Improvement Performances of Sensorless Control for PMSM Based on DTC Strategy Using SMO Observer and RL-TD3 Agent

One of the elements that contribute to increasing the reliability of the control system of a Permanent Magnet Synchronous Motor (PMSM) is the replacement of the speed transducers with software-implemented speed observers. Based on this, this paper concern on the sensorless control of a PMSM using the Direct Torque Control (DTC) control type strategy, in which a speed observer is used in combination with a Reinforcement Learning - Twin Delayed Deep Deterministic Policy Gradient (RL-TD3) type agent to increase the accuracy of the PMSM rotor speed estimation. The latter, after the training stage, can provide correction signals to the speed observer so that the estimated speed is as close as possible to the estimated speed. The control structures, control algorithms and operating equations of the PMSM, the DTC-type control strategy and speed observer are presented in this article. Numerical simulations realized in the Matlab/Simulink programming environment validate the superiority of the PMSM rotor speed estimation performance in case of using an RL-TD3 type agent in join with a speed observer, compared to the case of using the speed observer alone.