open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

In the 3D model I ask NUX, NUY and NUZ as variables for 3D graphic printouts.
I want to compare this output from the models with different number of vertical layers.
Could you tell me how these variables are calculated, with which formula?

Best regards,
Tatiana

Hello,

It depends on the turbulence model chosen. You can find how they are done in :

subroutine viscos.f (constant viscosity)
subroutine visclm.f (mixing-length, various sub-models)
subroutine viscke.f (k-epsilon)
subroutine vissma.f (Smagorinski)

They are all called in telemac3d.f

With best regards,

Jean-Michel Hervouet

Hello,

I compared the output of two runs: with 5 vertical layers and 15 layers. The calculated viscosities (NUX and NUY) are different in these simulations. However, there are no differences in the calculated velocities. I've attached Word with figures.
We use Smagorinski horizontal turbulence model and Nezu and Nakagawa vertical turbulence model.
Why the variation of the number of layers has an impact on the calculated viscosities?


Best regards,
Tatiana

Hello Tatiana,

When you refine on the vertical, you get steeper local gradients and they are used in the computation of viscosities, so it is normal to get different values. For Smagorinsky another reason is that this model is supposed to give the turbulent viscosity that occurs at sub grid level, so it highly depends on the mesh size by construction.

With best regards,

Jean-Michel Hervouet