A Wavelet-Based Fault Diagnosis Approach for Permanent Magnet Synchronous Motors

This paper presents a generic analytical tool to diagnose and classify faults in permanent magnet synchronous motors. The proposed method deploys wavelet transform to extract features for fault diagnosis, which provides a framework for studying nonstationary trends. Analyzing the stator current with wavelet raises a challenge since the energy of fundamental component spreads over different scales in the decomposition. An adaptive filter is used to estimate and remove the fundamental component in stator current. This filter predicts the main harmonic by processing the measured current data in real-time without any speed feedback. The proposed filter is designed in a way that it does not affect or suppress fault related harmonics. The estimation accuracy and convergence rate of this filter is tested and reported by error bounds, which exhibit an acceptable robustness. The validity of the proposed fault diagnosis approach is verified by finite element simulations and experimental results. The effectiveness of this algorithm is tested using two case studies including broken magnet and eccentricity faults. An average accuracy above 96% is obtained using experimental and simulation data. It is proven that the filtering scheme increases the overall accuracy of fault diagnosis.