A Virtual Voltage Vector Modulated Model Predictive Control of Modular Multilevel Converters

The model predictive control (MPC) method is widely applied in modular multilevel converter (MMC) control due to its multiple-object inclusion and fast response. But there still some challenges that exist such as high computational burden caused by the large number of submodule (SM), weighting factors tuning for various control object in the cost function and requirement of short sampling times. To solve these challenges, a virtual voltage vector modulated model predictive control (VSVV-M2PC) scheme for the control of the MMC is proposed. This scheme enables the MMC to output dual voltage levels, which improves the current tracking accuracy. Additional circulating current suppression (CCS) method and sorting-based modulator are designed, which reduce the computational burden, achieve fixed switching frequency, and avoid the weighting factor tuning process. Finally, the proposed scheme is validated by experimental results.