NRC-89-0213, Informs That Util Developing Program of PRA for Application to Facility,In Response to Generic Ltr 88-20, Individual Plant Exam Program for Severe Accident Vulnerability. Current Status of PRA Program Provided

From kanterella
Jump to navigation Jump to search
Informs That Util Developing Program of PRA for Application to Facility,In Response to Generic Ltr 88-20, Individual Plant Exam Program for Severe Accident Vulnerability. Current Status of PRA Program Provided
ML19327B644
Person / Time
Site: Fermi DTE Energy icon.png
Issue date: 10/25/1989
From: Sylvia B
DETROIT EDISON CO.
To:
NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
CON-NRC-89-0213, CON-NRC-89-213, RTR-NUREG-1335 GL-88-20, NUDOCS 8911030078
Download: ML19327B644 (6)


Text

n

. i ]+ : c - ,. .

p.>$6.

p # .s s

{[ e',. "M h[ * . ,

g, g gg lb E l M*' \hC* P't05"'

(.? '

~

6400 N6fth Dme i sh ay Newport %chipar. 40te3 p i

' (313) 586 4t50 n

w October 25, 1989 NIC-89-0213

,, y-( <

, lU. S. Nuclear Regulatory Cmmission A -

'Attna Document Cantrol Desk Washington, D. C 20555 Referencess; l)- Fermi 2

,, NBC Docket No. 50-341 NHC License No. NPF-43

2) Initiation of the Individual Plant Examination for Severe Accident Vuherabilities Federal Register, y Vol. 54, No.169, dated Friday, Septenber 1,1989
3) NUREG-1335, " Individual Plant Examination Submittal Guidance," Published by ti.N U. S. Nuclear r Regulatory Camission August,1989

,. 4) U. S. Nuclear Regulatory Commission Generic Letter l 88-20, "IPE for Severe Accident Vulnerabilities - i 10CFR50.54(f)", dated November 23, 1988 l L;

" l Subjectt Fermi 2 Individual Plant Examination Program for Severe l Accident Winerability - Generic Letter 88-20 l

Detroit Edison (DECO) is developing a program of Probabilistic Risk Assessnent (PRA) - for. application to Fermi 2. W e program when  ;

coupleted will meet the requirements' and objectives specified for the l Individual Plant Examination (IPE) in Generic letter 88-20 and is l

,, consistent with the supplemental guidance of NUREG-1335. % rough the i performance of the IPE, DECO will fulfill the stated objectives,  !

namely, (1) to develop an appreciation of severe accident behavior, (2)' to understand the most likely severe accident sequences that could i

. occur at Fermi' 2, (3) to gain a more quantitative understanding of the j overall probabilities of core damage and fission product releases, and i g ,

(4) if necessary, to reduce the overall probabilities of core damage ,

and fission product release by modifying, where appropriate, hardware

' '{N s

{

and procedures that would help prevent or mitigate severe accidents.

c.

? , his letter sunearizes the current status of the PRA program at DECO and

4

.1. Identifies the method and approach selected for performing the IPE,

2. Describes the method to be used, arri gool

~

I t

~L i 4

8911030078 891025 PDR ADDCK 05000341 l P PNU  !

w <

h "c.

    • 1 USNT '

, , October 25, 1989 2 1 NBC-89-0213 I' Page'2' L

L 1 I

i C 3. ' Identifies the milestones and schedules for performing the f' , IPE and submittal of the results to the NBC.

DECO has just conpleted a Ievel 1 PRA without external events or internal floods utilizing as its principal consultant Pickard, Lowe .

L and Garrick (PLG). Thus, the model is typical 'of recent PLG PRA modeling,' utilizing a suprntt stat 4 approach characterized by large  ;

' event trees and small fault trees, u.e latter actually depicted by "

. block diagrams and system equations. . We systen equations are supported by system notebooks for each system incorporated in the model. In order to assure adequate treatment of dependencies,

~

conprehensive dependency matrices were developed and utilized in the i modeling. Since Fermi 2 first began comnercial operation about the time the PPA work began, most of the conponent failure data and

, . initiating event- frequencies' were selected, from the PLG BWR generic ,

data base in conjunction with a. relevancy review. j

. Se Level'1 PRA sequences have been characterized in terms of their j plant damage state and functional failure type. Wis grouping provides for easy transfer of sequence information to the containment 3 ane. lysis and helps provide a basis for development of a subsequent accident management progran. As noted below, these groupings will be modified prfor to conpletion.of the IPE.

