open TELEMAC-MASCARET The mathematically superior suite of solvers

Dear All,

Can some one take a look at my file for time varying boundary condition (attached) and let me know why is this not working, I have got time varying stage at boundary-1 and time varying discahrge at boundary 2 and 3.

I am using this with Telemac3D in parallel mode with 4 processors.

The same file seems to be working just fine for Telemac2D in parallel mode.

The other thing that I want to point out is that the manual for Telemac-3D on the website is probably an old version corresponding to V5P8.

I say so because the keyword for time varying boundary is given as "FILE FOR LIQUID BOUNDARIES" for Telemac-3d but what actually doesn't give error is "LIQUID BOUNDARIES FILE" which is same as Telemac-2D.

I would really appreciate if some can give me some idea why time varying discharge is not recognized by Telemac3d.

Best regards,
Sumit

Hello,

I can't tell with this file only, but a hint would be that you need to have the keywords :

PRESCRIBED ELEVATIONS

and

PRESCRIBED DISCHARGES

given in your steering file, it is what triggers the choice between the different possibilities to enter boundary conditions.

You will tell me that it is not the case in Telemac-2D and you are right, there is a difference here. In fact what is done in Telemac-3D normally corresponds to the documentation, and I could do the same implementation as Telemac-2D (however the liquid boundaries file may not contain all the data and the missing data will be given by the keywords, so preventing user mistakes is not so easy).

I hope this helps,

Jean-Michel Hervouet

Hello Jean-Michel,

Thanks a lot for your kind help, it is indeed working, I was sure it was some thing minor on my part becasue the same file gave me great results with T2D.

Finally I am also attaching my steering file for T3D simulation, my case is having some 21000 nodes and 35000 elements in 2D and when I go to 3D I use 9 horizontal planes.

Would you have any more suggestion how can I increase my speed up in 3D by looking at the steering file. Is the speed up also dependent on the physical time duration that one is trying to simulate.

Thanks a lot for all your help.
Sumit

My steering file

Hello,

I see nothing obvious to improve in your case, however you could try to change :

PRECONDITIONING FOR PPE = 7 (this is Crout preconditioning, which is not the same in scalar mode and in parallel)

to 2 (simple diagonal preconditioning)

or 34 (exact solution on the vertical, sometimes very efficient, especially if you have large distortions between horizontal and vertical).

With best regards,

Jean-Michel Hervouet

Hello Jean-Michel

Thanks a lot for all your help.

Best regards,
Sumit