An evaluation method for a seismic-induced multi-unit initiating event extending an initiating event classification tree for a nuclear power station

Abstract

It is crucial to perform a multi-unit probabilistic risk assessment (MUPRA) to identify the risk of a nuclear power station (Hereafter referred to as site) with multiple units. External hazards play a significant role in MUPRA because they simultaneously affect multiple units. In Japan, an earthquake is a critical external hazard in site risk. In this study, we propose a method to assess seismic-induced multi-unit initiating events (SIMUIEs). To utilize insights and results from seismic single-unit probabilistic risk assessment (SUPRA), the proposed method identifies SIMUIEs and estimates their frequencies by extending an initiating event classification tree method (also known as a hierarchical event tree method) commonly used in Japan. The proposed method generates SIMUIEs by combining seismic initiating events considered in seismic SUPRAs. In the SIMUIE generation process, the following factors specific to a multi-unit site are considered: (1) failure of shared equipment among units, (2) cascading effects across units, and (3) correlation of responses and correlation of capacities of structures, systems, and components (SSCs) among units. Then, the SIMUIEs are identified by the combinations considering these factors. In quantification, a logical model is built for each SIMUIE, and they are quantified by considering the dependences among seismic-induced initiating events. In addition, a preliminary evaluation for two nuclear power reactors demonstrates how the proposed method can identify and quantify SIMUIEs.