Fault Isolation Based on Subjective Bayes´ Reasoning for Redundant Actuation System

Directing to the difficulties of fault isolation based on mathematical model, this paper presents a new fault isolation method that utilizes the limit apriori knowledge of expert, infers the failure process recursively and realizes the fault isolation with Bayes´ reasoning. Through dividing the system into several levels hierarchically according to their information flow, the system failure process can be reasoned from bottom to up with Bayes´ sufficient factor LS and Bayes´ necessary factor LN, and find the fault place with likelihood rate by the Bayes´ reasoning.. Because the subjective Bayes reasoning can strength the fault information and eliminate specious information, this method is suitable to determine the failure location for complicated system. Application of tri-redundant actuation system indicates that the subjective Bayes´ reasoning can accomplish the fault isolation effectively.