open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi again!

I just compared the results with a zone of 36km x 41km.They improve a little,but I still have the same problem as stated before.Thus I didn't attach my figures.

Regards,

Hanna

Hi Hanna
Could you post your model?

Regards

Hi Christophe,

Can you be more precise concerning which files you mean by model,because the data is much too big to post it here.

Regards,

Hanna

Hanna

In a first time, I thought to the latest geometry file and the boundary condition file with also the steering file and eventually if it exist the fortran file

The other files for tide are not necessary

regards

Thanks Christophe.

I attached the all the files you asked me for.









Regards,

Hanna

PS:Chi-Tuan, I think you have a problem with CCIND variable in your FORTRAN file, you posted to help me. It seems to me that it is used more than once which generates an error when launching the simulation.

Hi again,

You are right, you can add a DEALLOCATE(CCIND) at the end of CD2MSL_TPXO subroutine, e.g. just after DEALLOCATE( LON, LAT ). I used a different harmonic constants database to write this solution and was not able to see the bug, sorry (too quick, but it was a first attempt...). CCIND is also used in CONDI_TPXO or BORD_TIDE_TPXO and should only be allocated once if not locally deallocated.

Anyway, I have found the bug to solve the problem with time for JMJ database. If you still are interested in, you can find enclosed a temporary solution which will be integrated correctly in next version (the number of arguments of NODALUPV_PUGH must be changed, and because of interface_telemac2d, _2 suffix has been used). 2 subroutines have been modified, so please take both if you want to use the correction. Thanks a lot for raising this bug, it really helps me!



I will look at your model soon. Your case is very strange.

Regards,

Chi-Tuan

Hi Hanna,

After looking at your steering file more closely, I have a few things to suggest:
The first three ones should not lead to any differences, but I mention them here:
- 'FICHIER DU MODELE DE MAREE' is useless as you use tidal solutions coming from OSU,
- 'PAS DE TEMPS' = 6. (real) rather than 6 (integer), but TELEMAC changes the value automatically,
- 'FICHIER DES PARAMETRES' should not be mentionned to prevent from some bugs if you give a different name for this keyword and the steering file in which this keyword is mentionned.

Anyway, about keyword 'MAXIMUM D''ITERATIONS POUR LE SOLVEUR' = 150, as I am out of the office, I cannot check today, do you have any message in the output, such as 'NOMBRE D ITERATIONS MAXIMUM ATTEINT' (I do not remember the exact message) ? If yes, you can increase the 'MAXIMUM D''ITERATIONS POUR LE SOLVEUR'.

Compared to the test case provided in validation directory (123_tide), there are a few differences of keywords:
- 'MODELE DE TURBULENCE' = 1 rather than 4 (Smagorinski). I am not used to choose Smagorinski's model in 2D. Perhaps you can try option 1 (constant viscosity), just to check,
- 'OPTION POUR LA DIFFUSION DES VITESSES' = 1 (default) rather than 2, I am not sure that this can have a great influence,
- 'OPTION DE SUPG' = 0;0 rather than default option = 2;2 (in particular if you want to use all options to have positive depths),
- I suppose that you do not have any tidal flats over your domain to give 'BANCS DECOUVRANTS' = NO, otherwise TELEMAC-2D would crash.

Hope this helps,

Chi-Tuan

Hi Chi-Tuan,

thanks for your help.

I'll change my steering file as you suggested.
The 'NOMBRE D ITERATIONS MAXIMUM ATTEINT' is not reached during the simulation as I don't have such an output.

I tried 'MODELE DE TURBULENCE' = 1 before, but as I had bad results I changed it to 4.
As for 'BANCS DECOUVRANTS' = NO, you are right that there are no tidal flats in the zone I'm interested in.(And TELEMAC never crashed.)

I'll just try as you told me and hopefully it solves the problem.

Regards,

Hanna

Hi Hanna,
I have been following your topic closly in respect to the tidal Model and found it most helpful in applying a 2d model to tidal flows over Kish Sand bank in the Irish Sea off Dublin.

Using very similar inputs and the TPXO ES tidal model I have produced good agreement with 1 month of ADCPs velocity and elevations at two different sites along the sand bank.

For your situation it might be worth reducing bed roughness as you are using a Strickler coefficient of 27 in deep water which relatea to a manning coefficient of 0.038 (being Irish I use Manning ), Which for relatively deepwater and tidal applications might be too rough. I would have thought a mannings of 0.02 to 0.025 would be sufficient (Stricklier of 40 to 50). I also used the k-e model (turbulence model 3) which also helped somewhat in my application.

For your application given the significant local peaks in the bed bathymetry where you are interested in and where your ADCPs are located that 3-d modelling might be required to resolve more accurately the velocities.

Regards
Tony

Hi Tony,

I allready tried a Strickler coefficient of 40 and 50 before, but it didn't give any better results.I decided to take 27 because of the case of Chi-Tuan which is next to ours in the English Channel.

I also tried other models of turbulence.

I think i'll perhaps try a 3D simulation.(I didn't tried it before, because I thought that it won't be any better.)

Thanks a lot for your advises.
Regards
Hanna