ML20207L425

Forwards Info Re Advanced LWR Performance Goals.Info Assembled After Discussion W/Knowledgeable NRC Staff Members & During 880119 Meeting W/Erpi Representatives

ML20207L425
Person / Time
Site:	05000601
Issue date:	01/25/1988
From:	Crutchfield D Office of Nuclear Reactor Regulation
To:	Murley T Office of Nuclear Reactor Regulation
References
PROJECT-669A, PROJECT-671A, PROJECT-675A, PROJECT-676A NUDOCS 8810170292
Download: ML20207L425 (10)

Text

_ _ _ _ _ _ _ -

January 25. 1988 Docket No. 50 601 Project Nos. 669, 671, 675, 676 MEMORANDUM FOR: Thomas E. Purley, Director Office of Nuclear Reactor Regulation k

. t ,1h(

THRU: Frank Miraglia Associate Director for Projects &

FROM: Dennis M. Crutchfield, Director Division of Rsactor Projects - !!!, !Y, Y and Special Projects

SUBJECT:

ALWR PERFORMANCE GOALS Per your request, attached is an information paper regarding ALWR performance goals. This information has been assembled after discussion with knewledgeable NRC staff nenbers and during a January 19, 1988 meeting with ERPI representatives. WehaverequestedEPRI,GC,y,andCEtodefinetheirdesign goals in this .rea for their project (s) and will provide an update to this information paper when their formal responses are received.

orisinal siso*4 W ternisW.C M

• W N Dennis M. Crutchfield, Director Division of Reactor Projects - !!!, !Y, V and Special Projects CONTACT:

T. Kenyon NRR/PDSNP Ext. 18206 DISTRIBUTION:

Central File ' Pleech ' \

NRC FCR GYissirg '

PbSNP Reeding DScaletti TKenyon RWHeuston l LRubenstein JFartin ADR#4, JCollins MTaylor DCrutchfield . FMiraglia PDSNPU- PD [(

LRubinstein k

D DC 4

eld A

F is TK#hn:cw 01/j\/08 0171/88 01A.1/E3 01/g//88 01 8 gnoS8s?SSi8hjjg1

l l*

l ALWR Performance Goals The following information has been assembled af ter discussion with .

knowledger.ble NRC staff memters.

Large Radioactive Release Probability

, 1. Background Prior to the issuance of WASH 1400, it was generally held by the staf f and industry that ige probability of a large release due to core melt was on the order of 10 / reactor year. This number was based on design besis accidents, assuming complete core meltdown and a containment leakage rate of 0.1 (Volume) %/ day. ,

-5 Howaver, WASH 1400 stated that this probability was actually 5 x 10 ent PRAs have established this -

(etan value)/ reactor year and number to be on the order of 1 x 10 subsegg/ reactor year (median value).

n 1985-1986 time frame, Congressman Markey asked the NRC to estimate the r vere core damage accident in the next 20 years.

1 grobability Assuming a 3ofx a 10ig/ reactor year probability of severe core damage, the i l staff indicated the probability of a core melt in the next 20 years was ,

approximately 45% using ultra conservative simple ratio arithmetic. A I realistic esti.uate of 12% was given later. The ACRS determined this i number to ibout 17%. this probability l could be es high as 20%, Commissioner and concluded Asselstine that a 1 x believgg/

10 reactor year core insufficient for new advanced reactors and was of the melt probability view that a 1 x 10wgg/ reactor year core melt probability with a 1 x 210 /

reactor year containment failure probability should be an advanced plant safety goal. Af ter the accident at Chernobyl, Chairman Palladino and ,

Comissioner Asselstine proposed a performance 9_oa_l of an overall mean adioactive materials from a reactor frequencyofalargereleaseof,g/reactoryear.

accident to be less than 1 x 10 (Note that this is a <

l performance goal, and not an objective). The Safety Goal Policy does not define a large release, and work is underway within the staf f to do so.

Current NRC thinking defines a large release as_the amount of radioactivity 4

that has a 50% probability of causing an early fatality.

