problem about numerical scheme in LES.
 

dear all (1) i remember somebody said that he used TVD scheme in LES with eps==0. i did not find this type of papers. could you give some information about it? could i use Harten Yee TVD scheme to campture the shock in LES code? (2) I did my LES code with MaccorMack scheme. no matter what initial condition was set, the results were laminar flow! i set the CS in SGS model is zero, but obtained the same results.oh, the Smagrinsky SGS model is used. (3) the phenomena of odd and even oscilation occured.after added the 4th disspation term ,the viscous becomes greater.

in my code, the viscous terms were discreted by 2th center scheme in both pre-step and cor-step. i computed the channel flow, period BC in stream and spanwise. the average pressure gradient is set 2./re . in normal direction, d_p/d_n==0 is set. are there any question about it?

please give me some advice. best regards

Re: problem about numerical scheme in LES.
 

TVD scheme is suitable for capturing strong shock. Up to now no single schemes can habe both high accurate for LES and does produces oscillation in capturing strong shock. The minmod function in TVD is dissipative, in multi-dimension, vortices are dissipated quickly.

Re: problem about numerical scheme in LES.
 

There has been some recent work on so called implicit LES (ILES) where the subgrid terms are not expressed explicitly but arise naturally from the TVD scheme - basically a DNS code with a monotone convection scheme.

Re: problem about numerical scheme in LES.
 

could you give me some referrence about it? thanks.

Re: problem about numerical scheme in LES.
 

I think the original paper was

Ann S. Almgren, John B. Bell, William G. Szymczak, A Numerical Method for the Incompressible Navier-Stokes Equations Based on an Approximate Projection, SIAM J. Sci. Comp.,17:2, March 1996.

Also try searching google for "implicit large eddy simulation"