A sensorless direct torque control scheme for a permanent magnet synchronous motor based on unscented Kalman filter

A new sensorless direct torque control (DTC) and stator flux linkage estimator for an interior permanent magnet synchronous motor (PMSM) drive system is investigated. An accurate unscented Kalman filter(UKF) observer is designed for estimating the stator flux linkage, rotor position and angular speed based on the nonlinear model of the PMSM and a UKF. The observer provides recursive optimal state estimation of the PMSM which is a nonlinear, time variant and dynamical system in real time by using input signals that are corrupted by noise. The UKF is robust to parameter variation, model inaccuracy, process noise and measurement noise. Space vector modulation(SVM) is used for improving the system performance. The performance of the proposed sensorless control scheme is verified by the simulation results. The UKF-based PMSM-DTC sensorless drive system has good speed and torque response performance over a wide speed range, even for large initial position errors and initial flux linkage error.