Model Predictive Control for Grid-Tied Multi-Port System With Integrated PV and Battery Storage

This paper proposes a unified model predictive control (MPC) scheme for the integrated photovoltaic (PV) and battery storage system, where both of them are directly connected to the utility grid with high conversion efficiency through a multi-level neutral-point-clamped (NPC) inverter based multi-port interface. In such a system, the individual/unequal input voltage from each DC port raises control challenges, resulting in asymmetric voltage vector distribution and increased modulation complexity in the AC side. In this case, the finite-control-set MPC (FCS-MPC) scheme is proposed to make the power management be liberated from the modulator design thanks to the natural advantage, i.e., the direct control property without using a modulator. The multivariable-based cost function is designed for the AC side to regulate the injected grid current well to meet the IEEE 519-2014 standard. On the other hand, to proceed with the normal DC-side power flow, the capacitor voltage of each port is modeled, predicted, and also regulated through the unified cost function in the MPC framework. As a result, each PV array can work at the individual maximum power point (MPP) and the battery can be automatically charged/discharged to compensate for the power difference. The five-level four-port inverter-based simulation and three-level dual-port inverter-based experiment are conducted to verify the multi-mode operation of the integrated system and the advantages of the proposed controller.