Restrained search predictive control for five-phase dual-inverter supplied loads

Model-based predictive control (MBPC) techniques have been recently applied with success to electrical drives. Nevertheless, the real-time implementation of MBPC in multilevel and/or multiphase drives becomes difficult as the number of levels and phases increase because multiple iterations are required for optimal performance. At the same time, MBPC performance for low switching frequencies quickly deteriorates due to the absence of modulation within a sampling period. These shortcomings are especially relevant in multiphase dual-inverter supplied drives because the number of switching combinations is highly increased. This work examines the use of restrained search predictive control (RSPC) techniques for five-phase dual-inverter supplied loads proving by simulations that it is possible to simultaneously achieve good performance, low computational cost and low switching losses.