Variable-Vector-Based Model Predictive Control With Reduced Current Harmonic and Controllable Switching Frequency for PMSM Drives

This article proposes a variable-vector-based model predictive control (MPC) method for permanent-magnet synchronous motor drives. Different from finite-control-set MPC (FCS-MPC) assessing seven basic voltage vectors, the proposed method offers optimized candidate solutions considering one, two, three, and four vectors during one control period to obtain superior steady-state performance and controllable switching frequency. First, on the principle of current deadbeat control, the reference voltage vector is calculated based on an ultralocal model, which avoids the use of motor parameters. Second, the optimized candidate solutions are efficiently determined by adjusting the voltage vector combination and duration that are obtained according to the volt-second balance principle. Finally, the optimal solution is chosen by evaluating the designed cost functions, which include the current error with switching actions and even the harmonic current rms value. The experimental results confirm the effectiveness of the proposed method, which exhibits better steady-state performance, reducing the current total harmonic distortion by 37% and 12% at rated speed and rated load when compared with space-vector-modulation-based MPC and FCS-MPC at the same switching frequency.