DNS and LES resolution
 

Hi everyone
I am a bit newbie on the LES and DNS subjects and I am wondering if someone can help me clear out a couple of question.

First:
The kolmogorov turbulence length scale for DNS after what i have learned the the basis for grid size? Is the Taylor length scale the correct basis for LES?

Second:
To clear it out. How many cells are needed to describe the motion for the smallest turbulence length?

Third:
What is the common way to determine cell size for LES computation? I am specially interested in a wing transition problem.

Thanks