[Gmsh] Bug with Plug-in Crack? Elements on a surface are not separated

Messad Houamdi messad.houamdi at ensta-paristech.fr
Mon Jun 26 10:22:43 CEST 2017


Hello,

I just would like to report some bugs on the plug-in Crack.

I have to use this plug-in on a surface that represents a crack inside a 3D mesh generated with Gmsh. I've defined the 2D Physical Surface to duplicate and the Physical Line for the "OpenBoundaryPhysicalGroup" and, when I take a look at the .inp that is then used in Abaqus, I do have two surfaces instead of one, so the Plug-in seems to have worked. But then, I have to use Abaqus to apply a load on the mesh to check that the crack opens and I always have isolated elements that don't separate or part of the boundary line that won't open. The only way I've found to solve this issue was to make small (sometimes random) translations in Gmsh on some of the crack's geometrical points to make these elements separate, but it is done manually and the problem appears again each time that I change my cracks or my volumes. 

I'm joining a printscreen of what I get once I use Abaqus on the Gmsh cracked mesh so you can take a look, you can see that the elements "stuck" are really random.

First I thought it was due to the geometry of my cracks but this issue also appears for perfectly straight cracks.

Another small bug with the plug-in that I have is that it works (with the issue mentionned before) when I use the 2.x versions of Gmsh (I'm using 2.9) but, whenever I tried to use it with the version 3.2, Gmsh tells me that my Physical Surfaces and Lines to separate are empty.

Regards,
Messad
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Pb_gmsh_plug_in.PNG
Type: image/png
Size: 211377 bytes
Desc: not available
URL: <http://onelab.info/pipermail/gmsh/attachments/20170626/a2b87fff/attachment-0002.PNG>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Pb_gmsh_plug_in2.PNG
Type: image/png
Size: 401777 bytes
Desc: not available
URL: <http://onelab.info/pipermail/gmsh/attachments/20170626/a2b87fff/attachment-0003.PNG>


More information about the gmsh mailing list