Cost function optimization for multi-phase induction machines predictive control

Abstract

Control of multi-phase induction machines is evolving with the incorporation of predictive controllers that operate directly with the power converter without modulators. It is realized in discrete-time through an exhaustive optimization of the inverter states. The optimization uses a cost function that penalizes deviations from the reference and other undesired factors. The cost function design has a big impact on closed loop performance. In this paper, a new strategy for cost function design is presented as a contribution. The strategy defines and a performance index relative to the observed behavior. The cost functions are derived from an optimization of said index. Experimental and simulation results are included for a five-phase induction machine with a comparison with previous methods in the context of predictive current control.