Optimization of model predictive torque controlled induction motor drive system

In this thesis, model predictive torque control, which is a new method for high-performance control of induction motors, is optimized with conventional (single objective) and multi-objective genetic algorithms (GAs) and their control performances are compared through simulation studies. Despite the advantages of the model predictive torque control, it has problems with the selection of the weighting factor in the cost function. Although the trial-and-error method is generally used in the literature, few studies focus on determining these factors with multi-objective metaheuristic optimization algorithms. Unlike the current literature, this thesis focuses on optimizing model predictive torque control with conventional GA using speed errors, examining the effect of different decision-making methods on multi-objective optimization of model predictive torque control, and finally comparing the control performances of conventional and multi-objective optimized model predictive torque control systems. The simulation results show that the conventional optimization method is a powerful candidate against the multi-objective optimization method.