~

~ " -

2. Derive. ion of Numerical Value The 1 x 10-6/ reactor year large release probability is not without some theoretical support. It has been well established by PRAs that the median.g/

1 x 10 reactor year.ilculated frequency For most plants, onlyofa severe fractioncore of the damage calcu- is on the order of lated severe core damage sequences are likely to progress to a large-scale core melt. Of the core melt sequences, only 1 in 10 (or less) it expected to yield a large release of radioactive material (large release not yet defined). The probability of containment failure is approximately 1 in 10. Therefore:

l

. 1 a

1 2-10'4/ reactor year (core melt probability) x 10*I/ reactor year (large '

l

- *corereleaseprobability)x10'I/reactoryear(containmentfailure l probability)=10-6/ reactor year (large release to environment probability). l In addition, the staff believes that conditions at most sites (population, locaticn, weather, evacuation plans) are such that, even with a large ,

release, there is only 1 chance in 10 that an early fatality will result i

from such a release. Therefore, the staff has concluded that the risk of I an early fata11ty from a core melt is quite low. ,

1 ,

Note, however, that Comission Asselstine questiongd the staff's use  !

of the median frequency of severe core damage (10' / reactor year) since j

onstrated that severe core damage frequencies could be as .

i PRAshavedp/reactoryear.

high as 10* ,

1 Maxim m Severe Accident Dose 1 i

. 1. Background

• i i

IPRI's ALWR Requirements Document currently proposes a requirement that j the dose beyond 1/2 mile radius fron, an individual reactor shall not

• j exceed 25 rem (whole body) in the event of a severe accident. The exact  !

t origin of this number 1. a little unclear. During early discussions over

} the definition of large release, there apparently was a move by the staff to relate the maximum severe accident dose to 10 CrR Part 100, which .

states, in part, that, assuming a fission product release from the core I with an expected containment leak rate and appropriate meteorological l 3' conditions, the exclusion area should be

... of such size that an individual located at any point on its l boundary for two hours immediately following onset of the postulated  !

fission product release would not receive a total radiatien dose to I i

the whole body in excess of 25 rem...

.. (hote that most exclusion areas are within a 1/2 mile radius from the reactor). Certain staff members are of the opinion that EPRI basically took the position that the industry would meet 10 CFR Part 100 for severe accidents. First indications that the industry would propose this r"%er j

came during a GE presentation 'o the ACR$ stibcomittee on Advanced . actors i 3  !

! in 1986. l Derivation of Numerical Value l 2.

The biological effects of radiation exposure have been well documented. It j is expected that 100% of individuals receiving a dose of 500-600 rem (whole body) would not be expected to survive the exposure. Persons receiving a lesser dose on the order of 100 rem (whole sody) would be expected to show l i l

biological changes (such as drop in blood cowat), and a small percentage j would be expected tt' suffer early fatality. 25 rem is the whole body dose generally considen J to correspond to no observable health effects after

{ exposure, and has aistorically been used as a cut off point for such matters.

I l

a - - - -- - - - - - . - - - -- -

1 I'. .

3 The 1/2 mile radius from the reactor roughly corresponds to th'e exclusion l 1 zene typically used in current nuclear p ants.

Advanced Plant Vendor Views

] '

i The staff has requested EPRI, GE, W, and CE to precisely define their design

! goals in this area for their pro,iect(s). Attacied is a copy of the request '

1 for additional information.  :

q '

J During a January 19, 1988 meting, the staff discussed this utter with l I W Attached is a ccpy of EPRI's representativec of presentation. The followingEPRI, GE, Ts,aand CE.

sumary of EPRI's verbal response.  ;

< EPRI has proposed a public safety criterion that the dose at a distance of 0.5 I 4

mileshallbelesstian25 rem (wholebody)forggeidentsequenceswhose .

j a cumulative (mean) frequency value exceeds 1 x 10 / reactor year.  ;

i The E5 ren (whole body) was selected because it represents a very low dose with l

'no observable health effects" (10 CFR Part 100) ,

,. The 0.5 mile site boundary was selected because it is a reasonable site  !

j boundary distance.

