ML19256A587

From kanterella
Jump to navigation Jump to search
Responds to Request for NRC Views on Results of Survey of WASH-1400.Recommends WASH-1400 Use & Believes Safety Analysis & Acceptance of Nuclear Power Plants Should Be Done W/Probabilistic Event/Fault Tree Approach
ML19256A587
Person / Time
Issue date: 12/27/1978
From: Bennett G
NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
To: Budnitz R
NRC OFFICE OF NUCLEAR REGULATORY RESEARCH (RES)
Shared Package
ML19256A581 List:
References
RTR-WASH-1400 NUDOCS 7901090057
Download: ML19256A587 (1)


Text

.

/

DEC2 71978 10:

Robert J. Budnitz Deputy Director, RES FROM:

Gary L. Bennett, Chief Research Support Branch, RSR

SUBJECT:

STAFF VIEWS Otl RESULTS OF SURVEY OF USES OF WASH-1400 In response to your memorandum of December 22, 1978 on the same subject, I would like to corment on item (b) in Mr. Gossick's memorandum of December 20, 1978.

With regard to reconmendations for reconsideration of certain uses, I list the following:

1.

The safety analysis and acceptance of nuclear power plants should be done with a probabilistic event / fault tree approach.

2.

The introduction of regulatory requirements, (including regulatory guides, branch technical positions, standards and the Code of Federal Regulations) should only be done after a thorough value/ impact analysis which includes a probablistic event / fault tree systems analysis. This recommendation also applies to proposed backfitting of hardware.

3.

Formal decision analysis techniques, including event / fault trees, should be required for every allocation of resources.

While one can fault the WASH-1400 approach as an overextension of the methodology, to me the alternative is worse:

decisions made with no explicit consideration of the decision maker's values or the consequences of the decision.

I have sat in Regulatory Requirements Review Committee meetings when decisions to impose additional requirements were made on the basis that it made certain people feel better.

I emphatically disagree with this method of regulating a multibillion dollar industry.

f;RC must discipline itself to act in a responsible, thoughtful manner. We must focus our attentions on the real substantive issues and not have our efforts diluted by a myriad of "nonissues".

I believe that a disciplined application of the probabilistic risk analysis technicues can aid us to focus our limited resources on the key issues.

The ultimate advantage of the event / fault tree methodoiogy is that it forces people to look beyond their own narrow interests to the total system.

4 Gary L. Bennett, Chief f'

Research Support Branch, RSR 6

cc's (see next page) 790109 oM7

DECP 71c 9 cc:

Saul Levine Thomas E. Murley F. J. Arsenault A.-R.

Buhl R. M. Scroggins L. S. Tong C. fl. Kelber

.