Modulated Model Predictive Control for Modular Multilevel Matrix Converter with Enhanced Steady-state Performance

Modular multilevel matrix converter (M3C) is a direct ac/ac frequency conversion technology, which exhibits the characteristics of modularity, scalability, and redundancy. Conventional finite control set model predictive control (FCS-MPC) applies only a single switching state per control period, resulting in a variable switching frequency and significant current tracking errors. In this paper, a modulated model predictive control (M2PC) strategy is proposed, which fixes the switching frequency through modulation technology and reduces the tracking errors in branch currents by operating double switching states. Thus, the steady-state performance of the system is enhanced with the superior dynamic-state performance of the conventional FCS-MPC reserved. In addition, aiming at the inherent computational delay of the digital controller, a two-step forward prediction delay compensation scheme for proposed M2PC is employed. Finally, the feasibility and effectiveness of the proposed method are verified by simulations and experimental results.