open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everyone

I think I have a numerical problem, as shown in the pictures, there is a big alteration in some elements of the mesh.


I tried to use the keyword:
FREE SURFACE GRADIENT COMPATIBILITY = 0.1 (close to 0)
IMPLICITATION FOR VELOCITY = 0.8 (close to 1)
IMPLICITATION FOR DEPTH = 0.8 (close to 1)

To be able to make the problem more stable, but not better.

Does anyone else have any idea what the problem is?

PD: Posting in forum telemac2d and not in sisyphe because I think it is a problem numerical and not of sisyphe.

Attachment files

Joaquin

Hello Joaquin

This is strange that you have only local unstability. This looks more like non hydrostatic effects due to steep bathy gradient frontal to normal flow. I don't have the geo file to check it up.
Meanwhile, you can :

1- set mass lumping on velocity =1 (on water depth also)
2- change discretisation of velocities from 11 to 13
2- add some velocity diffusivity (default 10-6)

If it does'nt work, you can try with boussinesq to deal with non hydrostatic effects.

Before doing all these steps, I recommend to disconnect sisyphe, switch off variable time step.

I hope that this helps

with my best regards

Riadh

Hi Riadh, thanks for your reply.

I forgot to say, I first deactivated sisyphe and for my final solution obtained, I use it for my initial condition when active sisyphe. Only with hydrodynamics I do not have instability.

Likewise, add your recommendations (1 and 3) and I have exactly the same solution, now adding with sisyphe I do not know what will happen.

Regarding your recommendation 2, use the keyword "DISCRETIZATIONS IN SPACE = 13; 11; 11; 11" and it throws me error, so leave it in default (11; 11; 11; 11). I do not know if the version I'm using v6p3 will influence something: S.

I attach cas files, with and without their recommendations, which gives me the same result.

Joaquin

Well, it is really recommended to upgrade to the latest version. We are with version 7.2 and we will move to 7.3 within next few weeks. So you have at least 3 version backward.

Based on what you said, the unstabilities apprears when coupling with sisyphe, is it? So it looks like a problem with tidal flats or wetting/drying aspects.
in that case:
1- change your advection scheme to scheme 14
2- reduce the coupling period to 1,
3- and if not suffuicient reduce your time step.

with my best regards

Riadh