Gear Fault Diagnosis for Vehicle Electric Drive Systems Based on Stator Currents

With the popularization of electric vehicles, the safety performance of electric vehicles has drawn much attention. However, the gears of electric vehicles are more prone to failure at high speeds, which can affect the safety performance of the vehicle. This topic proposes a electromechanical coupling model, which is composed of a permanent magnet synchronous motor model, a vehicle longitudinal dynamics model and a transmission system model, and will be applied to gear fault diagnosis. First, the sensitivity of the gear fault to the stator current signal, the electromagnetic torque signal and the q-axis current signal is investigated based on the time-varying meshing stiffness obtained by the potential energy method. The discrete wavelet algorithm is used to decompose the stator current signal, and the d₁ component with obvious fault information is obtained. Then, the singular spectral entropy is selected to realize the feature extraction of the stator current signal by comparing the influence of the fault degree on different feature parameters. Finally, the support vector machine is used for the identification of gear faults, and the accuracy of the identification reached 93.3%. The results show that support vector machine can be used as a high precision algorithm for gear fault identification, thus improving the safety performance of electric vehicles.