A Hybrid Fault Detection and Diagnosis Algorithm for Five-Phase PMSM Drive

Five-Phase Permanent Magnet Synchronous Motors (5 phi-PMSM) are frequently employed for fault-tolerant applications such as aircraft, ship propulsion, electric vehicles, and locomotive traction due to their high reliability. To ensure the continuity and reliability of the service of the motors, the most critical factors considered are faults detection and diagnosis betimes. Inter-Turn Short-circuit Fault (ITSCF), the most dangerous fault in a motor, can spread to other serious faults and cause the entire system to shut down if it is not recognized and addressed in time. This paper proposes a hybrid fault detection and diagnosis algorithm (signal-based and knowledge-based) for ITSCF in a 5 phi-PMSM. Discrete Wavelet Transform (DWT) is applied for fault detection, and machine learning (Fuzzy Logic system) is used for fault diagnosis. Two cycles of stator current signals are considered for continuous analysis, and fault can be detected and diagnosed within two cycles of the stator currents.