An Improved Model Predictive Torque Control Based on Extended Control Set for PMSM

To address the issues of heavy computational load and large torque ripple in the conventional model predictive torque control (MPTC) forpermanent magnet synchronous motor (PMSM), this paper presents an improved MPTC approach based on an extended control set. Firstly, with the aim of lessening the computational overhead, a reduced-complexity search method based on the deadbeat principle is proposed to decrease the number of potential voltage vectors in the control set expanded by using discrete space vector modulation. The method put forward in this paper ascertains the optimal sector by leveraging the minimum cost function associated with the center voltage vector. It then divides the optimal sector into three regions based on the amplitude of the reference voltage vector, effectively narrowing the optimization range and potential voltage vectors, thereby simplifying the complexity of the suboptimal voltage vector selection process. Secondly, to reduce torque ripple, the suboptimal voltage vector obtained by the reduced-complexity search method is second stage optimization. After determining the region, it is combined with the virtual voltage vector set with variable amplitude according to its own region rules. The new vector set and the suboptimal voltage vector are optimized again to make the selected optimal voltage vector more closer to the ideal vector. Lastly, the validity of the proposed method is confirmed via experimental tests. The test results demonstrate that the method has lower computational complexity and better steady-state performance.