open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all!

I have tested several combinations of turbulence models for both horizontal (UV) and vertical (W) directions in case of flow behind a bridge. One of the best results compared to field measurements was obtained with Smagorinski in both direction (UV + W).

However, I cannot find a lot of references mentionning the use of Smagorinski model for vertical turbulence, though it is a mixing length model in its form. JMH also states in "Hydrodynamics of free surface flows" that "this model is best understood in 3D".

Does any user have experience to share regarding the use of Smagorinski for vertical turbulence modelling?

Thank you in advance!

Best regards
PL

Hello,

In some sense Smagorinski is like Large Eddy Simulation (some specialists will deny this but they are free to give us explanations...). So its efficiency probably depends on the mesh refinement. Having a different model on the vertical is a bit schizophrenic, as we have an assumption of isotropic turbulence behind Smagorinski and k-epsilon (hence "model best understood in 3D"), anyway it is true that with aspect ratios of sometimes a factor 1000 in our applications it is difficult to claim that turbulence is numerically isotropic.

Regards,

JMH

Hello,

Thank you for your quick reply.
This model seems then well adapted to my case which does not have such high aspect ratios (max 20).

Best regards
PL