Applying Bayesian networks in nuclear power plant safety analysis

Abstract

Over the last decade, Nuclear energy has become one of important energy. Nuclear power systems become more complex and traditional safety methods are hard to be applied. This paper presents a novel approach for nuclear power plant safety analysis which called Bayesian Networks(BN). The BN model is constructed based on the combination of Failure Mode, Effect Analysis (FMEA) and Fault Trees Analysis(FTA). The probability of the model’s root nodes is estimated by Bayesian estimation method and Monte Carlo simulation. Bidirectional inference and sensitivity analysis of the model is also researched. At last, we use a case study to show the method’s advantages compared with traditional methods in nuclear power plant safety analysis.