Model Predictive Control for LCL-Filtered DG-Grid Interfacing Inverters With State Variable and Input Disturbance Estimation

Abstract

Due to the high-order property and multivariable coupling characteristics of LCL filtered voltage source inverter (VSI), model predictive control (MPC) have received great favor with the advantage such as the intuitive concept, improved bandwidth, and multiobjective optimization. To suppress the inherent resonance problem, reject input disturbance and meet the multifunction requirement in distributed generation (DG) units, a model-based estimator is proposed in this paper, which exploits the previous four consecutive output information to simultaneously calculate the value of state variable and input disturbance in real-time. Specifically, with only grid voltage and current measurement, the inverter-side current, capacitor voltage, and input voltage disturbance can be obtained, which are used to generate the optimized control law. Constant switching frequency is achieved thanks to the modulator. Therefore, the proposed control law can improve the injected power quality, and also the system hardware cost can be saved. Simulation and experiment are realized to verify the proposed method.