open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello！
I want to run a Telemac2d simulation coupling with Sisyphe. The simulation reports an error later.

GRACJG (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 500 RELATIVE PRECISION: NaN
EQUNOR (BIEF) : EXCEEDING MAXIMUM ITERATIONS: 50 RELATIVE PRECISION: NaN

It seems that the calculation cannot converge.
A hydrodynamic simulation has been done before, by telemac-2d only.It works well.
I wonder if there were some problems with the boundary condition file? I'm not very clear about the boundary settings when considering TPXO tide data.
Here are the steering file and some information of my case.







Thankful for any help!

Gareth

Hello Gareth,

Can you change the extension of the PE00003-00001_2020-05-25.log e.g. with a .txt extension and upload it again? It cannot be read.

Anyway, you can also build a debug configuration with additional checks which could help tracking your NaN.
Depending on the compiler you use, you can have a look at the systel.edf.cfg configuration file in the configs directory (same level as sources and examples).
If you use gfortran compiler, have a look at S9.gfortran.debug or C9.gfortran.debug

Hope this helps,

Chi-Tuan

Thank you pham!

Due to my carelessness, I uploaded the wrong file format before. I uploaded the txt version this time.

I'm mainly in doubt about the boundary type setting in BC file when TPXO data is considered, because there is only one open boundary, which alternately works as the inflow boundary and outflow boundary. So, it is not possible to set the inflow boundary to 4 5 5 5 and the outflow boundary to 5 4 4 4, as suggested in the Forum.Can you give some advice on this question？

Thanks again for your kindly help.

Hello Gareth,

If using Thompson BC (OPTION FOR LIQUID BOUNDARIES = 2), you can prescribe both elevation + velocities in the BC file (code for the first 3 columns: 5 6 6). You can have a look at what is done in the tide example.

I would suggest you :
- to decrease your time step and see if your problem still occurs,
- to compile a release of telemac with more debug options as suggested in my previous post. The compiler may tell you exactly in which subroutine and at which line your error occurs (first NaN appears).

Hope this helps,

Chi-Tuan

Hellow pham!

I'll try your suggestion to see if it works.

Thank you for your kindly help!

Gareth

Hello Gareth,
As another general troubleshooting step, may I also suggest to run another computation that stops just before the first error occurs. Then you can inspect the results which may give you a good indication to the source of the error.

I have also noticed big evolutions just before the NaNs:If big evolutions occur near the boundaries, that could mean trouble to your boundary conditions.
Best Regards,
Costas

Hello cyamin.

I've done some tests. The problem really happened near the boundary. I will do more cases to test.

Thank you for your help

Hi Gareth,

When using TPXO boundary conditions, I use
5 4 4 ..... 5
in my boundary file.
so 5 4 4 for the hydrodynamics, which the TPXO data base will use to set levels and velocities at the boundary. So the bed evolution needs to be fix (5 for LIT in the boundary file), to not to cause errors of the water discharges through the boundary.

Michiel

Hello Michiel,

Thanks to your reply.I will try your suggestions.

I have another two questions to ask.

1.Should I only use one boundary file for telemac and SISYPHE?

2.According to SISYPHE's user manual, LIT (in boundary file)in telemac is replaced by LIE / LIC in sisyphe, which represents bottom elevation / concentration. But how do these two keywords work in the same position in a boundary file? Or, maybe they only work for "only bedload case" and "only suspension case" respectively ?

All regards
Gareth

Hello Michiel,

I left out a question.
Why use 5 4 4 for the hydrodynamics, not 5 6 6 which means imposed levels and velocities at the boundary?

Kind regards

Gareth