Robust Reconstruction of Current Sensor Faults for PMSM Drives in the Presence of Disturbances

The reliability and availability of a permanent magnet synchronous motor (PMSM) drive system are of increasing importance in industrial applications. Occurrence of current sensor faults in the system can lead to system degradation. Therefore, early detection of the faults has become an important problem. However, various exogenous disturbances seriously affect the accuracy of the faults detection. This paper presents a method of reconstructing current sensor faults for a PMSM drive containing unknown disturbances. An equivalent-input-disturbance approach is devised to reject the effects of unknown disturbances on system performance and reconstruction of faults And an integral observer is designed to approximately separate faults from the measured output to eliminate the interference to disturbance rejection. Then, a sliding-mode observer is designed for the disturbance-rejected system with current sensor faults. And the faults are reconstructed based on sliding-mode equivalent control methodology. The presented scheme has good performance in terms of faults reconstruction and is robust to time-varying unknown disturbances. Simulations and experiments validate the effectiveness of the method.