Simulation of Swept-Wing Receptivity to Distributed Roughness

Abstract

Simulations were carried out to model the receptivity and growth of stationary crossflow vortices from Distributed Roughness Elements (DRE) on a swept wing. A highly resolved Large Eddy Simulation (LES) numerical method was used for the study, the aim of the results were to achieve validation of the code to relevant experimental data and to gain a better understanding of the flow behaviour. The base flow for the simulations were based upon the experiment run by Hunt and Saric.1 The LES replicated the experimental setup using a WALE sub grid model and a streamline extraction process to only simulate the upper surface and reduce the overall computational expense. The WALE model is more suitable to modelling of transitional flows as it allows the sub-grid scale viscosity to vanish in laminar regions and in the inner regions of the boundary layer. Simulations were carried out for two spanwise wavelengths (λ = 6mm, 12mm) and for roughness heights ranging from 12 µm to 42 µm. The critical wavelength results showed, when comparing the stationary crossflow mode shapes, that the simulations at the larger roughness element sizes compare well with the experimental data. The control wavelength equally showed a good agreement to the experimental data at the larger roughness element sizes however the simulations over predict the amplitude of the smallest roughness element size. This can be attributed to either the simulation requiring a further refinement around the cylinder for the smallest roughness element size or to differences in the experimental and simulation roughness element shape. Overall the simulations successfully predict the receptivity from arrays of distributed roughness elements.