not working as expected

[pironman]

I'm usind this code with the module FEMbyGEN on freecad.
I try to optimize a motorcycle footrest, and this is the results:

so looks like the best shape is two plates. I think this is strange, so I made a simulation with 2 plates, and a simulation with 2 pipes, as must be correct:

using pipes the deformation is way lower, so stiffness way higher, and the weight is the same.
So I think something on this code don't work as expected

[pironman]

I add more information.
I try to optimize this shape:

and this is the result:

Basically, if I made the FEM, and with filter I remove the parts with low Von Mises stress, I obtain the same result:

So I'm sure the optimization is not working for some reason.

I'm using FreeCad 1.0.0 on Windows 11

[fandaL]

Can you upload the log file from the working directory to see detailed settings?

[pironman]

FEMMeshGmsh.log.zip

do you mean this file?

thank you

[fandaL]

Yes. Parameters seems analogical to what I have when using beso_fc_gui except domain_material values. In the input settings in beso_conf.py , there should be an input like this:

domain_density[elset_name] = [7.9e-15, 7.9e-9]
domain_material[elset_name] = ['*ELASTIC\n0.21, 0.3\n*DENSITY\n7.9e-15\n',
                               '*ELASTIC\n210000, 0.3\n*DENSITY\n7.9e-9\n']

Which corresponds to the "void" material with E = 0.21 MPa, nu = 0.3 and density 7.9e-15 t/mm^3
and "solid" material with E = 210000 MPa, nu = 0.3 and density 7.9e-9 t/mm^3.
The python syntax in the brackets is later used to write material inputs to the CalculiX files.

Of course you can change material values, just keep "void" material (first in the bracket) with much smaller elastic modulus and density (e.g. million times).

[pironman]

ok, so I think this is the problem: I've installed FEMbyGEN using the addon manager inside freecad, and there isn't the file beso_conf.py.
These are the files inside the addon folder:

[fandaL]

I use beso code from this repository, but FEMbyGEN uses modified code. The content of domain_material variable seems to be intentionally different, which means that the issue might be in something completely different.