Improved vector selection based model predictive torque control for IPMSM

Abstract

The model predictive torque control (MPTC) is an effective strategy for high performance motor systems. The strategy obtains the optimal voltage vector more quickly and accurately compared with the traditional direct torque control. However, some problems of the strategy are needed to be solved, such as few active vectors, difficult cost function design, and hard duty cycle regulation and so on. An improved MPTC is put forward for these problems in this study. The number of the vector is increased by constructing virtual vectors in the improved method. The finite control set under the rotating coordinate system based on the stator flux orientation is established to select the vector and reduce the computation load. The evaluation mechanism of the selected vector is set by combined with the duty cycle method, so the weight factor of cost function in the traditional method is eliminated. And the duty cycle can play a full role in the adjustment. The prototype experiment system is built for verifying the proposed method. The results demonstrate that the proposed method has better torque and flux control performances and effectively improve the above problems in the traditional method.