Model Predictive Control for Six-Phase Induction Machines with Insight into Past Current Errors

Multi-phase electric drives can act as a competitive solution for high-power applications. To take advantage of the multi-phase benefits, the design of high-performance control techniques is a crucial task. In this regard, the inherent features of model predictive control (MPC) have aroused the interest of the scientific community in the last decade. This control solution, based on the usage of a cost function, provides notable flexibility to the control designer to select diverse regulation goals. Although the use of some weighting factors is a common trend to prioritize control purposes, the internal behavior of each term of the cost function is commonly ignored since the cost function is only focused on minimizing the total error. With this issue in mind, this work firstly performs a detailed analysis about the behavior of each cost function term in a standard MPC. The mentioned study permits identifying a certain pattern of the internal mechanics of the cost function that can disturb the current quality. To solve this limitation, the cost function of the MPC scheme is redefined. The proposed strategy permits improving current quality without increasing the number of switching states applied per control period, as it is validated with experimental results in six-phase electric drives.