Comparative Study of Online Signal-Based Fault Detection Methods for Pmsms With Interturn Short-Circuit Faults

This paper compares two online signal-based fault detection algorithms for permanent magnet synchronous machines (PMSMs) with interturn short-circuit faults under closedloop operation. The first method analyzes the harmonic content of dq-axis voltage components, while the second focuses on the negative-sequence component of the three-phase voltage, as most fault-induced distortions appear in voltage signals during closedloop control. Both algorithms are implemented in MATLAB Simulink and validated through simulations and experiments on a 1 kW PMSM test bench, using both an ideal voltage source with no high-order switching harmonics and a voltage-source inverter with high-order switching harmonics. Results show that both methods effectively detect interturn short-circuit faults in both supply scenarios, with the negative-sequence approach being more robust to waveform distortion. These findings support the use of signal-based techniques for real-time fault monitoring in industrial PMSM applications. The main contribution of this paper is a rigorous evaluation of how high-frequency switching harmonics affect the fault detection algorithms, along with an experimental comparison on the same PMSM.