Questions about multipart FEM modeling for thermal analysis

About the development of the FEM module/workbench.

Moderator: bernd

Post Reply
mfreitas
Posts: 2
Joined: Wed Jan 18, 2023 12:19 am

Questions about multipart FEM modeling for thermal analysis

Post by mfreitas »

Hi folks,

I'm trying my first FEM modeling for thermal analysis and I'd like to ask for some help and advice. I have some previous experience with Solidworks workflow but I'm starting with FreeCAD now (congratulation to developers btw, very impressive work!)

I believe most of my difficulties are related to proper modeling the "assembly" for FEM simulation. I read the Transient FEM Wiki and seen some youtube tutorials, but this part is still a bit tricky. I don't know a canonical way of doing it and I end up with a lot of trial and error that are not really converging to an useful result. The file is here (don't mind the part shapes, the're just arbitrary):

http://hg.cptibr.org/~miguel/peca-mais-bloco2.FCStd

Image
http://hg.cptibr.org/~miguel/peca-mais-bloco2.png

1) I have succeeded in finishing the mesh and simulation by using "Compound" instead of "Boolean Fragments" (https://wiki.freecadweb.org/Transient_FEM_analysis say that both should work). However the two parts are not considered in contact for heat transfer. Where did it go wrong?

2) When I tried "Boolean Fragments" then I couldn't really select the parts individually for assigning materials. Empirically I have seen that the "Compound" that works for mesh and has assignable materials is shown as "Solids:2, CompSolids:0, Compounds:1". Is this the right shape contents "goal" for this kind of simulation?

3) In solidworks I remember using mechanical constraints for moving parts and have them in contact with each other. For FreeCAD what tutorials I have seen suggested are using third party workbenchs like "A2Plus". However that sort of workbench doesn't seem compatible with what FEM expects for a meshable compound (?). Is there any other resources for parts assembling in standard workbench that I'm overlooking, that might help producing a good "assembly" ("compound") for simulation? any pointers/tutorials are appreciated.

4) Finally, the simulation does produce a sensible result considering a model that has no contact between parts (each reaches a final constant temperature). What I don't understand though is why do I need the ResultPipeline to see these temperature variations in colorscale, if other results (like displacement magnitude) can be displayed just fine using "show result" dialog?

Thanks in advance!

regards,

Miguel
mfreitas
Posts: 2
Joined: Wed Jan 18, 2023 12:19 am

Re: Questions about multipart FEM modeling for thermal analysis

Post by mfreitas »

Quick update: attached alternative version of the same simulation but using BooleanFragments (Mode=CompSolid) and CompoundFilter. Mesh looks better (doesn't seem like separate solids).

This version, however, fails to simulate with the following errors:

*ERROR in calinput: no material was assigned
to element 1155
*ERROR in calinput: no material was assigned
to element 1172
*ERROR in calinput: no material was assigned
to element 1191
*ERROR in calinput: no material was assigned
to element 1205
(... many more ...)

highlighting some of those manually doesn't help much. Supposedly they belong to a solid that has been assigned material.

Gui.ActiveDocument.FEMMeshGmsh.HighlightedNodes = [1155,1172,1191,1205]
Attachments
peca-mais-bloco4.FCStd
(192.54 KiB) Downloaded 34 times
User avatar
NewJoker
Veteran
Posts: 3018
Joined: Sun Oct 11, 2020 7:49 pm

Re: Questions about multipart FEM modeling for thermal analysis

Post by NewJoker »

Regardless of the approach you use, if the mesh on both parts is not continuous (which can usually be seen by just looking at it, especially if there’s a difference in mesh density of each part) then tie constraint or thermal contact (contact with gap conduction) has to be applied. In FreeCAD, this is normally needed only when there’s a gap between the two parts, even very small.
Post Reply