m The effort to date has been divided about equally between consultants (PI4 and a third party reviewer) and DECO. While significant consultant help will still be required, future plans call for smaller I consultant participation since it is DECO's intention to produce a ,

risk assessment tool that is esmntially free of the need of j consultant support. j The aforementioned Level 1 PEA will form the basic tool for the front .!

end of the IPE. However, there are significant modifications and  !

additions to be made' to the Level 1 model. and its associated report before it can be considered as the docunented basis for the IPE. We principal enhancenents are as follows: (

o 7ddition of selected containnant systens to couplete all of '

the needed plant damage states i o Review and incorporation of the results, as applicable, of a recently conpleted multiple control system failure study and of a detailed a.c. load list compilation o Incorporation of relevant plant design changes made during the first refueling outage currently underway.

o Transfer of the model to the new PLG software now under development to facilitate model revisions and, more inportantly, to facilitate the utilization of the model

r .

3 ,

, , , L' * '
p. ,

0:

Y, Qf .' 3 m

-October 25, 1989 Y '

NRC-89-0213 Page.31

[y <

p .

,~

!.^ 4 , through sensitivity studles, identification. of M vulnerabilities, etc. Selection of functional sequences will be reviewed and possibly modified at this tine.

o Modification of the event trees using the new software to i h.- reflect final sequence adjustments made following a R

~ third party review of the inicial model.

L 2e modeling of' internal flooding will be perforned upon conpletion of ps the Level 1 analysis. Wis will allow the flooding essessment to L

benefit from the detailed plant model developed through the remainder F of the Level 1 evaluation. S e flooding analysis will include k consideration.of potential flooding sources, pathways of flooding to l key equipment,:the potential for flood induced plant accidents or K' transients, the likelihood of system failure due to flooding, and the

' ability to detect'and mitigate postulated flooding conditions.

a There has been essentially no activity on the containment analysis portion'of the IPE. A consultant and @propriate software have yet to be selected. . Se ID00R IPFM will not be used.

Se containment-evaluation will include a plant-specific containment j, - event tree consistent with Section 2.2.2.5 of NUREG-1335, " Individual l Plant Examination; Submittal Guidance." Se potential containment i failure nachanisms of. Table 2.2 of'NURB:1-1335 will be treated in the  !

containment evaluation.

In performing the containment evaluation, the approach will utilize applicable information performed for plants similar to Fermi 2 and plant-specific analyses as needed. Wherever reference calculations l are adopted for applicability to Fermi 2, an evaluation of the principal differences between Fermi 2 and the reference facility will be performed to provide assurance that the results are directly applicable or to identify the general inpact of the differences.

H While the details of the containnent assessnent have not yet been l

established, the expected principal elements are described below.

Changes may be made when the detailed nethodology is selected and as experience is gained during the review.

i

?- -

Particular design features or potential vulnerabilities y

identii' led through a plant walxdown and design evaluation will be assessed to determine any impact such features might ,

,; , have on containment performance.

Containnent ultimate capability will be developed by evalua6.lon of the insights from containment performance E evaluations performed for other facilities. Identified ultimate capability controlling features in such assessments will be compared to the Fermi 2 facility to characterize the degree of applicability of these assesoments to the facility.

a NWF n' .

F  ;

g 1, A,, <

2 . L. i ML ,j 74

, , gg g 6 1;

' October: 25, - 1989  :

NRC-89-0213  !

Page 4  !

.t , i L

4 i

t

, .- A centainment event tree will be developed that includes  :

sufficient detail to reflect- the effects of plant and '

emergency procedute changes, yet is concise encogh to effectively communicate the results. %e containment event 1 tree .will be analyzed for each plant damage state. Success paths for recovery of degraded core conditions within the

+ reactor vessel will be included. . Thus, eniergency procedures '

and operator recovery actions during the phases of: severe accident progression up to the point of postulated -

containment failure will be incorporated in the analysis.

,This will include consideration of systen recovery and- .

repair. ' % rough this process, realistic insights for future 1 accident management evaluation will be gaina3. W e nodes of  ;

the containment event' tree will be quantified, and the m , release scenarios identified will be characterized in terms of: timing and severity level.  ;

o' ,

The timing and progresLion of identified sequences will be  :

assessed based upon either plant specific or representative calculations for plants similar to Fermi 2. '

In order to characterize the severity.of sequences postulated -

through the containment event tree formulation, a series of source term code results will be conpiled. Se impact of li pathway,' magnitude of release, rate of release and mitigation

.' effectiveness will be included in the daracterization of each containment event tree release path. Similar sequences l l will be combined to develop a comprehensive set of release 11 likelihoods.. If deemed appropriate and/or necessary, direct Ll plant specific source terms will be calculated for selected ,

L sequences. ^

l'  ;

L .- Finally, uncertainties will be addressed by either l' sensitivity analyses or by comparisons between results of l different containment performance codes enploying different approaches to describing complex phenomena or a combination h

of both. Emphasis will be placed on uncertainties in ,

phenomena which directly impact accidant w.agenent considerations. Less emphasis will be given to phenonena l which, while they might affect the source terra, do not  :

f 's strongly inpact the selection of strategies for coping with

. severe accidents, u .

