Abstract:
Surface mounted permanent magnet synchronous machine (SPMSM) has been widely applied in the servo drives and other industrial fields, due to its simpler structure, higher power/torque density, and better control performance. However, the change of temperature and flux saturation may seriously affect the control performance of motor. So it is significant to identify the parameters of motor online. Due to the low order state model of permanent magnet motor, the speed and parameters can not be simultaneously identified by using only one observer. Hence, this paper presents a method to cope with the above problem by combining EKF and MRAS to estimate magnet flux and rotor speed respectively. Meanwhile, the comparison with the accurate MRAS model is made and some discussions on the selection of adaptive PI parameters are also given. Simulation results illustrate the feasibility and effectiveness of the proposed methods.