Research on Multi-Sensor Information Fusion Technique for Motor Fault Diagnosis

Motor fault diagnosis is very important for revolver. Data fusion method is introduced into motor fault diagnosis in this paper. Various running parameters from different sensors when motor is running, back propagation (BP) neural network for motors sub-partial diagnosis, the global integration for partial diagnosis results by Dempster-Shafer(D-S) evidence theory, are introduced for the implementation of accurate motors diagnosis. The experimental results show that the method is very effective because the credibility of diagnosis is significantly increased, and the uncertainty decreased.