Advanced Eccentricity Fault Recognition in Permanent Magnet Synchronous Motors Using Stator Current Signature Analysis

In this paper, a novel index is introduced for static and dynamic eccentricity fault diagnosis in permanent magnet synchronous motors (PMSMs). The proposed index is a linear combination of the energy, shape factor, peak, head angle of the peak, area below the peak, gradient of the peak of the detail signals in wavelet decomposition, and coefficients of the autoregressive model, which are extracted from the stator current signature analysis. Principal component analysis is applied to the features as the linear transform for dimension reduction and elimination of linear dependence between the features. In order to demonstrate the capability of these indexes to estimate eccentricity type and degree, the fuzzy support vector machine is employed as a classifier. Classification of the results indicates that the nominated index can be utilized to detect eccentricity occurrence, recognize its type, and determine its degree precisely. Since extraction of efficient indexes closely depends on precise computation of necessary signals, the time-stepping finite element method is utilized to model the PMSM under eccentricity fault and calculate the stator currents as a proper signal for processing. Simulation results are verified by the experimental results.