High Performance PMSM Control System Based on Extended Kalman Filter

Zedong Zheng,Yongdong Li,Fadel Maurice,Xi Xiao

DOI: https://doi.org/10.19595/j.cnki.1000-6753.tces.2007.10.004

2007-01-01

Abstract:Extended Kalman filter (EKF) is a nonlinear estimation method based on optimal control theory. The effects of system noises and measurement noises are considered. In this paper, a new state observer based on extended Kalman filter is proposed for the permanent magnet synchronous motors (PMSM). It considers the quantization errors of the optical encoder with limited accuracy as measurement noise, and could observe the precise rotor position and speed. The motor can start up without need of the initial position. The observer can also observe the load torque which can be used as the feed-forward compensation of the reference torque, so the speed control performance can be greatly improved during load torque changes. The stator currents of PMSM are very small with light loads, so the effects of the noises and disturbances in the current measurement are very obvious to the system performance. Kalman filter can remove the disturbances and noises from the measured motor currents effectively, and improve the performance of the system with light loads. Simulation and experimental results are given to verify the performance of the proposed system.