LES with near wall resolution
 

hi,

I am trying to do a LES (unstructured tetrahedral grid) on wall bounded flows (Re=1000) with const coefficient (Cs) Smagorinsky model for SGS. I have 2 doubts:

1. Will it be possible at all to capture the skin friction and other wall parameters with this kind of model, if my grid is fine enough??

2. What is the grid cell size I require near wall? I think it should be obtained by y+ < 1. Also is it necessary to have the same grid size in all directions, I mean should deltax, deltay and deltaz be equal?

Thanks,

Regards, Shailesh

Re: LES with near wall resolution
 

Hi

The answer of that, can you found in the lecture notes by Lars Davisson from Chalmers, Sweden at www.tfd.chalmers.se and find the course on turbulence modelling and his notes on the Web. It's funny that you've answered your own second question. For problems near wall, try to use a Hybrid modell, e.g. (V)LES Willems(from 1996) two scale model: filtered epsilon = RANS epsilon due to inertial range while filtered k = RANS k * (1-f(r)) where f(r) = 1 - (filter width/integral scale)^(2/3) when filter width/integral scale <= 1

0 otherwise

That's, a short version of Spetiale model. OK?

/Raul Machado