Evaluation of Component Failure Data of the Operating Nuclear Power Plants in Korea Based on NUREG/CR-6928

This paper focuses on ensuring the quality of component failure data. When performing data analysis in PSA, we have customized the component failure data based on Bayesian analysis using plant specific experiences and the generic data of Advanced Light Water Reactor Utility Requirements Document (ALWR URD). However, ALWR URD was established by collecting US nuclear power plant (NPP) practices from mid 1980s to early 1990s. We analyzed the component failure data using the raw data of component failures in Pressurized Water Reactor (PWR) plants by 2012. This paper presents the results from analyzing the component failure data based on the new generic data and the latest specific failure data. We also compare the new component failure data to the existing data of PSA models, and evaluate the risk impacts by applying the new data to the PSA models of reference NPPs in this paper. To apply the new generic data source to PSA models, we reviewed and compared NUREG/CR-6928 and the existing generic data source, ALWR URD. In addition, we analyzed the component failure data generated from 16 PWR plants by the end of 2012, and performed the Bayesian update with these raw data based on the new generic data source of NUREG/CR-6928. Also, we reviewed the PSA models of the reference NPP, and identified some important components to CDF. The failure data of the major components decreased in general by applying the new generic data and the latest plant specific data. As a result, the CDF of the reference NPP decreased over 30% compared to the value of the existing CDF