Predictive Control for Reduced Structure Multilevel Converters: Experimenting on a Seven Level Packed U-Cell

Abstract

Recently, a branch of multilevel converters is emerged, in which their ‘reduced structure’ topologies use lower number of devices compared to the available topologies. To get a cost efficient converter, lower number of components as well as high quality waveforms, multilevel converters with a ‘reduced structure’ (MCRS) are suitable for high/medium power systems. Also, utilizing the fast microprocessors available today, applications of predictive control in power converters are of very powerful and attractive alternatives to classical controllers. This paper proposes a finite control set model-based predictive control (FCS-MPC) for load current regulation and capacitor voltage balancing for a typical MCRS. A case study considered, three-phase seven level packed U-cell (PUC), which is among reduced structure multilevel converters. A discrete model of the system is derived, and a predictive model-based control is developed according to this model in order to predict the future behavior of the system for all possible switching states; then, the switching state that optimized the cost function is selected. The feasibility of the proposed FCS-MPC strategy for a seven level PUC is evaluated based on simulations with MATLAB/ SIMULINK. Moreover, experimental validation of the proposed control system on a 5 kVA PUC is examined through DSP implementation.