Mesh failure with complex fault geometry

Coreform Cubit Meshing
#1

Hi,

I am quite new to meshing, and am attempting to mesh the geometry in the attached journal file
geom_mesh_test.jou (44.4 KB)
. I believe I have modified the geometry properly to ensure connecting and intersecting fault surfaces are properly aligned and merged with each other and the domain volume, but the meshing still fails (despite providing additional sizing constraints for the surfaces). The meshing lines in the file begin around 860 - I have kept the geometry in a single file to make editing easier, so I apologize for its length. The errors I get from meshing are:

Command: mesh volume all smooth
ERROR: Mesh of Surface 31 has connectivity problems.
(Mesh has odd Euler characteristic 1.)
ERROR: Inconsistent edge to triangle count relationship.
ERROR: Mesh of Surface 31 is invalid.
ERROR: Found edge with nodes 12 45 intersects tri with nodes 36 14 25 near (9273.615892 -8962.483406 -7732.891269)
ERROR: Unable to mesh boundary
ERROR: Volumes 9 meshing unsuccessful using scheme: tetmesh
ERROR: 1 volume did not mesh : Volume 9

Any help would be greatly appreciated. I have been working on this problem off and on for months now!

Thanks,
Cassie

Could Someone Give me Advise with Meshing Complex Geometries in Coreform Cubit?
#2

Hi Cassie,

I recommend a slightly different process to represent the faults. I think the problem lies with the fault that extends through volume 9 from top to bottom. It looks like the boolean subtract creates a combination of manifold and non-manifold topology that the mesher doesn’t handle correctly.

Instead of doing the boolean subtract, I recommend meshing the faults separately and then using the tetmesh respect option as follows.

mesh surf in body with is_sheet 
volume with name "v_domain" tetmesh respect tri in surface in body with is_sheet

I’m attaching a modified journal file. Note that I used the is_sheet attribute rather than selecting specific ids. I prefer to avoid ids whenever possible. Ids may change over versions as the geometry kernel gets updated

Karl.
geom_mesh_test.jou (46.3 KB)

image

#3

Hi Karl,

Thank you so very much! I’ve been spinning my wheels for ages, and having this meshed is a huge relief. Now, I am excited to assign boundary conditions and put it to the test.

Best,
Cassie