Computationally Efficient Predictive Control of Photovoltaic Central NPC Inverter with Constant Switching Frequency

This paper proposes a computationally efficient modulated model predictive current control method for a three-phase neutral-point clamped (NPC) central inverter in the photovoltaic energy system. The proposed control method calculates the optimal sector number based on grid voltages, grid currents and maximum power point tracking and uses a subset of voltage vectors to reduce the number of calculations significantly, reducing the computational burden. The subset of voltage vectors are used to predict the future behavior of grid currents and DC-link capacitor voltages. A triangular-region based cost function chooses three optimal voltage vectors for a seven-segment switching method. The proposed control method provides low steady-state errors, fast transient response and a constant switching frequency. An 817 kW photovoltaic energy system is simulated to validate the proposed control method under transient and steady-state conditions.