Topology optimization

About the development of the FEM module/workbench.

Moderator: bernd

DMST1990
Posts: 7
Joined: Fri Jun 24, 2016 7:43 am

Re: Topology optimization

Postby DMST1990 » Wed Jun 29, 2016 10:27 am

fandaL wrote:Hi, welcome to the FreeCAD forum.
It seems good. I put it to my candidate list on which programs to look. ;)
I like your tool, its easier to use than mine. :)

I have added two pdf- files, which maybe can help you implementing the topology optimization (SIMP, BESO etc...)
and understanding the different methods.

https://github.com/DMST1990/ToOptiX

1. Masterthesis (German) (hard to read, you can just look at the pictures)
2. Introduction (German) (Easy to read, except the equations)

There are some parameter studies which explain the behaviour of the used parameters.
If you need some explenation or some help for implementing sth just ask me.
fandaL
Posts: 369
Joined: Thu Jul 24, 2014 8:29 am

Re: Topology optimization

Postby fandaL » Thu Jun 30, 2016 8:30 pm

Thank you.
I have some basics of German, so I think this will be a good motivation to practise.
It will take me some time to go through.
Samlavisse
Posts: 2
Joined: Thu Aug 18, 2016 9:33 am

Re: Topology optimization

Postby Samlavisse » Thu Aug 18, 2016 10:08 am

Hi, all.

How integrate that on Freecad ?
And how use it to ?
So, I'm not developper. I'm just a beginner user of Freecad (not beginner on CAO).
I have a personnal project on to developpe on CAO for 3D printing and that's the fonction I search for it.
I don't have find the good free tool for me at this time on the web.
(Blender don't have visual interface i can use for the moment : "my eyes bleed" when I attempt to use it, sorry :P
and I don't have test Z88Arion : I'm affraid to don't speak german enough but attempt anyway now).

Could you post a complete tutorial of each step (from install to final STL result), plz ?
(for the better tool as possible on Freecad, ...if it's possible ^^)
I think a lot of users (who won't use crack version of commercial tools and don't have money to paid it) could be really interesting by have that tool directly on Freecad.

Xcuz for my languages errors, that's not my born laguage :P
Thx a lot for your works :D
Sam...
Last edited by jmaustpc on Thu Aug 18, 2016 1:28 pm, edited 1 time in total.
Reason: Removed email address from post
jmaustpc
Posts: 9666
Joined: Tue Jul 26, 2011 6:28 am
Location: Australia

Re: Topology optimization

Postby jmaustpc » Thu Aug 18, 2016 1:32 pm

Samlavisse wrote:Hi, all.
Welcome to FreeCAD.

I deleted your duplicate post and removed your e-mail address. Please read the forum rules in the Help Forum as they will provide you with a lot of useful hints
fandaL
Posts: 369
Joined: Thu Jul 24, 2014 8:29 am

Re: Topology optimization

Postby fandaL » Thu Aug 18, 2016 8:12 pm

Hi Sam,
integration to the FreeCAD with gui is still far (it is probable that Z88Aurora* will be sooner translated to English). Some help how to run it is below

1) create a special folder for your analysis
2) follow tutorial from pdf file https://github.com/fandaL/beso

Currently it is not much comfortable for new users (especially if you are not familiar with python). I run it from external python console, but you can instead of it:
- install python 2.x, associate py files and run it simply by double click from your OS (hard to catch mystakes in input since window closes with error to me)
- it is theoretically possible to run it inside the FreeCAD python console, but changes in the beso_config file might need restart of the FreeCAD. Give this to FreeCAD with your working directory path

Code: Select all

my_working_directory = 'd:\\analysis_folder\\'

import os, sys
os.chdir(my_working_directory)
sys.path.append(os.getcwd())
and start by

Code: Select all

import beso_main
for next runs it will be enough

Code: Select all

reload(beso_main)
Now frd results which you drug and drop from your working directory are rather for inspiration, so when you get “good looking” result you need to model it by some cad and after that export in format for your 3D printer, because resulting mesh is not smooth and I didn't try to export it e.g. to Blender to apply some smoothing filter.

*EDIT: Z88Arion
Last edited by fandaL on Fri Aug 19, 2016 6:51 am, edited 1 time in total.
Samlavisse
Posts: 2
Joined: Thu Aug 18, 2016 9:33 am

Re: Topology optimization

Postby Samlavisse » Thu Aug 18, 2016 10:26 pm

Thx for your answers ^^
User avatar
NormandC
Posts: 18534
Joined: Sat Feb 06, 2010 9:52 pm
Location: Québec, Canada

Re: Topology optimization

Postby NormandC » Fri Aug 19, 2016 2:27 am

Bonjour Sam,
Samlavisse wrote:(not beginner on CAO)
CAO (conception assistée par ordinateur) means nothing in English, at least nothing that would be relevant here.

In English, the acronym is CAD (computer-aided design), as in... FreeCAD. ;)
fandaL
Posts: 369
Joined: Thu Jul 24, 2014 8:29 am

Re: Topology optimization

Postby fandaL » Thu Sep 08, 2016 1:09 pm

Support for other element types passed my testing - simple 3D model and 2D (different one) model with edge load. It should be possible use all element types in optimization domain except beam elements. Beam elements can be in the model but not in optimization domain.
What was clearly visible was computation time - for 2nd order elements 4-8 times longer than for 1st order elements for same number of elements. I didn't focus on differences in quality, but 2D models gave different topology for different element types.
results_with_tetra10.png
results_with_tetra10.png (31.9 KiB) Viewed 2518 times
results_with_hexa8.png
results_with_hexa8.png (26.34 KiB) Viewed 2518 times
results_with_quad4.png
results_with_quad4.png (7.46 KiB) Viewed 2518 times
results_with_tria3.png
results_with_tria3.png (16.71 KiB) Viewed 2518 times
results_with_tria6.png
results_with_tria6.png (16.75 KiB) Viewed 2518 times
In these cases all was one optimization domain, so edge load on the shell model is "in the air" after process.
User avatar
bernd
Posts: 10098
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: Topology optimization

Postby bernd » Thu Sep 08, 2016 3:49 pm

Very interesting that different elements give different results, even if you use hundreds of them. Would be interesting what the reason for this is. Mhh neiter nature nor materials does not know femelements. Which one is best ?! It is may be like a path. If the optimizer is on a specific path it just goes this way. There must be academic papers about this.

Mhh and tria3 give a better result than tria6 and quad4 (IMHO) really strange ...
User avatar
PrzemoF
Posts: 2983
Joined: Fri Jul 25, 2014 4:52 pm
Contact:

Re: Topology optimization

Postby PrzemoF » Thu Sep 08, 2016 4:26 pm

That looks very, very interesting!