EV Starter Motor Fault Detection Using DWT Skewness and Kurtosis

The present paper proposes a new diagnostic technique for determining short circuit conditions in automobile starter motor armature coils by computing the skewness and kurtosis of current waveforms filtered by wavelet transform. Using MATLAB Simulink, we developed a model to simulate an automobile system with incremental inter-turn errors in the starter motor, enabling the acquisition and analysis of signals at different fault stages using discrete wavelet transform (DWT). This method diagnoses faults early in the transient anomalies by analyzing signals in different resolutions. The skewness and kurtosis of wavelet coefficients, essential for fault diagnosis, provide critical information for distinguishing normal, pre-fault, and fault states. This new diagnostic approach enhances the precision of problem identification, improving safety and dependability in automotive repair.