Hi,
Please excuse my ignorance surrounding modelling of turbulence, however I am trying to gain a better understanding of the application of the Smagorinsky turbulence model to RANS simulations such as those of TELEMAC-3D.
From what I have managed to understand, the Smagorinsky model is typically employed during LES simulations, where turbulence scales exceeding the grid size are captured directly by the fundamental RANS equations, and scales below are filtered out through the use of a filter function. Turbulence diffusion for these smaller scale eddies is consequently accounted for through the use or an additional eddy viscosity provided by the Smagorinsky model. Although I may be wrong I believe that this value is typically smaller that that computed by a model such as K-epsilon (which accounts for turbulent diffusion across all length scales).
As I understand TELEMAC does not employ any low-pass filtering like the LES model and applies the computed eddy viscosity from the Smagorinsky model directly to the diffusion term of the RANS equations.
My question is what is the overall implication of this in terms of the results? Are these computed values in of eddy viscosity in fact smaller than say k-epsilon and what is the difference between TELEMACs implementation and an LES simulation.
I could well be missing something fundamental here and any help/ guidance on this would be greatly appreciated.
Cheers,
Toby
