Detection of Demagnetization Fault in Interior Permanent Magnet Synchronous Motors for EV Based on UKF

Abstract

In most applications, the permanent magnet flux linkage of interior permanent magnet synchronous motors is treated as a known constant value. However, high operation temperature, stator winding starting current impulsion, armature reaction and frequent field weakening control all may cause irreversible demagnetization of permanent magnet which has direct impacts on motor control performance. In order to implement demagnetization fault detection accurately, the nonlinear observation model is first set up allowing for the effects of magnetic saturation and cross saturation of IPMSM, and then the unscented Kalman filter (UKF) , based on dynamic data processing technology, is presented to implement flux linkage estimation and overcome the disadvantages of traditional extended Kalman filter (EKF). Finally, the simulation results are presented to verify the effectiveness of the proposed method, it is pointed that this method can estimate the permanent magnet flux linkage with the estimation error less than 3% in the full speed range of IPMSM, so it is fully competent for demagnetization fault detection, and consequently, can also prevent the deterioration of demagnetization and improve the reliability of drive system.