The US NRC's accident sequence precursor program: An overview and development of a Bayesian approach to estimate core damage frequency using precursor information

This paper presents an overview of the US Nuclear Regulatory Commission's (NRC) Accident Sequence Precursor Program. The objectives of the program are delineated: the screening, review and analysis methods for events involving an accident initiator are reviewed. The quantitative calculations are illustrated using an ASP event. A Bayesian framework for using ASP information to estimate annual core damage frequencies is then introduced. This method uses both failure and success information to estimate the core damage frequency. In addition, an alternative procedure is proposed to estimate the frequency of severe core damage accidents. The estimated occurrence rate is then used to define a retrospective accident sequence precursor measure of closeness to a core damage accident. The retrospective measure is the a posteriori probability of at least one core damage accident during a given period of time. Expressions for both the conditional and unconditional severe core damage probabilities are developed. The paper also shows that the sum of conditional severe core damage probabilities overestimates the rate at which severe core damage events occur.