open TELEMAC-MASCARET The mathematically superior suite of solvers

Okay, perfect I'm on it.

I checked all the processes and is failing when I add USER_TR3.f

Hello Juli,

You should have understood that your USER_TR3 is suspicious. :)

As you try to modify the surface boundary conditions, perhaps you need to refine your vertical mesh.
Otherwise, you should think if what you want to prescribe is good or not. In your case, you try to prescribe a Dirichlet condition, the same at every point, but varying in time. Perhaps a Neumann condition is better.
Have you tried an homogeneous Dirichlet condition (= 0 everywhere and everytime, or constant not equal to 0 at any time)? Do you have the same behaviour (surely)?

Hope this helps,

Chi-Tuan

Thx for your advice:
1. refine the mesh at the top.
2. Type of boundary conditions, I tried already with KLOG, KDIR, and KENT

About your question yes, first I tried by default: BC type Neumann and prescribed value equal to zero, later constant value for all the surface points and is working perfectly.
Even I tried the same with small domains free surface BC changing in time and is working.

But still, the problem is when I tried to run more the one core

Hello Juli,

Can you send the whole files to run your simulation and reproduce your issue?
Can you also explain what you expect what is your current problem?

Chi-Tuan

Hello, Dear Chi-Tuan.
I did a couple of things to see if my implementation (Free surface boundary conditions) worked.






Cas:

Geo:

cli:

fortran files:



When I changed the keyword PARALLEL PROCESSORS : 4
I got this


I tried changing adv schemes = 14 and the simulations start but crashed anyway.


Thank you very much

could you identify the problem?

Do you have any other advice?
Thanks in advance