How to fix (and apply loads) to nodes of FEM mesh

[tcdoe]

hi!
this is probably a newb question, but I have imported an FEA mesh successfully in .msh (gmsh) format and all I need to do is fix the lower nodes, and displace and/or especially apply loads to the upper part. I"m basically mimicking a tension/compression test. I want to use Elmer but whatever is the fastest FEA, just linear right now. Here is a sshot of the structure, I've added two cubes overlapping the base and upper regions.
So basically I just want to fix some nodes, and load-constrain/apply to other nodes.

I tried searching on this here as a newb but didn't see anything.
Any help super-appreciated.
I really want to run Elmer on this structure.

[Kunda1]

Moved to FEM subforum

[tcdoe]

Surprised no replies. Seems like simplest thing to do, but after quite more searching can't figure out. I seem to be able to make node sets, but can't assign conditions to them.
Can I use python for this?

[UR_]

In opposite to other FEM packages around, FreeCAD FEM uses not the mesh to apply constrains but the initial CAD model.
Please see examples on FreeCAD's start page.

[Jee-Bee]

No it isn't the simplest question. this is related to the FEM workflow
The normal workflow is that you create a model(geometry)
add constraints on the geometry (not on meshes).
create mesh from geometry
add material
send the bunch of info to the solver

Your model is combination of both geometry and meshes...
The combination of you current mesh have to fit with the two meshes of the endplates. and i'm not sure if this is even possible
The ones who maybe know are at there work.

The place where i expect some clues is here: https://github.com/FreeCAD/FreeCAD/blob ... shtools.py

[Jee-Bee]

In opposite to other FEM packages around...

To be ones i never have seen FEM packages who add constraints to meshes. I have only a breef experience with ansys and comsol.
Both work same method as FC
https://www.youtube.com/watch?v=vnpq5zzOS48

[-alex-]

I agree with UR_ you need a cad model to apply forces or boundary conditions. Furthermore multimeshes are not supported.
IMHO below is workflows I would advise to you:
Import your mesh -> convert it to solid (Part workbench, convert to solid) -> refine upper and lower faces (still Part WB) in order to get two single faces (if your upper and lower triangles are coplanar). Otherwise you could simply trim the upper and lower area of the cad model to get single flat faces.


In this way you should be able to apply easily constraints and loads as well. Furthermore your mesh should match with your cad model.

[bernd]

In addition to what was said before ...

Best would be if you would have the geometry as step, import this and use it for applying constraints.

But the question was Is it possible to apply constraints to the mesh and not to the geometry? No it is not possible with the GUI. It might be possible tu use an existing mesh in FreeCAD FEM and use this one for the analysis without having the geometry by Python.

As a start you could do a frequency analysis with ccx solver. For this you do not need constraints. Furthermore the mesh needs mesh groups, the nodes or elements one would like to apply constraints too. With this and Python and some FreeCAD FEM knowledge it should be possible to use an existing mesh without having the geometry. But I would not do this, it is just a to complicated workflow IMHO.

bernd

[tcdoe]

I agree with UR_ you need a cad model to apply forces or boundary conditions. Furthermore multimeshes are not supported.

Thanks! for this I am trying that right now. A step I am 'missing' is converting the mesh to solid.

[-alex-]

[A step I am 'missing' is converting the mesh to solid.

To get the solid:
Select your mesh -> Part WB: Part scroll bar: Create shape from mesh -> you get the solid
Then hide the mesh object
Select one triangle of upper face -> Part WB:Part scroll bar: Refine shape -> you should get an upper single face

I hope this solid will match enough with your initial mesh.msg...
Otherwise you could use the solid directly in the FC analysis then mesh it again with gmsh. I don't know if this final mesh would be exactly the same compare to the first .msg mesh. I assume it depends the geometry. You can try and give us feedback please.