open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello all,

I encountered an error with a coupled model setup that I want to share. I briefly describe all the steps I took until I got to the source of the error:

Initial setup was:
-Global results file is a result from a Tomawac-Telemac2d coupled computation
-Next computation is a Tomawac-Telemac2d-Sisyphe coupled computation
-DISSIPATION BY STRONG CURRENT =2

I am getting a segmentation fault error.


Then I try to continue the computation with Tomawac only and I got an error in LECSUI:
The number of frequencies in both cas files (the one that created the global results file and in the next computation) is 20. The number of frequencies in the created global results file is 20, plus 2 more entries (DEPTH and CURRENT-WIND). I change the next computation cas file to 21 frequencies and then I get a segmentation fault error after SEMIMP is called, which is a hint to disable DISSIPATION BY STRONG CURRENT.

Success. I set DISSIPATION BY STRONG CURRENT =1, it works again. I set NUMBER OF FREQUENCIES back to 20, and I get the same BAD NUMBER OF FREQUENCIES error in LECSUI.

To sum it up, it seems that there is an issue with the way frequencies are read from the previous computation file and the Westhuysen formulation while continuing a coupled computation. The second is easily overcome by disabling DISSIPATION BY STRONG CURRENT, or setting it to 1. However, for the first one I don't believe that increasing the number of frequencies is appropriate because it could mean that it is reading values from the DEPTH or WIND-CURRENT entries (which by the way do not contain DEPTH of WIND-CURRENT values, only small negative values in the order of 10E-4).

Can anyone verify this?

Best Regards,
Costas

Hello Costas,

I guess you found a bug. Reading lecsui and what you wrote, I guess it comes from a mixing up with the term couran which can means that you have a couran file or that you need a couran for couplage. It must be a case that we didn't think of. Could you please post your full configuration so that i can use it as a test configuration for this problem ?

thank you
regards,

thierry

Hello Thierry,

I forgot to mention that I am on the v7p1r0 version. No current files specified anywhere but the previous computation in Telemac2D, as I am trying to continue a coupled model.

Here are the cas files. They have changed a bit since I am trying to overcome the issue:


Please advise me if you address this issue.

Best Regards,
Costas

Thank you Costas

Yes i understood about the current files, that's normal.

Could you also post your geometry files, i could use another one but i think it is safer if i use the same as yours.

regards

thierry

OK, I will prepare a folder with the previous computations as well, so that you do not need to run the first if you don't have to. It took a few hours in my workstation. I will sent you the link to download in a pm.

Regards,
Costas

As you like costas, but i just wanted to have the
0.Geometry/geo_yfist_v2.slf and the 0.Geometry/geo_yfist_v2_544-4.cli files to run the cases on the same geometry a you.

up to you

regards

Thierry

Hello Costas

I am not sure that i have solved your problem but i solved the ones (at least some of the ones) i encountered with your case. It was about reading and writting next compputation. Thanks to your descriptions i found some error. As it was disturbing memory it can have different effect according to compiler.
The two files i attached will be in next release but if you want to check it on your case you can try to replace those in the sources of tomawac.

Hello Thierry,

Thank you for letting me know. At the moment I am having a big queue of v7p0r1 computations, but I will test it as soon as possible.

Best Regards,
Costas