How to account the subgrid scale turbulent kinetic energy in LES simulation??

You can model the transport equation for the sub-grid scale kinetic energy via its own balance equation - which is quite well defined. Much like the balance equation for the RANS TKE. In the case of the destruction term, epsilon, in the sub-grid case, one uses a mixing length tactic, and models the dissipation based on the local filter width and the local sub-grid k.

Much work has been done by Menon's group at Georgia Tech, and I would consider his work on this, at Industrial level apps. at least, the best in the world. However, his work is for the compressible case which is of interest to me, if you're working with incompressible flow then Lars Davidson is probably your best reference.

A last note, one can either model the constants dynamically, or not, and I would start by using the model constants rather than dynamically computing them...

Cheers

hi...cell "You can model the transport equation for the sub-grid scale kinetic energy via its own balance equation " you mean i need to build a equation to solve the k? but i have not see any reference talk about this. I also use the consant model, and how do you deal with the wall condition? how to solve the y+? thanks!

You're after quite a bit of information there. It has been done, 10-15 years worth by various people, and longer by others so it is out there. I would read the papers by the people I referenced to get a good understanding of the model, then worry about implementation details etc.

Look in AIAA for Menon, and look on his website for the references at Gerogia Tech, there are many papers. Like I said if you want to consider incompressible flow look at Lars Davidson.

Not sure what you mean by "I also use the consant model", I was talking about something quite subtle to do with test filters etc... I'd ignore that at the moment and get reading!!!