Dual-Vector Model Predictive Control for Modular Multilevel Converter With Low Calculation Burden

This paper proposes a novel dual-vector model predictive control (DV-MPC) method applied to modular multilevel converter(MMC). The traditional predictive control method generally aims to minimize the output current tracking error at the next sampling instant, while the proposed method minimizes the total harmonic distortion (THD) of the output current by quantifying the relationship between the THD and the output current trajectory within the next sampling period. The control of the circulating current is achieved by selecting proper total number of inserted submodules (SMs) in the upper and lower arms. By selecting appropriate SMs to insert or bypass, the voltage balance between submodules within the arm is fulfilled. This method accomplishes multiple control objectives without the need for cumbersome weighting factor design. At the same time, the adopted preselection method and total number of SMs selection method effectively reduce the calculation burden. Simulation and experimental results verify the superiority of the proposed method.