Improved Performance for PMSM Control System Based on Feedback Linearization and Computational Intelligence

Abstract

Starting from the nonlinear structure of the description equations of a Permanent Magnet Synchronous Motor (PMSM), this article presents the development of a control algorithm based on the feedback linearization (FL) method using Lie derivatives. The improvement of the performance of the PMSM control system when using a FL Controller is achieved by using computational intelligence algorithms. To optimize the coefficients of the algorithm implemented in FL Controller, a particle swarm optimization (PSO) and respectively a genetic algorithm (GA) are used. The improvement of the performance of the PMSM control system by using a Reinforcement Learning - Twin Delayed Deep Deterministic Policy Gradient (RL-TD3) agent in combination with FL Controller is also presented. The comparative numerical simulations of these PMSM control structures demonstrate the improvement of the PMSM control system performance by using computational intelligence algorithms.