Robust predictive control strategy for grid-connected inverters with ultra-local model based on linear matrix inequality

Abstract

To address the issue of poor robustness in the model predictive control of grid-connected inverters due to disturbances in load model parameters, this article developed an ultra-local model robust predictive controller based on Linear Matrix Inequality (LMI). First, to avoid dependence on model parameters, an ultra-local model of the system was constructed. Second, in the process of setting the state variable compensation gain, to simplify the complex computational steps of the traditional Lyapunov algorithm, the task of finding the gain that satisfies the Lyapunov stability condition was ingeniously transformed into an optimization problem of solving constrained LMI. Third, the optimized state variables are utilized to design a new cost function predictive model, thereby enhancing the precision of system control. Finally, the effectiveness of the proposed approach has been validated through simulations and experiments.