I'm writing my own CFD code. So far it is only 2D. If I apply the LES Smagorinsky model for trubulent channel flow, should I expect good results in 2D? Will the average velocity profile look anything line the DNS?

Thanks!

I think not. You can only perform a good LES simulation in 3D because the turbulent structures (vortices) break down in the third dimension. You can try, but I think that you'll obtain very poor results from a 2D simulation...

You will most certainly not get the same results. 2D turbulence is dominated by reverse energy cascades (small scale to large scale).

What you could do is perform your 2D simulation in 3D with periodic BC in the third direction. That way you might be able to compare your results with those in literature.

Hi,

for the beginning Turbulence is three dimensional. Anyway, in practice are also 2D turbulence simulations. What you cannot do is to compare 2D LES data with 3D DNS data...

regards, tele