A Honeycomb-Like Predictive Controller With a Reduced Computational Burden for Three-Level NPC Converters

This article presents a predictive control strategy to control a photovoltaic plant based on bifacial photovoltaic (BPV) panels connected to the electrical grid through a three-phase neutral point clamped (NPC) power converter. Electricity generation plants based on nonconventional renewable energies are affected by the availability of the natural resource and grid changes. To achieve optimal use of natural resources and equipment, BPV cells must operate at the maximum power point and the currents injected into a grid must be in phase with the voltages. Due to the high number of states, one important drawback of conventional predictive control in the NPC converter is its computational burden. The proposed controller avoids the use of a cost function and can achieve the same control objectives with only 20% of the computational cost of the conventional strategy. To achieve this reduction the proposed strategy exploits the operating area of the converter, separating the possible voltages in hexagons, which results in a shape like a honeycomb. Simulation and experimental results show the feasibility of the proposed method in different operating conditions.