Model Predictive Control of SRMs Based on Modified Multilevel Power Converter

This article presents a novel model predictive control (MPC) method for the switched reluctance machines (SRMs) based on a modified multilevel power converter to suppress the torque ripple. First, a modified multilevel power converter is proposed and the working principle is introduced. Then, a fast modeling method is presented to provide a data foundation for the MPC of the SRMs. The multilevel power converter brings more voltage vector options but also increases the computational effort of the MPC method. To reduce the calculation burden, a voltage vector selection rule is proposed and the number of states to be predicted in the commutation region is reduced from 25 to 12. Furthermore, the MPC method is presented in detail by constructing a loss function over the torque ripple and phase current. Finally, the experimental results under steady and dynamic conditions prove the effectiveness of the proposed MPC method.