Learning-based Adaptive Gust Mitigation with Oscillating Wings

Abstract

This paper investigates the application of a learning-based adaptive controller to mitigate the effect of gust on the lift generated by an airfoil in an unsteady flow environment. A high-order accurate CFD model is used to model the unsteady flow over a pitching airfoil. Open-loop simulations of the CFD model are used to ascertain feasible lift commands. The learning-based adaptive controller is based on the retrospective cost adaptive control (RCAC). First, RCAC is used to regulate the lift coefficient of the airfoil in a nominal case without gust. Next, the effect of the hyperparameters of the adaptive control on the closed-loop performance is investigated. Finally, we used RCAC to regulate the lift coefficient and mitigated the effect of gust on the airfoil.