Research Of Motor Fault Diagnosis Based On D-S Evidence Theory And Bayesian Network

Motor as the driving equipment most widely used in all kinds of industries since it hasbeen produced. Especially if the key parts of the motor problems, will have an importantimpact on the whole production line, and caused huge economic losses. Motor is anextremely complex integrated electrical system, and many types of electrical equipmentusually work in extremely harsh environments, therefore, motor fault diagnosis systemusually exist unascertained information. Therefore, the most important problem of solvingmotor fault diagnosis is to solve the uncertain problem through the research of motor faultdiagnosis technology, and it can guarantee the normal operation of electrical equipment.Bayesian network method in information fusion to become the best method forsolving uncertain problems of the motor fault diagnosis, and it based on Bayesian theory.Evidence theory can be combined for a variety of fault information at different levels. Itcan enhance mutual support among evidence, and then improve the accuracy of motor faultdiagnosis, so it has become a new hot spot in the current research of motor fault diagnosis.This paper provides a new fault diagnosis model which is the combining of Bayesiannetwork and Evidence theory, namely the Bayesian network-Evidence theory faultdiagnosis model. It can effectively improve the accuracy of motor fault diagnosis.A new kind of motor fault diagnosis method is proposed through the analysis of theuncertainty problems of motor failure, which utilize the parallel Bayesian network and D-Sevidence theory based on multi-source information fusion technique. In this method the setof motor fault features is divided into multiple fault sub-spaces and we uses differentdiagnostic Bayesian networks to locally diagnose motor faults for the fault sub-spaces.Then taking the result of local diagnostics as the independent evidence body, and finallyfusing the evidence bodies in the decision level according to the D-S evidence theory.Finally, the motor fault diagnosis result was given through certain fault diagnosis strategy.The method provides by this paper can solve the uncertain information of the motor faultdiagnosis system effectively; meanwhile, it can be obtained more accurate and efficient diagnostic results. Through test analysis verifies the fault diagnosis fusion model andalgorithm is effective and specific application.