Anticipating human errors from periodic big survey data in nuclear power plants

Human errors are the critical factor for nuclear safety in nuclear power plants (NPPs). A human error anticipation method for nuclear power plant is proposed in this presentation. The main idea of the method is to catch a human error omen from survey data, so-called work condition survey (WCS), that the NPP collects periodically. The WCS data is so various and much because it is collected through questionnaires. A database is designed to process the big WCS data efficiently. A method for human error anticipation based on statistical methods is proposed and verified using the database containing WCS data collected for last 3 years. It is expected that the method contribute to monitor human errors likelihood and to plan human error prevention strategies.