Switch open-circuit faults diagnosis of inverter based on wavelet and support vector machine

Abstract

The paper presents a technique of fault detection and diagnosis for open-circuit fault of power switch in inverter-fed motor drives based on wavelet and support vector machine (SVM). The output current of the inverter is processed by wavelet transform, the coefficient value of the recursive wavelet transform is obtained and the value can be used for fault detection. The layer detail coefficient and its energy are obtained by wavelet transform and the fault features are got by normalized the energy. Tri-class SVM is used to isolate switch faults. The classification accuracies obtained from the testing samples are 95.6 and 93.3% with SVM1 and SVM2, respectively. The simulation results show that the method can detect and isolate the faults effectively.