open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I’m trying to continue a Telemac 2D parallel simulation using a previous simulation.

My problem is that the last computation file (ex. rd2_name.slf) that I use to as “PREVIOUS COMPUTATION FILE” is huge (several Go) and so it takes several hours for Telemac 2D parallel to restart a simulation (i.e reading the file).

So I try to decrease the size of the previous computation file as recommended in chapter 4.1.3 of Telemac user manual (version 7.0) by extracting the last time step of the file used for the continuation. For information, I’m working with BlueKenue and Telemac v7.0.

So I have two questions :

1. Is there a simple (and efficient) way to extract the last time step of selafin result file using BlueKenue ?

For now, the only way I found is to manually extract each variables (WATER DEPTH, etc.) using Tools -> Extract surface -> Temporal maximum and selecting only the last frame as start and end option. Velocity UV also has to be manually exported as VELOCITY U and VELOCITY V.

2. When I used the new selafin file created with the method described above (file that include only the last time step), the pre-processing steps of Telemac parallel are very fast, but the simulation time restarts to zero. So, is there a way to indicate, for example in the new selafin file, the time of this last time steps ?

I have already tried to indicate this information in my Telemac 2D parameters file (t2d_name.cas) without success (I add the line “ORIGINAL DATE OF TIME = 1900;01;01 and ORIGINAL HOUR OF TIME = 00;20;00, since my first simulation stops after 1200 seconds).

Thanks in advance for your help !

Regards

Hi

for the point 1, it will be better to use python scripts
for the second point, with a restart, you could choose to reset time or not, there is a keyword for that)

Regards

Hello,

There are lots of solutions to your problem:
1-The easiest one, but not the most efficient, is to restart your simulation in sequential mode (not parallel) and run only one time step. This will generate a very small and efficient file which can be used as PREVIOUS COMPUATIONAL FILE
2- if you use fudaa-prepro, this latter extracts directly and easily the time step you want
3- with BK, I dont know a straight way to extract a time step, but you can build a new selafin object and then add layers (bottom, elevation,...) one by one
4- you can use Python scripts

You can use the following keywords :
INITIAL TIME SET TO ZERO = NO
you can choose the record number you want Telemac to use for restart
RECORD NUMBER FOR RESTART=


I hope that this helps

Riadh

Hello,

Thanks for your answers !

I will try your propositions to extract the last time step of my simulation. Until now, I use the proposition number 3 of riadh (adding layers one by one to a new selafin).

For my second question, by default the keyword “INITIAL TIME SET TO ZERO” is supposed to be “NO”, but I add it explicitly in my parameters file and it doesn’t solve my problem.

It looks like Telemac reads the sealfin file and its variables, but it restarts from zero (see attachment).





My parameters file looks like this now :

COMPUTATION CONTINUED = YES
PREVIOUS COMPUTATION FILE = r2d_nom_fichier_restart_depuis_1200.slf
INITIAL TIME SET TO ZERO = NO
ORIGINAL DATE OF TIME = 1900;01;01
ORIGINAL HOUR OF TIME = 00;20;00

If I use the same parameters file and I choose as previous computation file the complete (and huge) selafin file the simulation time works fine, but pre-processing is very (too) long.

Any other ideas ?

Thanks in advance for your help !

Regards

Michaël