Hello there,

has anybody of you made experience about a LES-simulation with the FEM method? obviously the advantages of this discretization method (flexibility in representation of the flow domain) implies problems when concerning the computational effort. anyway, i would be happy to hear some suggestions and probabbly some recommended techniques to reduce the expenses. thanks very much.

Hi ..

Look for the work by keneth jansen (center for turbulence research , stanford ... annual research briefs 1994,1995 and 1997) Another group is Chalot and Ducros in France ( AIAA 94-0814 by CHalot et al)

hope that is of some use

mayank

Hi,

FEM is known to work very well in elliptic problems where diffusion is much bigger than convection. For turbulent flows the convective terms are very important and classical Galerkin FEM are not suited to treat them. You must then use upwinding methods. Kenneth Jansen from the Rensselaer faculty has worked a lot on this topic and you can find his articles on his web page.

I am also working on the use of finite element (with and without upwinding) in LES but I am still in a development phase and I don't have enough results to tell you if FEM seems to be a viable way or not in LES. Therefore I am interested in any information you 'll find on this subject.