Deadbeat predictive control of three-level NPC inverter using in the photovoltaic chain conversion system

This paper aims to improve the performance of the Deadbeat (DB) predictive controller for a three-level neutral point clamped (NPC) inverter used in a photovoltaic energy conversion system (PVECS). The chain comprises a Buck-Boost power converter used as maximum power point tracking (MPPT) based on perturb and observe algorithm P&O controller providing robust tracking property against environment changes variations. Then, a three-level NPC stage for a PV-grid-connected system is used. A Deadbeat (DB) predictive controller is used to transfer solar energy to the grid side. The switching states necessary for the prediction and objective function optimization are chosen based on the position of the reference voltage vector in the space vector plane, the inverter current directions, and the state of charge of the DC-link capacitors. Because of this, choosing the best switching state is quick and simple and avoids the need to choose a weighting factor for capacitor voltage balancing. The proposed control approach eliminates the weighting factor for DC-link capacitor voltage balancing using redundant vectors. The results show that the suggested approach successfully achieves the DC-link capacitor voltage balancing while maintaining dynamic and steady-state current tracking.