A Simple Model-Free Solution for Finite Control-Set Predictive Control in Power Converters

The goal of this article is to provide a simple modelfree solution to the loss problem of accuracy in system model inherent in the existing finite control-set model predictive control for power converter systems with uncertain dynamics. To be more precise, the core idea behind the solution is to construct an invariance manifold-based unknown system dynamics estimator in a simple way by introducing a first-order low-pass filter into the finite control-set model-free predictive control framework. In the developed framework, the estimator is updated in real time to approximate the unknown nonlinear system dynamics, leading to considerable enhancement of robustness. Dissimilar to the state-of-the-art model-free predictive control methods, our development advocates a simple and straightforward structure since it does not require the widely applied function approximators and weighting factors, while ensuring satisfactory system performances than the existing ones. Further, the convergence analysis of the developed estimator is manifested. Finally, we underline its merits with different benchmark examples from the literature.