Association rules analysis of human factor events based on statistics method in digital nuclear power plant

Abstract

With human factor events rising in recent years, many researches begin to pay much attention to them. Especially, human factor events in nuclear power plant show more important than other human factor events. To effectively decrease human factor events, the authors propose the method of association rule analysis of human factor events in this paper. Association rule is one of the most popular data mining techniques applied to many scientific and industrial problems. Based on traditional methods, the authors propose a weight association rule based on statistics. Weight factors consist of inner and exterior human factors. In this paper, the authors propose a dynamic function and some methods with weight in order to assess support, confidence and correlation degree among human factor events. The proposed methods are tested by experiments. From results of experiments, we can easily find higher error rate events caused by human, higher confidence and correlation degree events among human factor events of steam generator tube rupture (SGTR) of nuclear power plant (NPP).