Fault Detection of Stator Windings for PMSM Based on Fast Fourier and Discrete Wavelet Transform

Abstract

Fault detection in permanent magnet synchronous motor (PMSM) is important and necessary to ensure reliability, performance and reduce downtime. In this study, a mathematical model is developed that describes the healthy and faulty states of the PMSM, including inter-turn short circuit (ITSC) faults. This dynamic mathematical model is simulated and its performance and behavior under different fault conditions are studied. Fast Fourier transform (FFT) technique is employed to analyze the frequency spectrum of motor speed and stator current signals, which allows us to detect harmonic distortions caused by short circuit faults in the windings. However, FFT technique is not sufficient for non-stationary conditions. To address this, the discrete wavelet transform (DWT) is utilized to analyze motor signals across various frequency ranges, which allows the detection of faults under non-stationary conditions and improves the accuracy of diagnosis under real operating conditions. The effectiveness of the studied techniques is demonstrated through simulation, with DWT showing superior performance in detection of PMSM faults.