Abstract:

A new equipment fault diagnosis method was proposed. In this method, we established the fault diagnosis Bayesian network through the combination of the fault tree with Bayesian network, and used joint tree to reason out the fault diagnosis based on the diagnosis fault Bayesian network. The method can solve the problems that modelling hardly and weakly reasoning ability caused by respectively using fault tree and Bayesian network. Fault diagnosis to a deck lighting system of a ship was carried out, and then the probability distributions of fault nodes were calculated, as a result, the fault of deck lighting system can be located quickly and accurately.

Key words: fault tree, Bayesian network, junction tree, fault diagnosis