Double-Objective Global Optimal Model-Free Predictive Control for SMPMSM Drive System Based on DSVM

Abstract

To achieve the global optimization of stator current and inverter switching frequency, a double-objective global optimal model-free predictive control (MFPC) is proposed based on discrete space vector modulation (DSVM) for surface-mounted permanent magnet synchronous motor (SMPMSM) drive system. First, the double-objective invalid optimization area of the conventional DSVM-based finite control set (FCS) model predictive control (MPC) is revealed, and the reason is analyzed that the global optimal voltage vector cannot be achieved. Then, according to the distribution of the double-objective invalid optimization area, the voltage hexagon is divided into three subareas. In each subarea, the candidate voltage vector with the best current control performance (BC-CVV) is obtained. Moreover, the inverter switching number of the three BC-CVVs is used as a criterion, and the number of voltage vectors evaluated online is significantly reduced. As a result, the double-objective global optimal voltage vector is achieved without enumerating all the feasible voltage vectors. Finally, experiments under different operating conditions are implemented and the effectiveness of the proposed method is confirmed.