Posted Wed, 17 Jun 2020 10:47:56 GMT by Guillem Buxo-Escapa

I have currently built a 3D geological model using Leapfrog Geo 5.0. I aim to create an unstructured 3D mesh in Feflow using the GM exported from Leapfrog (*.lfm), which consists of 6 volumes corresponding to the 6 diferent lithologies.

When I try to generate the unstructured mesh using "Supermesh import from maps", if I import the complete model (as 6 volumes-contact surfaces) , I get the following error from TetGen:  "[b]Tetgen meshing error: A self intersection was detected[/b]". If I import just one of the volumes (from a separate *.lfm file) the error disappears and my unstructured mesh is generated, obviously just for the volume that I have imported.

How can I adress it in order to generate a mesh for all my domain, including the 6 lithologies?

Also, once the mesh is generated, is there an easy way to select the elements that correspond to one lithology? I am thinking of groups of elements in the selection panel generated automatically from de *.lfm input file.

Thank you,
Posted Wed, 17 Jun 2020 11:32:46 GMT by Peter Schätzl Grundwassermodellierer
With respect to the selection, you can set marker points for TetGen (one for each lithology, just anywhere within the volume. I've never done the transition from Leapfrog to FEFLOW unstructured myself, but I can imagine that Leapfrog is not able to export the lithologies as non-manifold geometry, i. e. without overlaps and intersections. This, however, is a condition for the TetGen algorithm to work (FEFLOW only does very simple corrections like dropping duplicate triangles before handing over to TetGen).
Posted Fri, 03 Jul 2020 09:26:31 GMT by Guillem Buxo-Escapa
Thanks Peter,

I have solved the problem with the intersections now and I managed to generate the fully unstructured mesh directly from the Leapfrog Geological Model. The problem I have now is that the mesh consists of about 7 Million of elements, of wich 4.5 Million have a volume of 10 m3 or less (and 4 M, less than 1 m3).

To fulfill my purpose, I do not need such a fine resolution (except for some specific areas around wells, 10^6 m3 would be generally OK). Is there any option to generate the mesh with more flexibility so I can control the minimum volume of the elements?

Anyone understands why do I get so many, and so small, elements?

Thank you

Posted Fri, 03 Jul 2020 10:07:50 GMT by Peter Schätzl Grundwassermodellierer
As I said, I'm not very familiar with the Leapfrog - FEFLOW interface, but one option could be to re-generate the mesh for the different volumes in FEFLOW using TetGen. Whether this helps to reduce the number of elements significantly will, however, depend on the geometry of the volume-bounding interfaces which cannot be changed in FEFLOW.
Posted Fri, 11 Jun 2021 10:13:06 GMT by Fabio Canova Specialista Studi Ambientali
Hi Guillem,

i am tryng to built a 3d fully unstructured model starting from a Leapfrog GM as well. Unfortunately, when I try to generate the grid from the whole GM I got the very same error you had. How did you solved the problem? What procedure did you use in the end and was the fully unstructured model working smoothly and was it handy?

Thank you

Kindest regards

You must be signed in to post in this forum.