Low-Complexity Multi-Vector Model Predictive Current Control for PMSM Drives

Abstract

Multi-vector model predictive current control (MVMPCC) effectively reduces torque and current fluctuations compared to conventional model predictive control (MPC) schemes. Despite these benefits, its practical deployment is constrained by the substantial computational complexity arising from the need to evaluate all possible voltage vector (VV) combinations within each control cycle. To address this limitation, this paper proposes a low-complexity MV-MPCC that significantly reduces the computational burden while maintaining superior predictive control performance. With regard to the selection strategy of VVs in each sampling period, rather than sequentially traversing all active VVs to determine the optimal VVs, the proposed method identifies them by evaluating only three candidate VVs. This drastically reduces the number of cost function evaluations without compromising the quality of control. Furthermore, based on the principle of deadbeat current control, the duty cycles of the selected VVs are optimally allocated within each sampling interval. Theoretical analysis and results demonstrate the feasibility of the proposed control strategy.