A Predictive Submodule Choosing Algorithm for Soft-Switching Modular Multilevel Converters with Nearest Level Modulation Scheme

This paper applies active resonant commutated pole (ARCP) submodule (SM) in modular multilevel converter (MMC) to reduce the switching losses. In order to create zero-voltage-switching conditions for the main switches, the auxiliary switches need to be conducted before the state of the submodule changes. When MMC is with nearest level modulation (NLM) and SM is inserted by sorting balance algorithm, the advance conduction of auxiliary switches would introduce a phase shift error between the output wave and the modulation reference. A predictive SM choosing algorithm is proposed to overcome this problem. The voltage level message in the last cycle is stored to predict the level change at present moment, and the influence of the auxiliary switch advance conduction to the output wave is eliminated. Experiment results prove the soft-switching characteristics of ARCP MMC and simulation results validate the effectiveness of the predictive algorithm.