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

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