A Hybrid Fault Detection and Identification Method for Current Sensor Faults in PMSM Drives

Current sensors are essential for precise control of PMSM drives, providing feedback for torque, speed, and position control. Sensor faults can lead to inaccurate measurements, causing instability, performance degradation, or system failure. Effective fault identification is needed to reduce the risk of damage to motors, inverters, or other components. In this paper, a new hybrid fault detection and identification method is proposed. It combines model-based and data-driven approaches for fault detection and identification of the most common current sensor faults. The model-based approach is used to detect the fault and the data-driven network is used to identify the current sensor fault type. The effectiveness and accuracy of proposed method is validated experimentally.