Fault diagnosis approach based on fuzzy probability SDG model and reasoning

The fault diagnosis approach based on signed directed graph(SDG) has better completeness and explanation facility, and has the disadvatage of the lower diagnostic resolution. Therefore, the semi-quantitative fault diagnosis approach is proposed based on the model of fuzzy probabilistic SDG and Bayesian inference. The node variable is expressed as fuzzy variable. The cause-effect relationship between the nodes is described by conditional probabilities table (CPT). The set of failure source candidates is found out by using Bayesian inference and backtracking algorithm. Furthermore, the candidates in the set are ranked according to the rate of fault possibility. The primary electrical power supply system in certain a satellite is modeled with the proposed approach. The diagnosis simulation results show that the diagnostic resolution can be improved significantly, and the approach is feasible to be applied to on-board diagnosis for spacecraft.