Model Predictive Control of an Inverter in UPS System

This paper presents a model predictive control strategy of an inverter for uninterruptible power supply (UPS) system. The proposed method uses a predictive module to precalculate the behavior of the load voltage in each sampling period. Then, a cost function is developed to select the optimal switching state of the inverter by minimizing the error between the referenced voltage and the predicted voltage in next sampling instant. The reference voltage and predicted voltage are both transformed to d-q plane. In addition, a LC filter is cascaded behind the inverter to weaken the high frequency harmonics of the output voltages and currents. Simulation results under balanced resistive load, balanced resistive-inductive load, balanced load step and unbalanced load conditions are presented, verifying the satisfactory performance of the proposed control scheme.