Bayesian reliability evaluation method for the rotor

Abstract

This paper presents a reliability analysis method based on Bayesian network in order to determine the weak parts of the system and enhance its reliability. The Bayesian network model is converted from its corresponding fault tree representing the system conditions and failures, which is used to calculate the occurrence probabilities of all nodes under different conditions. Four kinds of familiar faults and their corresponding working conditions of the rotor are extracted, so the rotor's fault tree and Bayesian model are established. Using this model, the conditional probabilities of the faults occurred are computed when the observation of the rotor is presented. Meanwhile, the impacts on the system reliability due to changes of the states are analyzed. The experiment results prove the effectiveness of the proposed method.