Design perspectives of a system identification approach of the NATO AVT-316 multi-swept wing aerodynamics

Abstract

Air supremacy requires enhanced maneuvering capabilities at high angles of attack and even beyond the stall angle. High angle-of-attack (high-α) maneuverability can be achieved using swept wings and tailored leading-edge shapes to replace local and disorganized flow separation with controlled separation-induced leading-edge vortices (LEV). Strong leading-edge vortices delay the stall to higher angles and generate a non-linear lift increment up to the angle of attack where the jet-type flow structure of LEV changes to a reversed-flow bubble (vortex breakdown phenomenon). A multi-swept wing configuration has the potential to delay the vortex breakdown to even higher angles as the vortices formed over the front wing can energize vortices formed over the main wing and lead to vortex merging. This article is focused on understanding the unsteady interactions between multiple leading edge vortices formed over multi-swept wing configurations in the subsonic speed regime. Specifically, this article investigates the aerodynamic characteristics of wings being evaluated under the initiative of the NATO STO AVT-316 Task Group. Highly refined meshes, and the use of hybrid turbulence models such as Detached Eddy Simulations (DES) and Delayed DES are required to accurately resolve the vortical flows over these wings. Simulating all flow conditions of interest is a computationally expensive approach. A system identification method was therefore proposed to rapidly and accurately generate aerodynamic models of these wings. The method uses a novel piece-wise chirp (constant amplitude and increasing frequency) motion as an input signal (training maneuver). The motion computational cost is equivalent to the cost of six static CFD simulations, however it can predict aerodynamic responses of the wings over a wide range of angles of attack. The method has been tested for double and triple delta wings. Some design considerations are provided based on predicted flow features and aerodynamic data. Prediction results with different turbulence models, sting geometries, grid resolutions and an adaptive mesh refinement approach are provided.