A new model predictive torque control strategy with reduced set of prediction vectors

Major drawback of finite control set model predictive control (FCS-MPC) is its high computational burden. This paper proposes a new optimal vector selection strategy that reduces the computational cost of FCS-MPC technique. Considering two-level voltage source inverters (2L-VSI) utilized as motor drives, proposed strategy reduces the number of active prediction vectors from six to three. Hence, cost function is evaluated only for four vectors (three active and one zero). Moreover, between the two possible zero vector configurations, the one which avoids switching of the maximum current carrying phase arm is selected. Proposed control strategy has been validated by detailed MATLAB/Simulink models. Required computation time for the control algorithm has been reduced by 30%. The dynamic performance of the drive is not degraded with the reduction of active prediction vectors. Compared to the classical FCS-MPC, proposed algorithm offers up to 28% switching loss reduction (9.9% in average) especially in the high torque — low speed region. Simulation models have been made available as open access.