open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all,

My domains is a rectangular basin with 3 liquid boundary with imposed flow rates in upstream and 1 liquid boundary with imposed H at the down stream.

I have 20 plans on the vertical.

I would like create a vertical profil of the horizontal velocity for the 3 upstream bondaries. Indeed, I would like that the horizontal velocity is equal to 0.3 m/s in the plans 3 to 7 and 13 to 17 but 0 m/s in all other plans.

I found on this forum all the topics but I don't understand what is the modification that I will do in vel_prof_z to obtain the good result.

I wrote this code in the vel_prof_z:
IF(IPLAN.EQ.3 .OR. IPLAN.EQ.4) THEN
UBORL%R=0.30D0
ELSEIF(IPLAN.EQ.13 .OR. IPLAN.EQ..14) THEN
UBORL%R=0.30D0
ELSE
UBORL%R=0.D0
ENDIF

But the velocity profil seems homogeneous on the vertical, I don't understand this result ...

I join my files here.

Can you help me please? I think it's not very complicated but I'm new user of telemac ...

Hi
vel_prof_z in a function so you have to give the value to return to vel_prof_z, not to uborl

Regards

Thank you for your very quick answer.

I agree but when I changed vel_prof_z, I had vertical velocity not horizontal velocity and all flow rate was at the bottom, it's not homogeneous on the vertical ...

When I vizualise my results, I see that the velocity vectors are all on the z-axis ... is it normal ?

Thank you again for your help.

Gwendoline

Hi Gwendoline

Sorry but I don't understand what you mean...
In the boundary condition of the 2D mesh, you defined the liquid boundary with imposed discharge.
In vel_prof_z, you describe the vertical velocity profile which is used to impose the discharge.
According to your file and the correction you should have made, on the boundary, you see U and V set to 0 on some planes and non zero values on plane 3, 4, 13 and 14.
W is compute so it could be different from 0. Maybe if it's value is greater than U and V?
Try to join all the files (in a zip) in order to let us try to run the simulation.
Regards

Hi

Actually, I would like represent two flow rate inputs on the vertical.

For my first simulations I open all the wall on the vertical (via the 2D mesh) but it's not what I want. I would like 2 sources on the vertical for the 3 open input boudaries (via the 3D mesh) but I read that it's not consistent to change the kind of boundaries on the vertical (liquid and solid boundaries for example). Consequently, I look for another solution to represent the wanted model thus I think that imposing the horizontal velocity on the vertical I can obtain what I would like but the results are not very convaincing ...

I know that my configuration is not conventional and it's a few difficult to explain in english and by writting what I want to do.

Thanks

Gwendoline

Hi
I tested your model and in my opinion it gives the results you want.
This is a view of velocities on the planes


You could clearly see the 4-6 and 14-16 with velocities...

If you want different values in the plan you could modify the vertical profile. You should see the value you enter as a shape for the real values as the final velocities depends on the discharge and the local cell size

regards

Hi,

Thank you again for the previous coments.

I come back because I wondered me if it is consistent to use a constant viscosity rather than a k-epsilon model in the present case ?

Indeed, I remarked that if I used a turbulence model, my system is not stationnary and inhomogeneous on the vertical. However, when I consider a constant viscosity on the horizontal and vertical the results is satisfaying but is it the real result ?

Gwendoline

Hello,

If there is no wind in your domain I would advise to take a constant diffusion on the horizontal and the mixing-length model on the vertical, e.g. the Nezu and Nakagawa option. In our applications the mesh size is generally too large for a correct behaviour of the k-epsilon model and the turbulent diffusion is thus underestimated.

With best regards,

Jean-Michel Hervouet