open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi everybody,

I am trying to running a model of more than 400 000 nodes. Edge length is between 0.25 (near hydraulic structures) and 50 m.

When I use time-step over 0.5 s, the listing stays at initial time-step and it seems that computation does not work. With time-step under 0.5s, it works but the listing stops before the end of the calculation time (around 9900s).

So, how can I optimize time calculation and perform a more stable calculation.

You will find the attached files.


https://wetransfer.com/downloads/f4dcb40649e45e7f97ecb9682b54e6c920180905133250/aebaf2c7f508edd566f8a72a5d97183d20180905133250/f47267



Thanks a lot.

Nicolas

Hello

The time step can be managed with several different tools. For finite elements, the efficiency of the solver is the main indicators that can help to choose the time step. The iterative linear solver you are using should converge (i.e. not exceed the maximum number of iterations) and thus, if you see that it converges withn 40 to 50 iterations, this means that your time step is well chosen. If it converges in very few iterations, this means that you could increase it. On the other side, if the solver does not converge, this means that you should decrease it.

You have another option. You can use a varying time step by fixing the Courant number (around 1 for instance) and let Telemac do the choice of the value of time step. This is also the same stratégy for finite volumes.

I hope that this helps

with my best regards
Riadh Ata

Thanks Riadh for your answer.

My computation seems to work within 30-40 iterations, but after a while, the listing stops without error message:




What can be the origin of this problem?

Hello

I can not suggest any explaantions. Can you stop the run and then look at the result fil eand check if there is something that could explain this issue.
Otherwise, use this result file as a restart file and then see what happens.

kind regards

Riadh