Model Predictive Control of a Modular Multilevel DC Transformer under Quasi-square Modulation

This paper proposes a model predictive control (MPC) strategy for the modular multilevel DC transformer (MMDCT) under quasi-square modulation. The approach uses finite state machine principles and MPC optimization to determine the switching state group of the primary-side phase legs during level change instances that concomitantly controls the circulating current, maintains the balance of submodule capacitor voltages, and minimizes the total number of switching transitions. The MPC optimization also dynamically defines the phase displacement between the primary- and secondary-side bridges to manage the power transferred across the galvanic barrier and control the output DC-link voltage according to a target reference. The correctness and effectiveness of the proposed control technique are verified through dynamic simulation and experimental results obtained from an exemplary 2N+1-level MMDCT.