Compatibility for mesh and postprocessing in all FEM related wbs

A subforum specific to the development of the OpenFoam-based workbenches ( Cfd https://github.com/qingfengxia/Cfd and CfdOF https://github.com/jaheyns/CfdOF )

Moderator: oliveroxtoby

HoWil
Posts: 838
Joined: Sun Jun 14, 2015 7:31 pm
Location: Austria

Compatibility for mesh and postprocessing in all FEM related wbs

Postby HoWil » Wed Jun 14, 2017 3:42 pm

Hi,

Today I tested the CFDFoam-wb the first time and I really liked playing with it. I saw new, enhanced dialogs for selecting boundary conditions and tested new mesher like cfMesh.... pretty cool new stuff for me.

So I tried to use the mesh created in CFDFoam with a calculix analysis in FEM-wb. First I used a gmsh-object mesh in CFDFoam and copy pasted it to reuse the copy with calculix. After my first success running an analysis I tried the same with cfMesh-object. I also copy-pasted the mesh which creates also a copy of the original geometry and set up the simulation as in the gmsh case. When I now run the simulation I get

Code: Select all

'Missing prerequisit(s); FEM mesh has no volume elements, either define a shell thickness or provide a FEM mesh with volume elements'
I think this is only a small thing and hopefully can be "solved" easily.

What I ask for here is keeping elements for meshing (pre-) and postprocessing FC-wide as standardized, reusable or compatible as possible. From the perspective of a user I would love to reuse a mesh created for openfoam for an analysis in calculix or even Elmer. The same is for sure valid for dialogs. I also write this with some self-interest.... I am one of the mentors for the Elmer-integration and do not want to have our student to reinvent everything :D .

What do you think ... is this possible or are the different workbenches already fragmented to much.

BR,
HoWil
User avatar
bernd
Posts: 8232
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: Compatibility for mesh and postprocessing in all FEM related wbs

Postby bernd » Thu Jun 15, 2017 7:37 pm

Would you post an example FreeCAD file with such cfMesh ? Are your really sure it has volume elements. The foam guys work quite a lot with shell meshes.

It should make no difference where the mesh comes from, either manuall by python, by gmsh, by netgen or by any of the import filter or even this cfMesh. Internally they are all saved in the FreeCAD smesh structure, means SMESH does not even know where the mesh comes from.

Ahh for sure +1 no matter where the mesh comes from it should be usable with any other solver too !


bernd
HoWil
Posts: 838
Joined: Sun Jun 14, 2015 7:31 pm
Location: Austria

Re: Compatibility for mesh and postprocessing in all FEM related wbs

Postby HoWil » Sat Jun 17, 2017 8:10 pm

bernd wrote:
Thu Jun 15, 2017 7:37 pm
Would you post an example FreeCAD file with such cfMesh ?
Did add a CFDFoam analysis next to a calculix-gmsh one. Hope you can open it without CFDFoam-wb.
FemCalculixCantilever3D_cfMesh.fcstd
(108.96 KiB) Downloaded 27 times
Screenshot from 2017-06-17 22-04-19.png
Screenshot from 2017-06-17 22-04-19.png (308.96 KiB) Viewed 951 times
And the same with snappyhexmesh:
FemCalculixCantilever3D_snappy_hex_mesh.fcstd
(85.87 KiB) Downloaded 23 times
Screenshot from 2017-06-17 22-09-42.png
Screenshot from 2017-06-17 22-09-42.png (308.88 KiB) Viewed 951 times
BR,
HoWil
User avatar
bernd
Posts: 8232
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: Compatibility for mesh and postprocessing in all FEM related wbs

Postby bernd » Sun Jun 18, 2017 6:22 pm

as assumed, the meshes are face meshes only. Open one of your documents, activate FEM workbench, select the mesh right mouse click print mesh info. -> no volumes ... see screen. Means only the surface of the cantilever is meshed. What you could do is add a shell thickness of may be 50 mm and you have a hollow section profile ... and get results for this hollow section profile ... see scree attached ...

screen1.jpg
screen1.jpg (179.76 KiB) Viewed 931 times
screen2.jpg
screen2.jpg (131.82 KiB) Viewed 931 times
HoWil
Posts: 838
Joined: Sun Jun 14, 2015 7:31 pm
Location: Austria

Re: Compatibility for mesh and postprocessing in all FEM related wbs

Postby HoWil » Mon Jun 19, 2017 6:25 pm

I see.... you are completely right. But this makes no sense the CFDFoam guys also need volume-meshes!?
oliveroxtoby wrote:A short question
Is this the correct behaviour of cfmesh? Shouldn't it produce volume meshes? Did I install it in a wrong way?
EDIT: cfmesh should work I remeshed the elbow example and it could be solved (3D flow results of a 3d mesh model)!
Here a look at the related mesh data of the solved cfmesh elbow-example listing 0 Volume elements:
Screenshot from 2017-06-19 20-21-22.png
Screenshot from 2017-06-19 20-21-22.png (274.67 KiB) Viewed 907 times
Is only the mesh info wrong for cfmesh and snappyhexmesh?

BR
HoWil
User avatar
oliveroxtoby
Posts: 300
Joined: Fri Dec 23, 2016 9:43 am
Location: South Africa

Re: Compatibility for mesh and postprocessing in all FEM related wbs

Postby oliveroxtoby » Fri Jun 23, 2017 8:58 am

HoWil wrote:
Mon Jun 19, 2017 6:25 pm
I see.... you are completely right. But this makes no sense the CFDFoam guys also need volume-meshes!?
oliveroxtoby wrote:A short question
Is this the correct behaviour of cfmesh? Shouldn't it produce volume meshes? Did I install it in a wrong way?
EDIT: cfmesh should work I remeshed the elbow example and it could be solved (3D flow results of a 3d mesh model)!
Here a look at the related mesh data of the solved cfmesh elbow-example listing 0 Volume elements:
Screenshot from 2017-06-19 20-21-22.png
Is only the mesh info wrong for cfmesh and snappyhexmesh?

BR
HoWil
Sorry for the bad news, but unfortunately the mesh displayed in FreeCAD is only the outer shell of the full mesh. As cfMesh and snappyHexMesh write their results in OpenFOAM format, there is no functional need for us to read the full mesh into FreeCAD, and so the boundary mesh displayed is purely for visualisation. Someone would need to write an importer to read from the OpenFOAM format...
Please provide all the information requested in this post before reporting problems with CfdOF.