PMSM Current Sensor FDI Based on DC Link Current Estimation

Abstract

Sensors are significant in electric machine applications, such as electric vehicles, to maintain the safe operation of a system. The sensor reliability must be ensured to help avoid unexpected system failure caused by faulty sensor signals. To mitigate the machine current sensor fault impact on permanent magnet synchronous machines (PMSMs), this paper proposes a novel PMSM current sensor fault detection and isolation (FDI) method based on DC link current estimation. The fault detection is achieved by comparing the residual between measured and estimated DC link currents with a threshold value; the fault isolation is achieved based on the phase signals estimation and residuals examination. The proposed method is easy to implement without complicated modeling, not influenced by system imbalance and capable of distinguishing between machine current sensor and non-sensor faults. As compared to the existing PMSM systems without the proposed FDI method, the system failure risk under a machine current sensor fault is reduced when the proposed method is integrated into the PMSM controller. The effectiveness of the proposed FDI method is validated by simulation results in MATLAB.