Data-driven inference of low order representations of observable dynamics for an airfoil model

Reduced order modeling of fluid flow systems can allow for the application of more sophisticated mathematical analysis and control techniques that would otherwise be computationally prohibitive. Existing reduction techniques often fail when nonlinear terms dominate the flow (e.g., at large Reynolds numbers) necessitating the development of new techniques. In this work, we implement an adaptive isostable reduction strategy to obtain a data-driven reduced order model that captures the dynamics of observables in a computational model for fluid flow over an airfoil at moderate Reynolds numbers. The resulting model characterizes the response to both time-varying inflow conditions and Dirichlet boundary conditions on the surface of the airfoil meant to represent suction or blowing through the action of surface jets. The resulting reduced order model behaviors agree well with the dynamics of the full order model simulations in response to both open loop and closed loop inputs. This study provides a proof of concept that reduced order modeling techniques using adaptive isostable coordinates can be successfully used in realistic fluid flow models using geometries with practical relevance.