ML070330213
| ML070330213 | |
| Person / Time | |
|---|---|
| Site: | Vermont Yankee File:NorthStar Vermont Yankee icon.png |
| Issue date: | 02/21/2007 |
| From: | Dyer J Office of Nuclear Reactor Regulation |
| To: | Lochbaum D Union of Concerned Scientists |
| Stutzke M, NRR/DRA, 415-4105 | |
| Shared Package | |
| ml070460060 | List: |
| References | |
| TAC MD4116, Y020070019 | |
| Download: ML070330213 (3) | |
Text
February 21, 2007 Mr. David Lochbaum, Director Nuclear Safety Project Union of Concerned Scientists 1707 H Street NW, Suite 600 Washington, DC 20006-3962
SUBJECT:
VERMONT YANKEES PROBABILISTIC RISK ASSESSMENT
Dear Mr. Lochbaum:
I am responding to your letter of January 23, 2007, in which you used the phrase unique uniformity to characterize the Vermont Yankee probabilistic risk assessment (PRA) and to subsequently question its acceptability and usefulness. In addition, you asked what the U.S.
Nuclear Regulatory Commission (NRC) has done to verify that the Vermont Yankee PRA is using appropriate initiating event and failure rate data, whether from plant-specific or generic sources.
All nuclear power plant PRAs, such as Vermont Yankees, are unique in the sense that they have been developed to reflect the as-built, as-operated plant design and equipment configuration. PRA logic models consisting of event trees and fault trees and their associated reliability data, including the initiating event frequencies and component failure rates, are developed from plant-specific information such as plant-specific thermal-hydraulic analyses, general arrangement drawings, flow diagrams, electrical diagrams, operating procedures, training manuals, plant operating experience, and other similar plant-specific information. As an added step to ensure that the PRA is based upon the as-built, as-operated plant design and equipment configuration, the PRA logic model is checked through plant walkdowns.
As you point out in your letter, it is desirable to use the actual operating experience of the plant being modeled as the basis for estimating the PRAs data. However, the available operating experience may be sparse, that is, the plant may not have experienced some of the initiating events or component failures that have been included in the PRA. The lack of operating occurrences makes it difficult to develop statistically meaningful estimates of initiating event frequencies or component failure rates. For example, consider the difficulty in estimating the frequency of a large break loss-of-coolant accident since such an event has never actually occurred at any nuclear power plant. Similar to many nuclear power plant PRAs, the Vermont Yankee PRA uses Bayesian analysis to estimate initiating event frequencies and component failure rates. Bayesian analysis is a statistical method that combines generic initiating event frequencies and component failure rates with Vermont Yankees actual operating experience, for example, reactor trip events and component failures. The use of Bayesian analysis to support nuclear power plant PRAs satisfies the Standard for Probabilistic Risk Assessment for
D. Lochbaum Nuclear Power Plant Applications, American Society of Mechanical Engineers (ASME) RA-Sb-2005, which was developed by ASME and endorsed by the NRC in Regulatory Guide 1.200.
Bayesian analysis is a widely accepted statistical method that has been used in many scientific and engineering analyses, and its use is not unique to the Vermont Yankee PRA or to nuclear power plant PRAs in general.
There is no existing regulation that requires any current licensee to maintain a plant-specific PRA, and the NRC staff does not routinely or periodically review the licensees PRAs.
However, the NRC staff may review elements of a licensees PRA when the licensee submits a license amendment request (LAR) to the NRC. If the LAR is risk-informed and submitted in accordance with Regulatory Guide 1.174, then the staff will review elements of the PRA that are impacted by the LAR by using Chapter 19 of the Standard Review Plan (NUREG-0800). If the LAR is not risk-informed, for example an LAR for extended power uprate, then the NRC staff may, in accordance with Appendix D to Chapter 19 of the Standard Review Plan, review elements of the PRA in order to determine if the proposed LAR raises special circumstances that rebut the presumption of adequate protection afforded by compliance with regulations.
In 1995 the NRC staff reviewed the initiating event frequency and component failure rate data used in the Vermont Yankee PRA as part of its review of Vermont Yankees individual plant examination (IPE), which was submitted in response to Generic Letter 88-20. This review examined the process used by the Vermont Yankee licensee to collect and analyze the data, compared the data to generic sources, and concluded that the data used in the IPE was consistent with the data typically used in other PRAs and IPEs. Starting in 2002, the NRC staff reviewed Vermont Yankees request for an extended power uprate (EPU), which considered in part how the EPU would affect the initiating event frequencies and component failure rates used in the PRA. In fact, the licensee identified one initiating event frequency that required revision due to changes resulting from the EPU. The staff did not identify any additional changes to initiating event frequencies and component failure rates due to the EPU.
In conclusion, consistent with the Commissions PRA Policy Statement (Use of Probabilistic Risk Assessment Methods in Nuclear Regulatory Activities; Final Policy Statement, 60 FR 42622, August 16, 1995), the NRC staff does not make regulatory decisions based solely on the numerical results of PRAs. Rather, PRA results and their associated uncertainties are inputs into our risk-informed regulatory decisionmaking process.
Should you have any further questions regarding this matter, please contact John A. Grobe, Associate Director for Engineering and Safety Systems, Office of Nuclear Reactor Regulation (301 415-1274; jag@nrc.gov).
Sincerely,
/RA John A. Grobe for/
J. E. Dyer, Director Office of Nuclear Reactor Regulation
(Package): ML070460060 ADAMS Accession No. (Incoming): ML070300617 ADAMS Accession No. (Response): ML070330213 NRR-106 OFFICE APLA/DRA BC:APLA/DRA D:DRA AD:ADES/NRR D:NRR NAME MStutzke MRubin
/RA D. Harrison for/
CHolden JGrobe JDyer
/RA J. Grobe for/
DATE 02/01/07 02/09/07 02/15/07 02/21/07 02/21/07