[Panda School]

Hi,
i'm new to large-eddy simulation and trying to simulate the flow around a wing section using LES. I found no clear criteria for computational grid of LES.

is there any books or papers you can recommend about this topic?

Thanks in advance,
Z.F.

[siw]

ANSYS software presentation at http://www.tfd.chalmers.se/~lada/com...4_for_Lars.pdf. See Slide 11 onwards.

[Panda School]

Wow, that really helps, Thanks a lot!

[FMDenaro]

Originally Posted by Panda School

Hi,
i'm new to large-eddy simulation and trying to simulate the flow around a wing section using LES. I found no clear criteria for computational grid of LES.

is there any books or papers you can recommend about this topic?

Thanks in advance,
Z.F.

You cannot use LES around a 2D wing section but you need to extend the 3D spanwise direction. You can approximate a small part of the geometry using periodic conditions otherwise you need a full 3D geometry and the grid has to be very refined

[Panda School]

Thank FMDenaro for your reply.

I intend to extend the model in spanwise direction and apply periodic boundary conditions. But I don't know why LES cannot be used to simulate the flow around a 2D wing section. As I know, there are papers about 2D LES calculations.

[flotus1]

You can of course use LES in a 2D domain. But only if your intention is to simulate 2D turbulence. Its behaviour is fundamentally different from 3D turbulence. So if the physical domain is 3D, you can not run your LES in 2D and expect correct results.

[Panda School]

I think I got it, thank Alex for your explanation.

[FMDenaro]

Originally Posted by Panda School

Thank FMDenaro for your reply.

I intend to extend the model in spanwise direction and apply periodic boundary conditions. But I don't know why LES cannot be used to simulate the flow around a 2D wing section. As I know, there are papers about 2D LES calculations.

The issue is not in the lack in the LES formulation but in what is the aim of using LES (or also DNS). Using LES means you want to describe the physics of large vorticla structures of the turbulence as better as possible. In a 2D assumption, the vorticity dynamics is forced to have no stretching action, which is a fundamental mechanism in turbulence to create the energy cascade. Thus, you could also perform a 2D LES but the solution is not a physical description of the real turbulence.
Maybe you have read paper (I also have written some papers of 2D LES) in which LES is used for geophysical flows in a quasi-2D turbulence, that is a flow problem where two flow is developed more in the plane than in the third direction. That is not possible for a wing section.

[Panda School]

Now I know the difference between 2D and 3D LES, thank you again for your detailed explanation.
I'd like to take this opportunity to ask you whether LES is an appropriate method to analyse the wake behind a wing section at moderate Reynolds number, say, about 10e5. I know DNS could be a better choice but it is quite expensive computationally at such a Reynolds number.

[FMDenaro]

Originally Posted by Panda School

Now I know the difference between 2D and 3D LES, thank you again for your detailed explanation.
I'd like to take this opportunity to ask you whether LES is an appropriate method to analyse the wake behind a wing section at moderate Reynolds number, say, about 10e5. I know DNS could be a better choice but it is quite expensive computationally at such a Reynolds number.

A pure (wall resolved) LES around a 3d wing section will be computational expensive almost as much as a DNS ... Generally, this flow problem is solved using DES

[Panda School]

Thank you for your answer and patience.

[FMDenaro]

Originally Posted by Panda School

Thank you for your answer and patience.

If you are interested in a deeper understanding of the LES topics, you could have a read to my lectures about LES and the references addressed there



https://www.researchgate.net/publica..._Part_I_and_II