Comparison of DC link current and stator phase current in inverter switching faults detection of PMSM drives in HEVs

Abstract

In this paper, a robust method for detection of inverter short circuit faults in field oriented control (FOC) of permanent magnet synchronous motor (PMSM) drive that is used frequently in hybrid electric vehicles (HEV) thanks to high power density and high efficiency is introduced. In the proposed method, a fault diagnostic and protection technique is developed by using discrete wavelet transform (DWT) and artificial neural network (ANN). The Symlet2 wavelet is selected to perform stator current analysis and the DWT coefficients of the motor currents are used as inputs of ANN for detecting inverter faults in PMSM drives. In HEVs, motor faults can cause permanent damage and accidents. So the early detection of PMSM drive faults in HEV will be provided by the proposed method. For implementing this method, the Matlab program is used to process DWT of signals. DC link and phase currents are compared in inverter switch faults detection and the results show that the proposed technique is very effective for fault diagnostic. Keywords: Inverter switch fault, DC link current, DWT-based fault diagnosis, Hybrid electric vehicle, Artificial neural network, Field oriented controlled PMSM