Induction Motor Fault Diagnosis Based on SSA-SVM

In response to the problem of low fault identification rate of induction motor, An induction motor fault diagnosis method based on the combination of fast overall average empirical modal decomposition (FEEMD) and support vector machine (SSA-SVM) optimized by sparrow search algorithm is proposed. First, the stator current is decomposed into intrinsic modal components (IMFs) of sequentially decreasing frequency by FEEMD, and then the IMF components with larger correlation coefficients are selected by correlation coefficient method and the energy entropy and sample entropy are calculated as the eigenvectors, which are then inputted into the SSA-SVM model in order to receive the diagnosis results. The results show that the fault diagnosis accuracy of SSA-SVM model reaches 96.7%, which has higher accuracy and shorter time compared with the two models of Grey Wolf Algorithm (GWO) optimized SVM and Particle Swarm Algorithm (PSO) optimized SVM, which verifies that the method is a reliable method for fault diagnosis of induction motors.