Detection of Different Bearing Faults in a Permanent Magnet Synchronous Motor using PWT and LVQNN

This paper deals with the detection of bearing-related faults in permanent magnet synchronous motors. First, the motor current was sampled and analyzed using multi-level packet waveform transformation and then based on different features from the transformed signals, which resulting from the primary waveform, the LVQ neural network is trained for doing the classification. The outer-ring, inner-ring, cage, lubricant loss, and imperfect ball fault classes are considered in this paper. The proposed approach for fault detection shows high accuracy for practical cases.