open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello,

I'm trying to run a TELEMAC2D simulation using control sections to monitor the xsectional dicharge. I am running parallel mode (1 node with 12 processors) on a Sci. Linux cluster at the university. The problem is, the simulation gets stuck after the first time step. If I run without the control sections, the simulation run without a problem.

As suggested by jmhervouet in post #18215, I have checked and made sure there is no obstacle between the control section points.

Here are the summary of my simulation
*I'm trying to simulate a river flow at a specified discharge
*mesh: 0.5m approximately equilateral triangles, ~4.5 million nodes
*time step: 0.1s
*initial condition: constant (free surface) elevation
*boundary condition: [boundary#1 - rating curve (outlet)], [boundary#2 - prescribed inflow (inlet)]
*the river flows from northeast to southwest

I have attached the steering file, output file, partel log and external control section file for your review. I will provide the geometry/boundary condition file as well but it is rather heavy so they will have to be emailed.

Any help is greatly appreciated.

Best regards,
Sho

Hello Sho,

Can you upload the sections input file to have a look?

Regards,
Costas

Hi Costas,

Here is the section input file. Thanks for looking into the issue.

Best regards,
Sho

Hi Sho,

I remember encountering some trouble too when using nodes to define sections. Using node coordinates instead, solved the problem for me and was more straightforward to define.

Can you try that and get back? Just replace nodes with coordinate pairs, and remember to set -1 in the 2nd line to a positive number.

Regards,
Costas

Hi Costas,

I changed the section input file to coordinate format (section input file attached) in the following format
section_1
x1 y1 x2 y2

and set the 2nd line to "+1", but the problem seems to persist (output file attached).

I have found a way around this though I'm not sure if this is the right thing to do. I changed the parameter

in the flusec_telemac2d.f to NSEMAX=600 and added as FORTRAN FILE to the steering file. My logic here is that there is about 219 nodes across the control section, which exceeded the limit of nodes allowed for computation of cross section flux (NSEMAX=200).

Is this something safe to do?

Best regards,
Sho

Indeed, your error leads to the NSEMAX limit. I don't know the purpose of this limit (to limit memory usage on older systems maybe?), but in my opinion it is safe to increase it a bit.
Costas

Hi Costas,

I appreciate your input.

Best regards,
Sho