1:00 p.m.
FEETING WITH NRC STAFF 5:00 p.m.
RUOURN Topics to be discussed include risk assessncnt progr ms for:
1.
Reactor Applications (1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />)
Naa-Radiological Transportation Accidents.
a.
b.
Fenrgency Planning.-
c.
Inspection and Enforcenent.
2 Fuel Cycle Applications (30 min.)
a
, a.
Fuel processing.
b.
New Initiatives e
3 Other items of interest (30 min.)
i 4
t 4'i i
e-em
- e. e o e
e o
..w emmyse.
...ew
-en
.