The online detection of incipient faults for permanent magnet synchronous machine (PMSM) has been a great challenge in recent decades, where the fault features are extremely trivial and the machine fault type is normally unknown in advance. An online fault detection and discrimination of incipient faults for PMSM, including stator interturn short-circuit and resistive unbalance fault, is proposed in this article. First, zero-sequence voltage signal is collected and the zero-sequence voltage component (ZSVC) is used to detect the abnormal status of the PMSM. Then, high frequency (HF) signals are injected to the PMSM to discriminate the fault type. It shows that the amplitude of the HF component in ZSVC is proportional to the frequency of the injected signal under interturn short-circuit while it is constant under the resistive unbalance fault, and due to this difference, the fault type may be discriminated. Finally, the effectiveness of the theoretical analysis is validated by the simulation and experimental results.
Read full abstract