Distributed Continuous Control Set Model Predictive Control Using Nash Equilibrium for Flying Capacitor Multilevel Converters

This article presents a novel distributed model predictive control (DMPC) method for multilevel converters with multiple switching configurations. The central aspect of the control approach is to organize the dynamic game among multiple subcontrollers and achieve the Nash equilibrium to realize the deadbeat control of the multilevel converter. Compared with the conventional lumped structure finite control set model predictive control (FCS-MPC) method, DMPC is more flexible in structure and easy to expand. Moreover, the output of DMPC is duty cycle for modulation resulting in fixed switching frequency and improved ripple. The DMPC method has been experimentally validated with a 100 kHz GaN-based five-level flying capacitor (FC) converter. Experimental results show that compared with FCS-MPC, the current ripple, the voltage ripple, and the total harmonic distortion (THD) under DMPC control are reduced by 80%, 70%, and 90%, respectively. At the same time, the computational burden of DMPC is only 16% of that of FCS-MPC. These advantages would promote the implementation of fixed frequency and deadbeat nonlinear control of the FCMC at high levels.