ML16217A369

Draft Policy Statement on the Use of Probabilistic Risk Assessment Methods in Reactor Regulatory Activities

ML16217A369
Person / Time
Issue date:	05/11/1994
From:	Kress T Advisory Committee on Reactor Safeguards
To:	Taylor J NRC/EDO
References
D940511
Download: ML16217A369 (3)

Text

D940511 Mr. James M. Taylor Executive Director for Operations U. S. Nuclear Regulatory Commission Washington, D.C. 20555

SUBJECT:

DRAFT POLICY STATEMENT ON THE USE OF PROBABILISTIC RISK ASSESSMENT METHODS IN REACTOR REGULATORY ACTIVITIES During the 409th meeting of the Advisory Committee on Reactor Safeguards, May 5-7, 1994, we reviewed the current draft Policy Statement on agency usage of probabilistic risk assessment (PRA).

We had the benefit of discussions with representatives of the NRC staff. We also had the benefit of the documents referenced.

We are in general agreement with the Policy Statement. It appears to present an appropriate position on the use of PRA in the regulatory process. We are, however, concerned with some aspects of the Policy.

Some provisions of the Policy Statement are crafted in rather weak language. For example, we believe that in Item (2) of Section II, Policy Statement, the word "may" ought to be replaced by "should" to make a commitment to increase the use of PRA to help eliminate unnecessary conservatism associated with current regulatory requirements.

The Policy is very general and does not provide any specific guidance or plan for the expanded use of PRA in regulatory activities. This has apparently been relegated to an "implementation plan" which is referred to in the Policy Statement.

We hope that this plan will provide some specific and definitive elements to guide the use of PRA in the regulatory process. We recommend that the implementation plan be submitted for public comment along with the Policy Statement.

The draft Policy Statement seems to draw a distinction between the traditional regulatory process (commonly known as "deterministic")

and the PRA approach. This common perception causes some in the regulatory arena to be skeptical of and reluctant to embrace the PRA approach. However, we believe that treating the PRA approach as a distinct and unique method compared to the traditional approach is inappropriate and misleading. We believe that the PRA approach should be considered as an extension and enhancement of traditional regulation rather than a separate and different technology. Certainly, the deterministic approach is replete with implied elements of probability, from the selection of accidents to be analyzed (e.g., reactor vessel rupture is too improbable to be considered) to the requirements for emergency core cooling (e.g.,

safety train redundancy and protection against single failure).

The PRA approach enhances traditional approaches by considering risk in a coherent and complete manner, thereby providing a method

to quantify the overall level of safety.

We agree that there are uncertainties, limitations, and omissions with the PRA approach. However, we think it is important to understand that these uncertainties are derived from knowledge limitations. These knowledge limitations were not created by PRA, but rather were exposed by it. These limitations existed during the traditional regulatory approach, some were unknown, others only vaguely understood. Attempts were made to accommodate these limitations by imposing prescriptive and what was hoped to be conservative regulatory requirements. The PRA approach has exposed these limitations and has provided a framework to assess their significance and assist in developing a strategy to accommodate them in the regulatory process. We are pleased that these issues are identified in the Policy Statement and that they are being addressed in the implementation plan.

One of the more important shortcomings of PRA use was not identified in the Policy Statement. This is the misuse and misapplication of PRA results stemming from an incomplete and/or flawed analysis. While those in the nuclear regulatory arena have done an excellent job in many instances in applying and using PRA, there have been examples where this has not been the case. Among the more important of these are some of the cost/benefit analyses for backfits. We recognize that these analyses are difficult. We urge the staff to assign high priority in the implementation plan to improving and adding consistency to cost/benefit analyses.

We further believe that the implementation plan needs to address the need for PRA research to help assure that the PRA state-of-the-art is at a level consistent with the intended PRA usage in the agency. We intend to further consider the area of PRA research needs in the near future.

In conclusion, we reiterate our support for the overall thrust of the PRA Policy Statement and the allocation of resources to implement it. We would like to be kept informed of the progress in developing the implementation plan.

Sincerely, T. S. Kress Chairman

References:

1. Memorandum (Undated) from James M. Taylor, Executive Director for Operations, for The Commissioners,

Subject:

Draft Policy Statement on the Use of Probabilistic Risk Assessment Methods in Reactor Regulatory Activities, received May 5, 1994 (Predecisional)

2. Memorandum dated April 14, 1994, from Martin J. Virgilio, Office of Nuclear Reactor Regulation, to John T. Larkins, Executive Director, ACRS,

Subject:

PRA Draft Policy Statement, with Predecisional Enclosure

3. U.S. Nuclear Regulatory Commission, Policy Statement dated January 18, 1979,

Subject:

NRC Statement on Risk Assessment and The Reactor Safety Study Report (WASH-1400) In Light of the Risk Assessment Review Group Report