Novel ripple reduced Direct Model Predictive Control of three-level NPC active front end with reduced computational effort

Abstract

Three-level neutral-point clamped (NPC) power converter seems promising for high power grid-tied renewable applications. Direct Model Predictive Control (DMPC) is an attractive control method, in particular for multi-level converters. However, relatively big ripples of the control variables and heavy computational efforts are regarded as two of the shortcomings for DMPC schemes due to its cost enumeration and one-vector-per-control-interval characters. This work proposes a computational efficient ripple-reduced DMPC scheme for three level NPC Active-Front-End (AFE). By combining a deadbeat concept the targeted switching vectors are allocated efficiently and the respective actuating times of the vectors are on-line optimally calculated. Compared to the classical DMPC scheme, computational efforts are reduced efficiently and much smaller ripples of the control variables are achieved. Simulation results emphasize the effectiveness of the proposed scheme.