This work numerically models an airfoil geometry inspired by the downy coat of the barn owl and contrasts its aerodynamic and aeroacoustic performance with the baseline airfoil. The owl-inspired geometry simulated here was suggested by Clark et al. (2014) which used ``fences\u27\u27 near the trailing edge of the airfoil to simulate the canopy effect of the owl feathers. The simulated geometry is not an exact replica of the experimental model and the aerodynamics and aeroacoustics consequences of these differences are discussed. Implicit large eddy simulations are performed with low-pass filtering of the solution using the extensively validated, high-order accurate Navier Stokes solver FDL3DI. A baseline NACA 0012 airfoil is compared against the same model with an array of fences at the trailing edge. Both models are simulated with three different flow conditions.The owl-inspired airfoil geometries do not significantly degrade the aerodynamic performance but lead to reductions in unsteady surface pressure by up to 4.0 dB compared to the baseline. Unsteady surface pressure is the primary aerodynamic noise source in this problem where the flow Mach number (0.2) is small. The fences give a reduction in the overall sound pressure level (OASPL) of unsteady surface pressure in the entire region where the fences are located, except around the fence leading edge where the OASPL is found to increase. It is hypothesized that this observed increase in the simulations is due to the differences in the modeling of the leading edge of the fence between the experiments and the simulations.Results of noise source diagnostics are reported. Contours and profiles of turbulence kinetic energy obtained from the simulations clearly show that the turbulence is lifted off the surface and relocated above the fences. The observed surface OASPL reductions are believed to be due to the increased separation between the sound sources (turbulence) from the scattering surface, primarily the trailing edge of the airfoil. The sharp leading edge of the fence is a singularity which is most effective at scattering hydrodynamic energy in the turbulence into acoustics. There does not appear to be a significant difference in either the spanwise coherence or the far-field predicted noise between the baseline and fence geometries. The lack of far field noise reduction and spanwise coherence reduction in the simulations suggests a direction for future work -- removing the geometric differences in the experiments and simulations and repeating the analysis.
展开▼
