Using CEEMDAN algorithm and SVM fault diagnosis methodology in three-phase inverters of PMSM drive systems

Abstract

In this paper, a fault diagnosis method of three-phase inverters of permanent magnet synchronous motor(PMSM) is proposed, which is based on the complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN) and support vector machine(SVM) according to the measured α — β phase currents. This method can get higher diagnostic accuracy and have preferable against disturbances compared with other methods. Previously, wavelet denoising is introduced to three-phase currents for denoising and preprocessing before Concordia transform. Then the feature of these proposed phase currents are extracted by CEEMDAN algorithm and Hilbert transform, and the faults are detected and diagnosed using SVM approach. Simulated data are used to train the fault diagnosis model, as well as validate the proposed fault diagnosis methodology. The simulation results are presented to validate the effectiveness of the method.