Switching State Selection for Model Predictive Control Based on Genetic Algorithm Solution in an Indirect Matrix Converter

Abstract

Model predictive control (MPC) has emerged as a promising control scheme for power converters and motor drives. One of the major advantages of MPC is the possibility to control several system objectives with a single control law (cost function). However, the minimization selection of the cost function and its corresponding switching state is often achieved by using the enumeration method based on a simple element comparison that can lead to a local optimum solution and is time-consuming. In this article, a genetic algorithm (GA) optimum technique is applied to enable a global optimal solution to the cost function and its corresponding switching state in parallel implementation of MPC that ensures good performance in terms of the input reactive power and output current in an indirect matrix converter (IMC). Moreover, a performance comparison between using a GA and a conventional enumeration method is described. Meanwhile, a hybrid weighting factor searching method combining branch and bound algorithms with a strategy of precise searching over a small weighting range is proposed. Finally, the operation principles and validity of the proposed algorithm are analyzed and verified by using simulation and experimental results.