The 1 x 10'0/ reactor year accident frequency was selected because:

' t

1) it indicates a high level of protection to the public 'l in a f

million chance, i

2)itissufficientlylowforutilityinvestmentconsiderations,ag  ;

j 3) it is attaiaable, while a frequency criterien lower than 1 x 10  !

I would be difficult to analytically demonstrate.  ;

i The 1 x 10-6/ reactor year accident frequency is consistent with a design geel in the l

1 x 10,gWR requirements

/ reactor years. This nuder document that the is believed by the (man) industry coretodamage be a frequency (CDl, j factor of 5 to 10 times better than most currert plants and is believed to be }

j achievable. To analytically demonstrate a lower core damage frequency would (

! be difficult, since consideration of difficult te-define events (e.g., cominon i l pode failures, human interactions) for which good analytical data is scarce would have to be included in the evaluation.

EPRI reccgnizes the containment failure rate w ighted over credible core

' damage sequeness can be inferred to be 1 x 10'q/ reactor year. Therefore: )

]

1 x 10-5/ reactor year (CDF) x 1 x 10'I/ reactor year (containment failure l l l rate) = 1 x 10-6/ reactor year (accident frequency).

l Note that, by using this approach to define the ALWR perform nce goal, it is j jl not necessary to define what is mant by a large release nor is it necessary te relate this goal to the health objectives of the Safety Goal Policy since j EPRI believes compliance with the ALWR performance goal is more restrictive 1

I I 1 i

!_ l

4 than compliance with the Safety Goal Policy. It is necessary, however, to define the meteorology analysis methods to be used when determining compliance of specific designs with this goal. EPRI intends to utilize the CRAC2 i consequence calculation rethodology, so that it will average weathei conditiens for a j

specific site over a one year period, but take no credit for the wind rose.

Definitions and assumptions used by EPRI in definin this performance goal are d

provided in the attacied slides. We have not recei ed a ferral response to the request for additional information yet, and will provide an update regarding the response when it is received.

i 1

0 l

t i

i l

l i

4 I _ _-- _ _ _ _ _ _ _ - _ _ - _ _ - _ _ _

1. 'q,, UNITED STATE 8

+

f r NUCLEAR REGULATORY COMMISSION 3 i W ASHINGTON, D. C. 30Se4 Docket No. 50-601 Project No. 676 Mr. W. J. Johnsen -

Nuclear Safety Department Westinghouse Electric Corporation Water Reactor Division Box 355 Pittsburgh, Pennsylvania 15230

SUBJECT:

RE0 VEST FOR ADDITIONAL INFORMTION REGARDING DESIGN GOALS ADDRESSING SEVERE ACCIDENT RELEASES i

) As a result of our review of your ALWR application, we have determined the

, need to request additional information regarding your design goals that address large radioactive releases resulting from a severe accident. Attached are our questions. ,

l Please respond to this request for additional information within 30 days of 1 the date of this letter. If you have any questions concerning this matter. .

l contact the project manager for your application. i l

l 1 Sincerely, 1

j (ester S. Rubenstein. Director Standardization and Non-Power i l Reactor Project Directorate l l Division of Reactor Projects !!!, !Y. l i V and Special Projects j Office of Nuclear Peactor Regulation

Enclosure:

l

! As stated '

ec

See next page l l

r

l l

l REQUEST FOR ADDITIONAL IhFORMATION

. l in Chapter 1 of the EPRI ALWR Requirements Document, EPRI proposes the following requirement:

In the event of a severe accident, the dose beyond a half-mile radius The expected frequency of from the reacter shall not exceed 25 rem.

' occurrence for high off-site deses shall be less than once per million reactor years, considering both internal and external events.

1 CE has indicated they int >nd to comply with this criteria.

GE has defined their compliance with this position in its Licensing Review Bases.

W has not comitted to this position for its RESAR SP/90 application.

~

Rather, it has comftted Ittoisshowing a core melt however, our understanding, frequencythat of the W -

approximately 1x10" .

