Loading…
Large eddy simulation of crossflow vortices on an infinite swept wing
Large Eddy Simulation (LES) was used to model the formation of a crossflow vortex packet in a 3-dimensional swept-wing boundary layer. The capability of LES to model the fine structures near the wall was investigated. An experiment by Chernoray et al.1 was used as a base case and the solution domain...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Default Conference proceeding |
| Published: |
2012
|
| Subjects: | |
| Online Access: | https://hdl.handle.net/2134/10270 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Large Eddy Simulation (LES) was used to model the formation of a crossflow vortex packet in a 3-dimensional swept-wing boundary layer. The capability of LES to model the fine structures near the wall was investigated. An experiment by Chernoray et al.1 was used as a base case and the solution domain replicated their experimental setup with a C-16 airfoil at sweep 45◦ and Rec = 390, 000. Two sub-grid models were used for the investigation: the standard Smagorinsky and the Wall-Adapting Eddy Viscosity (WALE) model. The WALE model is more suitable as it allows the sub-grid scale viscosity to vanish in laminar regions and in the inner regions of the boundary layer. Stagnation streamlines at airfoil leading and trailing edges were taken from a full C-16 grid and used to define the lower boundary of a smaller solution domain which included the wing upper surface. This allowed, for a given computational resource, additional refinement in the area of interest. Results from the full grid matched well with that of the streamline defined domain. The laminar base flow for two LES grids of size 38 million & 161 million nodes was compared against the experiment and the results agreed well although the LES results slightly over-predicted the boundary layer thickness compared to the experiment. Stationary crossflow vortices were generated by strong continuous suction through a 1mm hole. The LES successfully captured the generation and growth of the crossflow vortex packet as well as the breakdown to turbulence on both grids. The fine grid performed better in modelling the growth of the vortices and the location of onset and growth of a dominant ‘z’ mode secondary instability. It was concluded that with suitable grid resolution LES is capable of successfully capturing the onset and development of crossflow vortices at a lower computational cost compared to DNS. |
|---|