Stress Result in a cylinder Shell

[jake77]

Dear all,

Mighty work, thank you all for your efforts.

When modelling a simple cylinder with external pressure on it, I think I got some qualitatively wrong results, namely: when requested to display the max principal stresses (PS1), the stresses shown are the SXX components.
On the attached fig the stresses should be uniform along the circumfence.
Running the same model on calculix directly and displaying the results using cgx the results are as expected. I guess this is just some typo in the way results are imported from the datasets. Could you please confirm this?

[bernd]

Welcome to FreeCAD FEM. I moved your post in a separate topic, since it seams not related to the problems of the other topic.

your zip contains the *.inp file. Would it be possible to post the FreeCAD file and your FreeCAD info.

It seams your file was created with Version 0.17. We are just befor release of 0.18. It would be good if you could check your problem with 0.18 developer release available on github https://github.com/FreeCAD/FreeCAD/releases

[jake77]

Hello @bernd,

Thanks for the prompt answer.
I tried to attach the FreeCAD files but they were too large. What is the limit? How can I send larger files?
I use 0.17 as being a beginner using the stable release seems to be the right choice How stable is the next candidate? Is there some information on the new features that might crash? I use sketcher -> part to create a shell and then the GMSH mesher and finally create the ccx inp. That is then modified (*STATIC -> *BUCKLE) and run in ccx.

Right now I have some specific problems to solve (this includes checking the results against other results), so I'll use FreeCAD as a preprocessor and run analyses in the ccx environment.

[bernd]

- select you FEM mesh in tree view --> clear mesh will clear the mesh but not delete the parmeter.
- purge the results

Your file should than be small ...

[bernd]

How stable is the next candidate?

For FEM very much better than 0.17. There where fixed lots of issues. I do not even have 0.17 installed anymore (would install it only if someone would find a bug we really would need to fix still in 0.17 release, but we did not have this in 0.17 release cycle ... )

[bernd]

(*STATIC -> *BUCKLE)

We should think about adding this buckling stuff to FreeCAD as an analysis type ...

[jake77]

How stable is the next candidate?

For FEM very much better than 0.17. There where fixed lots of issues. I do not even have 0.17 installed anymore (would install it only if someone would find a bug we really would need to fix still in 0.17 release, but we did not have this in 0.17 release cycle ... )

Shall I use the pre-release from April? I'm on Win, dont't think compiling from source is an option.

[jake77]

(*STATIC -> *BUCKLE)

We should think about adding this buckling stuff to FreeCAD as an analysis type ...

Based on the 0.17 version I had quite a list for improvements, but I'm positively surprised on the advances since 0.16; if the improvements in 0.18 are similar this will be quite a good tool.
I can write python stuff and I'm interested to make certain workflows better, but have very limited time. Let's find out how I can help.

[jake77]

How stable is the next candidate?

For FEM very much better than 0.17. There where fixed lots of issues. I do not even have 0.17 installed anymore (would install it only if someone would find a bug we really would need to fix still in 0.17 release, but we did not have this in 0.17 release cycle ... )

Shall I use the pre-release from April? I'm on Win, dont't think compiling from source is an option.

OK I see you have it regularly updated. I'll fire it up then.

[bernd]

OK I see you have it regularly updated. I'll fire it up then.

you can get it here for windows, just unpack and run it https://github.com/FreeCAD/FreeCAD/releases

btw. compiling on windows int not that difficault, since we gone provide precompiled library pack.