A Simple and Efficient Direct Predictive Control Framework for Multiparallel Power Converters

Parallel connection of power converters is a widely adopted approach to increase power capacity and enhance fault tolerance. This article presents a simple and efficient direct predictive control framework for multiparallel two-level and three-level neutral-point-clamped converters. By using the number of available switching states as control variables, rather than the switching states themselves, the candidate set is reduced to no more than four, significantly lowering the computational burden and enhancing scalability. Additionally, a novel current sorting method is introduced to suppress circulating currents independently, allowing more flexibility for output current and voltage balancing optimization. Both simulation and experimental results validate its effectiveness.