Model Predictive Control of a Five-level Active Boost Neutral Point Clamped (5L-ABNPC) Inverter for Transformerless Grid-Connected PV Applications

Recently, an improved version of active boost neural point clamped (ABNPC) inverter with the capability to produce five output voltage levels using only six power switches has been presented in the literature. The presented modulation strategy of such inverter is based on an open-loop voltage-oriented control approach. The major advantage lies in its inherent voltage balancing of the integrated flying capacitor with the leakage current elimination capability for transformerless grid-tied photovoltaic applications. Using a fast and robust current controller-based technique, the injected grid current of the 5L-ABNPC can be tightly controlled under various dynamic conditions of the local grid. The aim of this paper is to apply two improved versions of the Finite Control Set (FCS) and Continues Control Set (CCS)-Model Predictive Control (MPC) strategies to control different operations of the discussed 5L-ABNPC inverter. The proposed solutions present an optimal switching state scheme with less computational burden for FCS-MPC and also a fixed switching frequency spectrum for the CCS-MPC. To verify the correctness of the inverter operation and its control platform, simulation and experimental results are also presented.