A new risk assessment method based on belief rule base and fault tree analysis

Abstract

Risk assessment methods are often used in complex industrial systems to avoid risks and reduce losses. The existing methods have not effectively solved the problems of lack of evaluation data and the interpretability of the entire evaluation process. This paper proposes a new risk assessment model based on the belief rule base (BRB) and Fault Tree Analysis (FTA). The FTA algorithm overcomes the difficulties of traditional BRB model in obtaining expert knowledge, clear indicators, and establishing logical relationships. This method establishes FTA rules based on the BRB model and expands the knowledge base through the FTA algorithm. A Bayesian network is applied as a conversion bridge between the FTA and BRB model. In addition, the model is optimized to reduce the uncertainty in the model. The method proposed is described by a case and its effectiveness is verified.