' AP-600 will cocply with the EPRI requirements.

1. a) Provide a precise statement of your design goals that address large radicactive releases resulting from a severe accident. This statement shoold be defined in terms of probabilities as well as large releases.

Items to define include the number of sequences to be considered, the use of internal and/or external events, consideration of sabotage (insider / I outsider threat), etc. This statenent should clearly indicate whether 1

the values used are median / upper bound / lower bound / average /etc. values, b) Provide your definition of core damage (clad tee;,erature/p(ercentage ofleakage claddir.g failure /compitte meltdown), containment failure .

total release), and lar,* release (threshold value) (as app /windropriate).

Specify your/assumptions direction wind speed regarding

/ wind meteoroloadverse shift probability or expected weather),

/gy (plume stre f l

1 population distribution (probability of individual seeing plume / location 1 of individual (s) during release), and time of exposure (as appropriate).

2. How do you propose to show the NRC that you meet this objective?

i I I

i i

! l l

l l

t

a

.e

~

EPRI Al.WR REQUIREMEN,TS DOCUMENT

. ALWR CORE DAMAGE FREQUENCY CRITERION In addition to meeting all other licensing design basis requirements, mean annual CDF s.1 x 10 5 Believed to be a factor of 5 to 10 better than most current plants Believed to be sufficiently low for protection of utility investment A lower value would present a problem in making an anaiytical demonstration Function level PRA models are being developed for ALWRs Work to date indicates . 7t the ALWR requirements specified to date will result in a plant that is likely to meet the 1 x 10.s target value l

l wrt!1/19'ntC 2

e EPRI At.WR REQUIREMENTS DOCUMENT ALWR PUBLIC SAFETY CRITERlON .

7 -

+ Additional check beyond meeting other licensing design basis requirements.

- For accident sequences whose cumulative frequency mean value exceeds 1 x 10 8 per reactor year, the dose at a distance of 0.5 mile l shall be less than 25 Rem whole body.

. This criterion is viewed as an extremely demanding target worth reaching for.

The criterion was selected based on a number of considerations

- A desire on the part of utility sponsors to define an ALWR that is excellent in all respects

- An accident frequency less than 1 x 104 per reactor year is low enough 10 '.'atisfy this desire for excellence and the public perception

- A frequency criterion lower than 1 x 10 5would present a problem in making an analytical demonstration l

- 25 Rem at the site boundary is a very low does with "no observable health effects" (10CFR100)

• Preliminary estimates based on existing plant PRAs indicate that an improved ALWR whose dominant accident sequence frequencies are reduced has good likelihood of meeting this stringent criterion

• For design that meets the 1 x 105 CDF target, the containment function needs simply to add one decade of additional frequency reduction  :

(

- Consideration may be given to raising the 25 Rem target value if the improved, optimized design is unable to meet the criterion as presently i l stated

'l t

wtv1/19 nic 3

' EPRI Al.WR REQUIREMENTS DOCUMENT ALWR PRA KEY ASSUMPTIONS & GROUNDRULES

- Being developed, reviewed and submitted as a supplement to the CDF and Safety criteria contained inChapter 1 of the ALWR Requirements e Scope: Internal and External Events (exceot Sabotage)

- Core Damage defined as

- RCS collapsed level uncovers active fuel, and .

Realistic analysis shows cladding temperature >2200oF

~

- CDF criterion evaluated based on point estimate comparison with mean value (expected value) and cualitative uncertainty analysis

- Safety Criterion evaluated by comparing mean CCDF for whole body dose to the 1 x 104,25 Rem t'oundaries l

- CCF evaluations using EPRt/NUREG methods and data recently developed and benchmarked

- Human Interactions evaluated using EPRI SHARP analysis framework

and

- TesVmaintenance events quantified using THERP procedure i j (NUREG/CR 1278)

- Procedural actions and Recovery actions quantified using EPRI l HCR correlation (NUS 4351 and EPRI RP28471 Interim Report) l

• Hardware f ailures to run for "indefinite": evaluated based on 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> l mission time t

i j

wrv1/19'nic 4