Meshing Brep list - Stitch Seams

I am trying to create a FEA model through grasshopper. I create a shape of a structural steel connection as surfaces (breps) and then mesh them. The problem is that mesh algorithm doesn’t always create common points (nodes) where different parts are connected. If I am not mistaken, stitch seams option is supposed to take care of that but it doesn’t work (or at least not as I would like it).

The solution I have found is instead of meshing a list of Breps and then combine the meshes, I should join all Breps to a single Brep and then mesh it. However, joining Breps just doesn’t work consistently. Anyone other suggestions?

Connection Breps.3dm (266.5 KB) Connection test.gh (12.0 KB)

This usually indicates a surfacing/intersection problem and/or a problem with tolerances. You need to make sure that surface trimming tolerance is as strict as, or more strict than the join tolerance. Or, in other words, if you have created trims at a given tolerance, reduce it by a factor of 10 before joining. Or in even other words, set your tolerances very strict and perform the surface trimming. Then reduce it slightly and join trimmed surfaces in Grasshopper.

That sounds like a good point to start my investigation from. Thanks for the tip!

Update: Changing the tolerance in Rhino doesn’t seem to make any difference.

The issue is that you have non manifold edges. Therefore you can’t join all surfaces.

But meshes can have non manifold edges.

As of now, your joint is built from surfaces without thickness. If all the surfaces were closed polysurfaces, you could join them into one BREP.

PS: for future uploads you could make the life of people like me a little easier by internalizing input geometry. That would only require downloading and opening one file.

1 Like

Thank you very much for your reply it was helpful. You are right in terms of non manifold edges, this is the problem with joining BREPs, I didn’t know about that. However, meshing algorithm fails to identify my complex geometry at most of those points. For now, I will say there is no solution to the problem. Get the mesh and manually merge the wrong nodes together in your analysis software.

For anyone else trying to mesh surfaces to get a FEA grid, here are some tips:

  • avoid non manifold edges if possible (having T shapes)
  • join BREPs to one BREP if possible (that will ensure grid continuity)
  • create edges at intersection locations (split the BREP at those locations)
  • do not forget to “clean and combine” the generated meshes
  • Karamba3D mesh widget seems better that the stock one, but not perfect too (it can even identify some intersection points, gives a more uniform mesh, uses only triangular elements so use quadrangulate widget)

If anyone has more tips, let us know!

Maybe the command MatchMeshEdges can help you somehow. I haven’t tried it on your case, but I have used it in the past to get conformal edges on two mesh patches.