Application of Particle Swarm Optimization to PMSM Stator Fault Diagnosis

Permanent magnet synchronous motors (PMSM) are frequently used to high performance applications. Accurate diagnosis of incipient faults can significantly improve system availability and reliability. This paper proposes a new scheme for the automatic diagnosis of turn-to-turn short circuit faults in PMSM stator windings. Both the fault location and fault severity are diagnosed using a particle swarm optimization (PSO) algorithm. The performance of the motor under the fault conditions is simulated through lumped-parameter models. Waveforms of the machine phase currents are monitored, based on which a fitness function is formulated and PSO is used to identify the fault location and fault size. Simulation results in MATLAB provide preliminary verification of the diagnosis scheme.