I am trying to get obtain an accurate Strouhal number for turbulent flow of 4e6 over a cylinder. If anyone has done this type of problem and has been succesful in getting the correct frequencies, please advise me of the set up. I am using LES, and when I calculate my Strouhal Number, i get 20% error (with reference to Schlictiing)

did you try to shift your Smagorinsky constant ?

if i remember well, i've seen such works where oscillations behind a cylinder were much more well computed using a CS=0.13 CS instead of the default. However, Germano dynamic model would be a great progress in fluent.

I was reading up on the Cs constant in the Smag.-Lilly sgs model. The default value of .1 is good for transition flows, but the original proposed value is .23 and I believe that might work better for my problem.

I am unfamiliar with the Germano dynamic model. What is it?

the germano uses double filtering of different size.

The problem is solved in LES using a test filter twice as large as the original including a smagorinsky model for SGS. Then the problem is solved with the usual filter (of lower size) with the same smagorinsky model BUT it makes the assumption that the SGS have the same properties for both filtering SO in the gap between the test and the normal filter you have a correspondance between the SGS model of the test and the solved scales of the normal filter. You can here make a correspondance and then evaluate a value for the CS witch is locally dependant of the flow properties.

Of course, this is a simplified approach of the original germano model but you have the key idea. many improvements were made since espaceially in the sgs modelisation.

see reference below :

@ARTICLE{GER91, author = "\textbf{Germano M., Piomelli U., Moin P., Cabot W. H.}", title = "A dynamic subgrid-scale eddy viscosity model", journal = "Physics of Fluids A", year = "1991", volume = "3", number = "7", }

for more exhaustive list you should check the papers made by Paviz Moin (AMES laboratory i think...)