open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

The message from "rescue" means that Fudaa treats the initial conditions as a restart from a previous computation file, if you have not defined an initial free surface, it may be 0. I don't like this too much, as it could be that your depth computed as free surface minus bottom is also 0. I would then suggest that you remove :

SUITE DE CALCUL : OUI

from your parameter file. Your other parameters and files seem well done, so it should work, and if not we'll see why.

With best regards,

Jean-Michel Hervouet

Hello,

Thank for the help ! I just had to set a SL condition (supercritical flow), and now this works.
JM. Hervouet : what is the alternative to the restart file ? I tried to do it but it does not work. I don't know how to define local initial water height without creating a initCond.ser file (fudaa)...

PL

Hi,

The calculation works fine, but i don't have the output listing file (.sortie). It is strange because my steering file contains all the relevant keyword to define it...
Do you have an idea ?

Other question : i have built by the past a 1D model on this dam break. When comparing the results, the peak flow attenuation is very different when obtained from both methods (for identical K coefficient) : 1D attenuates peak flow ca 2 times more than T2D.
Does someone has any feedback on similar cases ?

Thanks in advance,
PL

Hello,

On comparisons between 1D and 2D you should look for publications on the CADAM european project, especially the comparisons on the Malpasset dambreak. What we find is that results by 1D models are much more scattered than 2D results. One of the reasons is that headlosses due to bends must be modelled in 1D while they are naturally reproduced in 2D. This is contradictory to your own result, generally if you add no headloss in 1D the floodwave goes too fast because the bends are not "seen", this is specially the case in Malpasset because there are two 90° bends.

With best regards,

Jean-Michel Hervouet

Hello,

For the output listing, if you do nothing you will get only the listing on the screen, otherwise you have to run the programme like this :

telemac2d -s name_of_parameter_file

Is it what you do ?

JMH

Hi,

Forget my remark about 1D/2D comparison, i did not have the same K configurations (the K modification using fudaa did not work...). I'll work it out. But when comparing identical configurations, i logically have the same trends you mentioned.

Thanks for the listing file command. But is there a way to get a 'result file' containing, for each node / element (such as RUBAR's .env file) :
X;Y;B;MAXZ;......
?
Or is there any other easy way to export easily the maximal water height to a GIS software ?

Thanks in advance,
PL

Hello,

You can get these values in your listing with the keyword :

VARIABLES FOR LISTING PRINTOUTS, which works like the VARIABLES FOR GRAPHIC PRINTOUTS

was this your question ?

Otherwise you can use the subroutine utimp_telemac2d.f to write what you want on a given file.

With best regards,

Jean-Michel Hervouet

Hi,

I already included this in the VARIABLES FOR LISTING PRINTOUTS, but the log only contains each parameter value for each node, identified with its number.
To manage with the data with a GIS software (eg for a maximal extents flooding map), I would need to have directely the parameters value on each node's coordinates X and Y.
I had a look at the subroutine, but i am absolutely not familiar with fortran...
Maybe fudaa can do such a thing ?

Otherwise, i have a question regarding the results i got from the computation. I have hydrographs characterized by a non-realistic peak flow. This occurs for the firsts "wet" time steps. You have a picture enclosed.
When comparing with the SL profile on fudaa, i can see that to the high peak flow value corresponds a high water surface. Thus, these high peak flow values don't seem to be generated by the integration method (fudaa).
I have also enclosed screenshots of the mesh around the 2 profiles with this feature at the first "wet" time step.
Maybe the problem might be that my mesh is not refined enough? Or can it be solved by computing using specific numerical configuration ?

Thank you inadvance !
PL

Hi,

Regarding my hydrographs, I think the problem comes from my cells size in a narrow section. Between two time steps, the wet cells pass from several to only one. In order to conserve the discharge or water height, telemac should then affect a abnormal high velocity value, which i can see on the velocity profiles.
Then, i assume telemac takes this value into account for the following time steps / cells, which, for a coherent water height, leads to overestimate the discharge. Can it be smthg like this ?
If yes, what would be the minimum number of wet cells under which we should not go to avoid this problem ? Do you have any feedback on this ?
Furthermore, up to how many cells can telemac compute with reasonable computation time (even though it also depends on the cells size) ?

Thanks in advance !
PL

Hello,

With keywords you can only have a constant depth or constant elevation. The other way is to program subroutine CONDIN.F, see e.g. the test case called gouttedo, where a gaussian hill of water is programmed as an initial condition.

JMH