open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I have some questions about the calculation of COEFFICIENT FOR HORIZONTAL DIFFUSION OF VELOCITIES in Telemac3D and for VELOCITY DIFFUSIVITY in Telemac2D.

Firstly, do these two parameters have the same physical meaning ?
Then, I read on the Telemac3D and on the Telemac2D user manuals that their default value is set by default as 10-6 . This value corresponds to the physical value for molecular + turbulent viscosity of water.
But I also read in Telemac3D reference manual that the coefficient sets the value of the global viscosity (dynamic + turbulent viscosity).

So based on this last definition (T3D ref manual) and the value of the dynamic viscosity for water (1.002x10-3 kg/m/s), the coefficients named COEFFICIENT FOR HORIZONTAL DIFFUSION OF VELOCITIES (T3d) and VELOCITY DIFFUSIVITY (T2d) should be set as 10-3.

So is there a mistake in the ref_manual ? Did I missunderstand the calculation of that coefficient ? Or did I make a mistake ?

Thanks for your attention, Julien.

Hi,

I just wanted to know if someone has an idea about these questions.
I would say that there is a mistake in the T3D reference manual (dynamic instead of molecular viscosity).

I also can't have a value for the turbulent/eddy viscosity which is the other term in the coefficient. I only know that eddy viscosity can reach 10-3 at the surface of oceans. This viscosity is divided by a "closure coefficient" in the transport equation for turbulent kinetic energy and dissipation rate which are the governing equations of the k-epsilon model.

Thanks for your attention, Julien.

Hello,

I can't really answer you regarding TELEMAC-2D but for TELEMAC-3D here coms an answer:

- the COEFFICIENT FOR HORIZONTAL DIFFUSION OF VELOCITIES is the kinematic (molecular) viscosity of the fluid. In case you use a turbulence model with a constant turbulent viscosity it is the sum of the molecular+turbulent kinematic viscosities (in m2/s);
- in case you use a 1-equation or 2-equations turbulence model (mixing length, k-epsilon), you should only specify the molecular kinematic viscosity there and TELEMAC will calculate the total viscosity itself (which will be space-time variable). k-epsilon is a widely-used turbulence model, which is more general than the mixing-length model. In case you don't know what to use as a mixing length model, and you have a fully turbulent flow, I would suggest trying the k-epsilon model.

I hope this answer will help you.
Best regards,
Agnès

Hello,

thank you for your very clear answer.
I am currently using the k-epsilon model for the turbulence.
Now I understand the role of this coefficient better.

Regards, Julien.