Using Extended Kalman Filter and Adaptive Filter for sensorles predictive torque control of PM-Assissted Synchronous Reluctance Motor

Abstract

The Permanent Magnet-Assissted Synchronous Reluctance Motor (PMA-SynRM) drive has become one of the most interesting replacements for the high efficiency variable speed drive. Herein, sensorless predictive torque control of a PMA-SynRM with non-sinusoidal back electromotive force (Back-EMF) is introduced. In order to control PMA-SynRM, finite control set-model predictive control (FCS-MPC) is implemented by means of a two-level inverter. Furthermore, an improved form of FCS-MPC, i.e., direct mean torque control (DMTC), is utilized as a second method to control PMA-SynRM. For improving the sensorless the combination of Extended Kalman Filter (EKF), Adaptive Filter (AF) and quadrature Phase-Locked Loop (PLL) are used for better estimation of the non-sinusoidal back EMF, elimination of the high order harmonics, and the accurate estimation of position and speed rotor, respectively. The simulation result in effectively minimizing torque ripples compared to conventional FCS-MPC. The outcomes of the observer simulation are successfully guaranteed the accurate estimation of speed and rotor position.