Overview of model predictive control for induction motor drives

Model predictive control (MPC) has attracted widespread attention in both academic and industry communities due to its merits of intuitive concept, quick dynamic response, multi-variable control, ability to handle various nonlinear constraints, and so on. It is considered a powerful alternative to field oriented control (FOC) and direct torque control (DTC) in high performance AC motor drives. Compared to FOC, MPC eliminates the use of internal current control loops and modulation block, hence featuring very quick dynamic response. Compared to DTC, MPC uses a cost function rather than a heuristic switching table to select the best voltage vector, producing better steady state performance. In spite of the merits above, MPC also presents some drawbacks such as high computational burden, nontrivial weighting factor tuning, high sampling frequency, variable switching frequency, model/parameter dependence and relatively high steady ripples in torque and stator flux. This paper presents the state of the art of MPC in high performance induction motor (IM) drives, and in particular the progress on solving the drawbacks of conventional MPC. Finally, one of the improved MPC is compared to FOC to validate its superiority. It is shown that the improved MPC has great potential in the future high performance AC motor drives.