Comparison of two approaches for seismic probabilistic safety assessment

Abstract

Seismic probabilistic safety assessment(SPSA) is one of the methods for evaluating the safety of nuclear power plants for the seismic event. It consists of the accident sequence and consequence evaluation based on the hazard analysis and fragility analysis to estimate the core damage frequency. In general, analysts develop a primary event tree and secondary event trees in the accident sequence evaluation. The primary event tree considers the seismic fragilities of the major structure, system and component(SSCs). Corresponding to the initiating events, each sequence in primary event tree can be assumed to result in direct core damage or be linked to a secondary event tree for more detailed analysis. There are mainly two ways to perform SPSA described by Reed et al.(1989) and guided by USNRC(2008). The methods can be distinguished according to how the hazard and fragility curves are applied. The first method is that the fragility curves of SSCs are integrated with the hazard curve. In Korea, the software called Probabilistic Risk Assessment of Systems for Seismic Events(PRASSE) is widely used to compute the plant level fragility curve and perform the convolution integration over ground motion levels. However, when a detailed analysis is required and linked to secondary event trees, PSA software packages such as AIMS-PSA are used. In the second method guided by USNRC(2008), the ground motion levels are partitioned into several bins to perform the discrete calculation. The former may calculate some of sequence frequencies more accurately, while there exist disconnections between the primary and secondary event trees which result in difficulties in reviewing the minimal cutsets. Although the latter produces more conservative results, it has advantages in dealing with dependencies among failure events. The information of a failed SSC from the primary event tree can be linked to the secondary event trees since the primary and secondary event trees can be combined. Therefore, the method can be more convenient in assessing the results and reviewing the minimal cutsets. It is also applicable to multi-unit risk assessment because the method is easier to connect one model with others. In this study, we provide our experience in SPSA with the two methods. © ESREL 2020-PSAM15 Organizers. Published by Research Publishing, Singapore.