L Upon conpletion of the Level 1 and containnent performance i

assessments, an evaluation of significant vulnerabilities and insights I will be performed. His evaluation will consider those sequences nesting the screening criteria defined Ir' Generic Ietter 88-20. The evaluation will include consideration of uncertainties to enhance the insights developed. Selected USI and GSI topics will be inc]uded, but at a mininum, USI A-45 aealing with decay heat renoval will be l

ga ,

f.,<.cg- l

t ' '.

A, + ,.-, . i USNE 4

e ..

. October 25, 1989

[ .NBC-89-0213 l

[- Page 5  ;

b ,

i O resolved through the-identification and evaluation of any decay heat s e, removal vulnerabilities. Se appropriate PRA insights referenced in #

I Appendix 5 to Generic Letter 88-20 will be considered in this effort.

K, '

DHCo will evaluate each potential accident vulnerability identified J through the performance of-this evaluation. Any identified h" '

' vulnerability that warrants correction will be thoroughly evaluated  ;

. for potential improvenents. Serious considerarion will be given to

. the proposed Mark I containnent perfornance inprovenents described in s, ,

, Enclosure 2 of Supplenent 1 to Generic letter 88-20 (Note that Fermi 2 l - already utilizes Revision 4 to the BWR Owners Group EPG, and DECO has 0

agreed to proceed with a hardened wetwell vent as discussed in the e response tu Generic Letter 89-16, NBC-89-0216, dated October 20,

  • 1989).

. We results of the review will be submitted to the hT using a two s tier approach. As described in Generic Intter 88-20, *he infornetion ,

submitted will includ) the results of the examination ar.i a stunn.ary of the insights gained. Detailed ckcunentation will be retaped by DEo end utilized in the ongoing risk evaluation and safety assescient program. . %e report submitted to the NBC will be consistent with the

~

general guidance in Table.2.1 of NUREG-1335. Referenced attached ,

i appendices.may be utilized to facilitate the use of do:unentation ,

't.1 ready available through the current Ievel 1 PRA.

Generic Ietter 88-20 notes that the " quality and comprehensiveness of the results derived from an IPE depend on the vigor with which the utility apolies the nethod of examination and on the utilities ,

consnitment to the intent of the IIE." DECO entarked on the current [

PRA program based on its own internal desire to evaluate the facility

, and to produce a risk-based tool to help evaluate proposed changes and to aid in the prioritization of resources. %e program is being  ;

developed by a dedicated DECO PRA group with assistance from i

additional site personnel and consultants as appropriate. Sus, the s , results of.the risk evaluation will become DECO products. Work is and p will continue to be reviewed by plant oprations personnel and system I.

engineers for consistency with design and operation.

t De level of involvenent and detail associated with the DECO progran

% requires careful and consistent developnent. The program la designed D to proceed at a pace that assures neaningful involvement by L appropriate operational and engineering personnel and to allow parallel utilization of the current model to support risk assessnent activities indeperdent of the IPE. De schedule telow is consistent with this level of involvenent in developent of the PRA products and evaluation of insights developed therefrom.

i>

o Completion of enhancenents - Nov. 1990 to the Ievel 1 analysis listed on page 2 (excludes internal flocds) .

h

DV, -

, x. < .

g,. ,

v.. , . , ,

c ,

s e ^' ..  :. .

g  : 5'.

  • 4 ' ' ggy3c - l October 25, 1989- i D. . NIC-89-0213 '  !

'L, 'Page'6-  ;

(

o Conpletion of Level 1 analysis for -

April 1991' ,

' internal floods. (

o~ Capletion of containment performance -

Oct. 1991 i analysis..

.o_ cmpletion of final assessnents and -

April 1992 M recomendations and submittal of the s ,

IPE report. +

r As the PPA prograin continues, DECO will' inform the NIC of any

'p '

significant findings and will provide brief summary progress reports upon coupletion of each of the three milestones that precede the final l J-L report..

.If you have any questions, please contact Mr. Arnold Jaufmann at (313) 586-4213.

Sincerely,

{

cc: :A. B. Davis- .

R. C. Knop

, , W.'G. Rogers e J.'F. Stang D. Modeen (NUMARC)

,a ,

e, s g

-,