Application Of Information Fusion Technology In Motor Fault Diagnosis

Motor is one of the most widely used power equipments in modern society production. The study of motor fault diagnosis has great significance for industrial production and our normal life. Information fusion is a new subject, and it is a technology of high synthesis and practicality. With the rapid development of basic theory which integrates system theory, control theory, information theory, D-S theory of evidence, and use technology which integrates computer technology, network technology, sensor technology, technology of information fusion is also widely used in the field of fault diagnosis. This paper studies technology of information fusion based on neural network and its application in motor fault diagnosis system.In order to reflect the operating status of motor from more aspects, automatic identity and accurate diagnose motor failure mode, motor fault diagnosis system is built through combination of data fusion technology and neural network. According to large number of references, field survey and operation principle of motor, this paper analyzes the common motor faults and its mechanism systematically, determines characteristic parameters of various fault status; and acquire a large number of experimental data from various fault simulation experiments, it is used to build fault diagnosis system. First of all, fault features are classified by the signal separation capability of neural network; To obtain independent evidence, fault feature level fusion and the locate motor fault diagnosis is carried out by Three-BP neural network; Secondly, all evidences are integrated by D-S evidence theory fusion algorithm. Therefore the accuracy of motor fault diagnosis is achieved; the reliability of diagnostic results is improved.Finally, this paper designs a motor fault diagnosis system in accordance with theoretical analysis and experimental data of simulation, the system can be used to diagnose and analyse motor fault.