Cost function optimization for predictive control of a five‐phase IM drive

Abstract

SummaryModel predictive control has been used, for some time now, as a method to directly control power converters in electrical systems. The usual practice is tuning the cost function of the controller to obtain a certain compromise solution over the whole operating range of the system. This method is extended here to consider multiple, locally optimal, and tunings. The design objectives (tracking error, switching frequency, etc) are used to define a unique performance index that is locally optimized. In this way, the parameters of the cost function are linked to the current operating point. The tuning at each operating point is obtained numerically solving the optimization of the performance index. Although the idea can be applied to induction machines with any number of phases, in this paper, a five‐phase induction motor is considered for presentation. This system is a demanding case due to the extra number of phases compared with the traditional three‐phase motor. Simulation and experimental results are presented to assess the proposed predictive controller.