A Reference-Variant-Based Model Predictive Torque Control Scheme for PMSM

Finite control set model predictive control (FCS-MPC) has been widely used in permanent magnet synchronous machine (PMSM) drive systems. FCS-MPC takes advantage of the discrete nature of power converters and selects an optimal switching state based on an evaluation of all possible ones. However, FCS-MPC suffers from an unsatisfactory steady-state performance due to the lack of input variants and suffers from a low computational efficiency due to the evaluation of all possible switching states. In this paper, a reference-variant based model predictive torque control (MPTC) is introduced. In the proposed strategy, by transforming the reference torque and reference flux to a reference voltage vector, an optimal solution can be obtained in a fast way. However, due to the single reference point and the lack of input variants, additional constraints cannot be included. Thus, reference variants are constructed based on the original reference point, thus resulting in more optimal solutions. A MPTC scheme with flexible structure to handle constraints is achieved. Simulations results show the improvements by comparison of the proposed MPTC and the single reference voltage vector based MPTC.