Development of new performance shaping factors' taxonomy for a human reliability analysis

Abstract

Analyzing human reliability is significant to evaluate plant safety risks. The value of human error probability assessed by an HRA becomes a basic event of the probabilistic safety assessment (PSA). Therefore, it is necessary to determine human error probability accurately to assess the core damage frequency. Many HRA models have been developed for determining human error probabilities accurately but most existing HRA methods have been developed for moderate situations. Analyzing human reliability under extreme conditions has not been performed because most accidents have not brought about extreme consequences. However, human performance was revealed as a key problem in the Fukushima accident. Therefore, comprehending factors that affect human performance under extreme conditions becomes significant. In this study, we scrutinize key differentiating characteristics between moderate and extreme situations during an accident. In addition, we are going to derive dominant performance influencing factors from task analysis, human psychology, and human situation awareness studies. Selected dominant performance influencing factors will become key elements of the new HRA method under extreme situations. Precise human error probabilities assessed by the HRA method will be greatly helpful for probabilistic risk assessment for the nuclear power plant.