open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello all,
I have noticed that GLOBAL RESULT FILE will not get created under certain circumstances. The WACRBI file(s) in the temp working directory are zero-sized. The circumstances under which, my model failed to create it are:

• Small wave spectrum energy in sheltered areas
• Bigger timesteps

Taking measures to increase spectrum energy and reducing timestep (e.g. from 4s to 2s) helped, but since all other result files (results, spectrum points) were fine, it was hard to troubleshoot.
I was wondering if it is possible to correct the issue, or implement a warning message for unsuspected users.
Best Regards,
Costas

Hello Costas,

What you describe is really strange, i mean i would see a reason for timestep, since we check time to know if the result should be printed but i do not see any reason for sheltered area.

Anyway that is a bug to correct. Could you please send your case so that i can check and correct. It seems you are in a very particular case.

kind regards

thierry

Hello Thierry,

I understand it is sounds strange. I think the main reason is the lack of wave energy. On one occasion, increasing wave height and/or adding wave growth solved the problem. The wave prescribed at the boundary was small, ~0.6m, and the domain included nearshore sheltered areas where wave energy was minimal. Minimum element size was ~2m at the shallows.
On another occasion, the GLOBAL RESULT FILE wouldn't populate no matter the wave height, unless I reduced the timestep. On every occasion, frequency discretization was tailored to describe the boundary spectrum.

I will assemble a test case that illustrates the issue.

Costas

Hello Thierry.

I found at where my problem is. I thought that in coupled mode, the NUMBER OF TIME STEP keyword in TOMAWAC was not required. Wrong. It needs to set to reflect the number of timesteps set in the Telemac steering file. If this conditions is not met, GLOBAL RESULT FILE fails to to be created.

If this shortcoming cannot be fixed, at least a warning message should be printed when one asks for GLOBAL RESULT FILE in COUPLED mode.

Best Regards,
Costas

To complicate things, COUPLING PERIOD FOR TOMAWAC also has an effect. If the number of timesteps is not a multiple of the coupling period, GLOBAL RESULT FILE won't create.
Costas

Hello Costas

Yes you have to have a time of duration in tomawac equaled to the duration in telemac3d.
As it costs quite a lot to record a global result, we avoid to do it each time we leave tomawac, so we will have a global result only when the final time of tomawac is equal to the time incremented by telemac3D calls :

at the end of wac.F you will find the test :
IF(ABS(AT-AT0-TOTAL_ITER*DT).LT.1.D-6) THEN

It is necessary otherwise you won't have the same time when you restart

I have started to write something in lecdon_tomawac to check that duration are equalled at the beginning. I understand how it can be frustrating after a long run to not have your results. I will inform you when tests will be done

Best regards

thierry

Hello Thierry,
Thank you for looking into it. The frustrating part wass troubleshooting why the GLOBAL RESULT FILE is not created. And even if you are aware of the issue, tweaking duration here, number of timestes there, timestep, coupling periods etc, at some point you 'll miss something and you 'll only know after a few hours...
So its good to do a check in advance.
Costas

Hello Costas,

I added a check in the subroutine lecdon_tomawac.f, i also had to modify homere_telemac2d.F to make it work in 2D.
I just pushed them into the trunk.

Best regards

Thierry