Double-Virtual-Vector-Based Model Predictive Torque Control for Dual Three-Phase PMSM

To overcome the drawbacks of large torque ripples and high harmonic contents in a dual three-phase permanent magnet synchronous motor (PMSM) used in electric vehicle drive systems, a double-virtual-vector-based model predictive torque control (DVV-MPTC) strategy was proposed in this paper. Firstly, 12 virtual voltage vectors were constructed to minimize harmonic interference as much as possible. Then, the DVV-MPTC strategy is proposed to solve the problem of large torque ripples caused by single-virtual-vector-based MPTC (SVV-MPTC) method. On the other hand, an enhancement to the cost function was also introduced to resolve the challenges of tuning weight coefficients. Experimental comparisons between traditional direct torque control (DTC), SVV-MPTC method, and the proposed DVV-MPTC strategy were carried out, which show that the latter achieves significant improvements. In particular, it can reduce both harmonic components and torque ripple compared to traditional control strategies, resulting in a more efficient and stable performance for the electric drive system.