Arm-current sensorless model predictive control for grid-interfacing modular multilevel converter with reduced switching frequency

Model predictive control (MPC) emerges as an attractive alternative for the control of modular multilevel converters (MMCs) owing to its superior dynamic performance and ease of implementation. Nevertheless, conventional MPC's performance could be further improved by reducing dependence on weighting factors, switching frequency, and computational burden. To achieve these objectives, an arm-current sensorless MPC is developed for a grid-interfacing MMC in this article. The MPC is integrated with an arm-current sensorless reduced switching frequency voltage balance algorithm that requires only m+1$$ m+1 $$ switching vectors (where m=$$ m= $$ number of submodules per arm) to optimize the cost function. This method does not require knowledge of arm-current direction to decide the charging and discharging state of the submodule capacitors. The developed control not only reduces the switching frequency of devices but also eliminates the need to sense the arm currents. As a result, it reduces the system's cost, complexity, computational burden, switching losses and provides a more reliable control method for MMCs with superior dynamic performance. Further, the effect of parameter mismatch on the developed MPC's dynamic performance and stability is studied. The developed MPC's effectiveness is validated through both simulation and experimental results and also compared with existing methods under different conditions. In this article, an arm current sensorless reduced switching frequency model predictive control is developed for grid-interfacing MMC, which operates submodule switches' of the MMC at the reduced switching frequency and does not require knowledge of the direction of the arm current. Further, model predictive control helps in achieving faster dynamic response.image