Data-Driven Based Model-Free MPC for DTP-PMSM with Optimized Space Vector

Finite control set model predictive control (FCS-MPC) has been extensively investigated owing to its numerous disadvantage, such as fast response and easy implementation. However, FCS-MPC is highly sensitive to system modeling and parameters, and its performance deteriorate when the system has unmodeled dynamics and parameter mismatches. To address this issue, a data-driven based model-free predictive current control (MFPCC) for dual-three phase permanent magnet synchronous machine (DTP-PMSM) is proposed, which use only the input and output data of the system instead of the accurate mathematical models and parameters. In addition, the virtual voltage vectors (V3s) is utilized to reduce the harmonic current. More specifically, the space vector is optimized for reducing the torque ripple. Simulation investigations on DTP-PMSM platform validate the superiority of the proposed strategy.