add new function to fem module

[bernd]

Wow looks great. I like that one. What ever you like. We could have two examples too. It would just be cool if we had at least one example we could calculate the load by a calculator.

A lateral torsional buckling would be cool too. I have spent on this lateral torsional problems of I-Beams more than a year during university ... but long time ago ... more than 20 years. A person who could do a lateral torsional analysis with shell elements could make great papers at that time. Today with FreeCAD we can do this in just a few minutes ...

[twangrt]

Creating an example took more effort than I initially thought, but very educational. Have a look at the PR https://github.com/FreeCAD/FreeCAD/pull/4379

I´ve made a LTB example with a simple supported steel beam. Had some issues making it "hinged" at both ends so I divided the beam. I think I didnt break any structual laws, but please enlighten me if I did.

Bernd, can you explain what you mean with calculate the load with the calculator? The buckling load factor multiplied with the actual load for the criticial linear buckling load?

Is there any limit for the mesh size? Would be nice to make a more refined mesh. But I see the problem, the file size of the mesh can get quite large.

[Jee-Bee]

probably similar to
https://amesweb.info/Beam/Column-Buckli ... lator.aspx and
https://www.johannes-strommer.com/engli ... f-columns/ ?

[bernd]

probably similar to
https://amesweb.info/Beam/Column-Buckli ... lator.aspx and
https://www.johannes-strommer.com/engli ... f-columns/ ?

yep, the above is based on this AFAIK ... https://en.wikipedia.org/wiki/Euler%27s_critical_load

[bernd]

the lateral torsional buckling moment can be calculated with a calculator as well for some cross sections. See https://www.uni-kassel.de/fb14bau/filea ... ssteel.pdf

[twangrt]

the lateral torsional buckling moment can be calculated with a calculator as well for some cross sections. See https://www.uni-kassel.de/fb14bau/filea ... ssteel.pdf

Yeah, this i know

But to be super clear, you want the examples are going to be "ordinary" structual textbook problems so they can be calculated with an external calculator? But you want to include, somewhere, the critical ltb moment from the calculix calculation(Resulting buckling factor multiplied with the moment due to load in model).

Did some checking with another fea program with shells and my committed example and it seem I broke a structual law. It performs a bit worse than an simple spanned beam. I think its due to no rotational resistance in midspan in my example. The beam is a steel H-profile
so its rotational stiffness is quite low, so the difference is quite small, but its wrong.

Ill try to fix this next, and add the unit tests to the example.

[bernd]

But you want to include, somewhere, the critical ltb moment from the calculix calculation(Resulting buckling factor multiplied with the moment due to load in model).

If that would work it would proof the calculix buckling factor It is difficault to add this to the code. May be post the calculation here in the forum, or post a result from another shell analysis as you did here and add a link to the code like this: https://github.com/FreeCAD/FreeCAD/blob ... tie.py#L33

[bernd]

BTW: great to see you back. Just realised you have been used FreeCAD FEM 2.5 years ago already.

[twangrt]

BTW: great to see you back. Just realised you have been used FreeCAD FEM 2.5 years ago already.

Thanks, feels great to be back! Even tough I have been lurking in the background for quite a while

Update: I have created a manual calculation of the critical moment in the example that uses the input from the created geometry. Due to meshing issues with fillets I have based the calculation on a welded I-beam. The manual calculation outputs the result as a text-file in FreeCAD.

The problem I am having right now is that the result does not seem to converge to the target buckling factor with a finer mesh.

I have crosschecked this in another FEM software, but with shells, and quadratic mesh, and it should converge to a buckling factor around 1.6. Which is consistent with the manual calculation within 5% margin. In FreeCAD the buckling factor converges to 1.8 with a finer mesh. Tough I am having issues creating finer meshes as well in Ubuntu as well.

Also want to try to create a hex-mesh in GMSH, to see the difference.

I have not got around to making the unit test yet. Sorry for the slow pace.

[bernd]

thanks for keeping us informed. FreeCAD motto is "It's done when it's done!"