open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi,

I am trying to make a model with a .... meters long weir.
Normally, the weir is not horizontal but always 30cm above the topography. I did not manage to make the model work by the way, so I divided my long weir in 8 horizontal weirs but it still doesn't work, with the message :

---

ETAPE DE CONVECTION

---

ETAPE DE DIFFUSION-PROPAGATION
GMRES (BIEF) : ECHEC DE L'ALGORITHME R= 0.000000
SI LA MATRICE EST DIAGONALE, PRENDRE
COMME SOLVEUR LE GRADIENT CONJUGUE.
GMRES (BIEF) : 1 ITERATIONS, PRECISION RELATIVE: NaN

In attachment you can see all the files I used to launch the simulation.
My method of resolution if the n°7 (SOLVEUR = 7 : gmres).

I am trying to launch the simulation with SOLVEUR = 1 (gradient conjugué) but I am not sure it will work.

Does anybody has a solution ?

Regards,

Gaëlle.

Hi Gaëlle,

If the system you try to solve is symmetric, the conjugate gradient method is recommended (e.g. with TREATMENT OF THE LINEAR SYSTEM = 2 - wave equation). the GMRES method works for non symmetric system (e.g. with TREATMENT OF THE LINEAR SYSTEM = 1 - coupled).

By the way, are sure that the problem does not come from your mesh and topography? Do you have any vertical slopes (that is not allowed for TELEMAC simulations)?

Hope this helps,

Chi-Tuan

I do not know a lot of thing about the numerical parameters, but mines are :
TREATMENT OF THE LINEAR SYSTEM = 2 - wave equation
SOLVER = 7

I after tried to change the SOLVER keyword to 1 (conjugate gradient) but it did not work.


What do you mean about vertical slopes ?
You can see the topography upstream from my weir in attachment : indeed it is not very regular, do you think that could be the origin of the problem ?


Thank you for your help.

Gaëlle.

Hi Gaëlle,

By "vertical slope", I mean bathymetry that is very steep. For example in your last figure if you can make a zoom around 1,050 m or 2,000 m (in particular), the bathymetry seems less flat than along the first 1,000 m.

Can you attach "all the files I used to launch the simulation" from your first post. I have not seen any attached files in this topic.

Regards,

Chi-Tuan

Hi,

Indeed it seems to be a problem of topography, because I tried to launche the simulation with a topography very regular near the weir and there is no failure of the simulation now

Anyhow, I give you the files I used for the simulation in attachment. Some of them are missing, I will dowload them in the next post since I can attach all of them in the same message...














of topography

The other files :

Hi

Two remarks:
1- your fortran file shows reference to old subroutine. I suppose this is nt really compatible with the telemac version you use.
2- the geometry file is missing

Regards

That is true that the versions of the subroutines are older, but I have always used these ones and it has never posed problem.

Sorry for the geo file, I was pretty sure to have uploaded it...

Hi Gaelle

The subroutines works because there is no arguments so no changes since the old version but take care about such methodology because this is not a global rules for all subroutines.
But in such case, you will obtain compilation error.

I also noticed that the weirs file is missing.
(Take care about the size and also the file format, if necessary zip the file)

Regards

I try to Zip the geo file...
Hope it will be uploaded.