On-line Particle Swarm Optimization of Anti-Windup Speed Controller for PMSM Drive System

Abstract

In this paper, we describe a new optimum design procedure for speed controller considering anti-windup for PMSM drive system. Tuning of speed controller is complicated by nonlinearities. Usual practice obtains controller settings with conventional linear analysis methods and then tunes the settings using trial-and-error methods during commissioning. An alternative approach, using particle swarm optimization algorithm devised to test and compare different parameters of speed controller, we search for the anti-windup speed controller to achieving the optimal compromise of variable speed at full load. The search is performed on-line, on the physical hardware, by continuously downloading and testing new solutions on a DSP running the control algorithms in real time. The controller obtained through PSO significantly outperforms alterative schemes obtained with conventional design techniques.