A robust model predictive control for grid-connected voltage-source inverters

Abstract

Model predictive control is a new alternative in the control of power converters. The practical application of this controller demands a correct compensation of the computation delay. This compensation requires an estimation of the load current, usually attained in an open-loop manner, i. e. without error feedback of the model parameters. However, this strategy gives rise to diverse perturbations in the waveform, when there is a certain amount of parameter variation. In this work, the effects of this variation are analyzed and a way to reduce the distortion is proposed, using a Luenberger observer. This new approach is called Robust Model Predictive Control. The proposal is applied to a grid-connected voltage-source inverter. To validate the proposal, simulations are included in the paper.