FLIC as DNS/LES

Nick

Hello! Could you tell me if it is possible to use methods like FLIC (or large cells method etc...) for DNS or LES modeling when NS equations are solved explicitly? I'm trying to solve a back step problem in 3D in explicit way for Re=150000 but the results are poor on boundaries. One more thing – how are the boundary conditions should be implemented in this case? A wall boundary should be periodic? What are the methods for NS LES or DNS more suitable? May be it's better to use LES on Euler's equations with assumption for Re=(asymptotical infinity)? Thank you.