From reading a bunch of books I know general recommendations for LES are pretty big, here are 2 things I can recommend: Boundary Layer - ridicoulus refinement necessity. Growth rate of 1.1 and atleast 20 cells for an adequate simulation, good results are obtained at 30 to 40 layers Turbulence - for turbulence you can use a quantity to display in your fluid field, which relates your cell size to eddies. You can then use this to determine if your flow field is resolved enough. For this: its "Ratio of integral length scale to grid length scale R_L" if you obtain l0 your integral length scale then you can divide this by your mean cell dimension which is delta = (cell volume)^(1/3) For LES its recommended to have this value not be smaller than 10, so you can create a visualization of your cells and you can threshhold anything above 10, anything below would need refinement. The ressource here is S. Pope