open TELEMAC-MASCARET The mathematically superior suite of solvers

Hello there,

at the moment I try to refine my mesh, but somehow it always creates over constrained triangles. I tried to rise the amount of attachment points in the area where this triangles occur, but without any success.
What would be the best way to get rid of them?
I attached my mesh in any case.

Thanks in advance.

Post the t3m file here to make it easier to help.

Martin

You can find the file attached, but I needed to change the ending since the forum does not accept .t3m files.

- Clemens

Hello everyone,

I wrote a small python script to find/view the overconstrained triangles easily in BK: export_constrained_point_location.py

Process goes in a few steps in order to generate a .xyz with the nodes coordinates:
1/ run a 1 time-step simulation in TELEMAC with `CHECKING MESH=OK` and save the listing output in a log file:2/ In export_constrained_point_loc.py change mesh file name with your .slf, and run it
3/ Open .xyz created in BlueKenue

The script should be available here: gitlab.com/phmusiedlak/public/-/blob/mai...on.py?ref_type=heads

Hope it helps

How does your new script compare to the "Auto-Adjust Mesh" functionality that currently exists in Blue Kenue?



Kindly,
Alan

Alan,

The Auto-Adjust Mesh function "smooths" the mesh by tweaking node locations. That is, it reshapes the triangles to be closer to the ideal equilateral shape.

It doesn't remove or replace overconstrained triangles.

Just another feature that was "on the list" that never saw the light of day.

NOTE: A lot of the ground work is there.

In BlueKenue the only current option is to manually adjust the overconstrained triangles via use of the "Swap Edge", "Split Edge" and "Split Element" functions.
What remains to be be done is to automatically identify which function or sequence of functions to apply to target triangles.

Hope this helps... Martin

Thanks for the clarification, Martin

Kindly,
Alan

Hello,

I found a way to avoid the first mesh checking step (since it fails sometimes) with the log file by directly loading the .slf with matplotlib and finding the boundaries. Thanks to Ian_Thomas1 from discourse.matplotlib.org/t/boundary-edge...-set-of-points/15397

The (Python) script loads the .slf, and uses mesh x,y coordinates and ikle table. Then, matplotlib neighbors property is used to find segments with just one neighbor.

The process is:
0/ mesh you domain in BK => mesh.slf
1/ run the script (change directory and file name)
2/ open the generated over_constrained_triangles.xyz file in BK
3/ Remove triangles manually

Should help you to find them






The script is called :and is available here : gitlab.com/phmusiedlak/public


Hope it helps

Pierre-Henri

You may also wish to try the new resampling method called "Segment Splitting" that will be in the next alpha created after 26 Jan 2024. This may help to avoid creating over constrained triangles when starting from a 2D polygon.

See also #44017

Kindly,
Alan

Blue Kenue 3.12.12 - alpha is now available for download with this new enhancement.

Kindly,
Alan