open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello

I'm using the Telemac software to study sediment transport along the Belgian Coast. I internally coupled Telemac-2D and Sisyphe and obtained satisfying results. Now I want to include waves but I want to use the chaining method instead of internally coupling the 3 modules to reduce the computation time. I've added the following settings:

Telemac-2D steering file:
- NAMES OF CLANDESTINE VARIABLES = 'MEAN DIRECTION' ; 'PEAK PERIOD TPR5 S' ; 'WAVE
HEIGHT HM0 M'

Sisyphe steering file:
- EFFECT OF WAVES = YES
- HYDRODYNAMIC FILE = Tomwacresults.slf (results of Telemac2D-Tomawac simulation
using the same mesh and containing the following variables:
HM0,DMOY,SPD,ZF,WD,VX,VY,FX,FY,FMOY,TPR5,TMOY,TM01,TM02,TPD,BETA,UX,UY)
- STARTING TIME OF THE HYDROGRAM = 0.

After initializing and copying all files, I get the following error message:
'Missing wave data
Wave height HM0
Peak period TPR5
Mean direction'

I've tried a lot of other settings I found on the forum, but nothing seems to help. I've attached my 2 steering files. Does anyone know what I am doing wrong?

Thanks in advance!

Hi Laura,

I'm not really sure that this chaining method is planned in the "classic" cases of use..
I think Sisyphe is looking for the wave variables in the result file provided at each time step by Telemac, and does not consider the "hydrodynamic file" that you provide. Maybe you can try with the "wave file" keyword in the Sisyphe steering file, but it's possible that you'll have the same error message or only the first (or the last) time step will be read..

Another safer solution is to manually program the reading of your wave results file directly in condim_sisyphe.f called at each time step, either from the "wave file" keyword from Sisyphe or with the clandestine variables you directly give at Telemac.

Hope this helps.

Regards,
Laurent

Hi Laurent,

The manual says indeed that in case of adding a "wave file", only the last time step will be read. But since I have a unsteady wave field, this is not a good solution for me.
If I add the subroutine condim_sisyphe.f to my fortran file, do I have to change anything in the code? And then I just have to add "binary data file 1", "names of clandestine variables" and "wave driven currents" to my telemac steering file?

Hi Laura,

The morphological changes will eventually render your wave calculations invalid. That is why online coupling is the only proper way for long calculations but it takes too much time.

Regards,
Costas

Hi Laura,

If you only want to consider wave influence on sediment transport, I think the simplest way is to program into sisyphe by the way of the subroutine condim_sisyphe.f that does nothing by default. A simple example with no waves is presented in the subroutine, with comments. You have to program something here for reading your wave data from the "hydrodynamic file" (SISHYD) or your "wave file" (SISCOU), at the right time step corresponding to your t2d/sis simulation, and eventually interpolate them between 2 time steps (depending on how often you did graphical printouts from your previous t2d/tom simulation). Your wave data have to be stored in HW, TW, THETAW arrays, and you needn't to manage depth and currents, as t2d will do it for you.
In this way, "clandestine variables" and "wave driven currents" are useless and not taken into account.


However, If you also want to consider "wave driven currents", you have to provide the wave file (toma_BCG_v711_mercator_sim9.slf) to t2d steering file (thanks to the "binary data file 1" for instance, and by activation of "wave driven currents"), and to program the reading of your wave data to provide the driving forces FX, FY to the subroutine prosou.f to fill the FXWAVE and FYWAVE arrays at the right time step, in order to t2d can compute wave driven currents.
If you choose this solution, maybe you can directly access the wave data (Hs, Tp, dir) into sisyphe.F (without providing them by the "hydrodynamic file" or "wave file") and put them in the HW, TW, THETAW arrays.


If you also look at the morphological changes, Costas is right, you have to be careful with the validity of wave results, as long as the bottom changes are not too high.
Moreover, I'm not really sure that you're going to save too much time by doing a first computation t2d/tom (long) and then a t2d/sis with results of tom (fast). The 3-way coupling may be simpler for you and probably not much longer.

Regards,
Laurent

Hi Laurent,

Thank you for this information! I will try to program the subroutine condim_sisyphe.
I have to calibrate some parameters for the sediment transport, while the wave field will not change. So I only have to do the Tomawac simulation once and then I can use the results in all the test runs for Siysphe. That's why I'm not very fond of 3-way coupling. As a final test I will off course couple the 3 modules.

Regards,
Laura

If you just want to calibrate the parameters of the solid transport formula, another solution is to do a sisyphe simulation alone, by reading the full hydrodynamic results (H, U, V, Hs, Tp, Dir) from a previous t2d/tom simulation, and then no need to program.

But be aware that in a rigorous way, the morphological changes require to regularly update the hydrodynamic conditions (by coupling) which will then modify the solid transports, and so on... So maybe that the optimal parameters for the formulation will not be exactly the same with and without coupling.

Laurent

Laura,

For completion: when running uncoupled, SISYPHE will read the wave information from the wave file, to be specified as (WAVE FILE= ). It does not look for wave information in the hydrodynamic file.