Static Eccentricity Fault Diagnosis in Permanent Magnet Synchronous Motor Using Time Stepping Finite Element Method

This paper introduces a new index for noninvasive diagnosis of static eccentricity in permanent magnet synchronous motors (PMSM). Use of this index makes it also possible to precisely determine the eccentricity degree. The index is the amplitude of the harmonic components with a particular frequency pattern. Occurrence and increase of the fault degree cause the rise of amplitude of the harmonic components which can be used to diagnose the fault and determine its degree. To evaluate the ability of the proposed index for static eccentricity detection and estimation of its severity, the correlation between index and eccentricity degree is calculated. Then a three-layer artificial neural network is employed to classify the current and torque profiles to one of the four possible classes of eccentricities. After all, a white Gaussian noise is added to the both measured current and torque and robustness of the proposed index is analyzed with respect to the noise variance. A PMSM under static eccentricity fault is modeled using time stepping finite element method. This modeling includes all geometrical and physical characteristics of the machine components, non-uniform permeance of the air gap and non-uniform characteristics of the PM material. Use of this precise modeling makes it possible to access the demanded signals for a very high precision processing.