[general] about the future of the FEM workbench

About the development of the FEM module/workbench.

Moderator: bernd

User avatar
uwestoehr
Veteran
Posts: 4961
Joined: Sun Jan 27, 2019 3:21 am
Location: Germany
Contact:

Re: [general] about the future of the FEM workbench

Post by uwestoehr »

Jee-Bee wrote: Wed Aug 17, 2022 8:19 pm
uwestoehr wrote: Wed Aug 17, 2022 6:10 pm Does not happen here. I just have to remesh to update the mesh t the new geometry.
I'm kind of curious is that only because of ccx vs elmer? because i have the same experience as thschrader. If i can do things better...
1. I took the Tutorial 1 example STEP that comes with Elmer "pump_carter_sup.stp", imported it to FC. There I closed one of the holes by adding an additive cylinder:
ElmerStep.FCStd
(124.25 KiB) Downloaded 30 times
2. Then i exported it to a STEP using STEP format 203 (maybe 214 works, but did not test.
For Elmer it is important that the STEP is a signle body, therefore I made the Fusion in FreeCAD.

3. I start Elmer and load the last project
4. in Elmer I go to menu Open and load the changed STEP

result: Elmer automatically remeshes it with its basic settings:
ElmerGUI_6Ga9qKFCoG.png
ElmerGUI_6Ga9qKFCoG.png (77.44 KiB) Viewed 1680 times

So you can see that the analysis settings are kept.
You have to reset the faces on which forces act because the new mesh will have another shape. (Toponaming is a general issue one cannot easily resolve ;) )
Resetting all faces can become time consuming and that's why I wrote for optimizing geometries, FreeCAD is better. For a single model that you just analyze you are quicker with ElmerGUI directly. However, since our Elmer support is still basic, you often have no other chance than to use ElmerGUI.
User avatar
johnwang
Veteran
Posts: 1345
Joined: Sun Jan 27, 2019 12:41 am

Re: [general] about the future of the FEM workbench

Post by johnwang »

uwestoehr wrote: Wed Aug 17, 2022 5:24 pm Where do you see this?
That was done long time ago. I think ElmerGUI is not under active development. Most people who have experience in FEA research coding, he will not have more energy to do GUI coding.
hfc series CAE workbenches for FreeCAD (hfcNastran95, hfcMystran, hfcFrame3DD, hfcSU2 and more)
User avatar
HarryvL
Veteran
Posts: 1283
Joined: Sat Jan 06, 2018 7:38 pm
Location: Netherlands

Re: [general] about the future of the FEM workbench

Post by HarryvL »

uwestoehr wrote: Wed Aug 17, 2022 5:44 pm While I have you here, can you please reply in this thread since I made a PR just for you and need your input:
https://forum.freecadweb.org/viewtopic. ... 66#p612366
Sorry for upsetting you. This PR was to fix a two-year old regression that several people complained about. I will have a look at your roll-back of the fix when I have access to a computer. What is a few weeks on a two year regression?
uwestoehr wrote: Wed Aug 17, 2022 5:44 pm … I think it doesn't matter in what field you are. The task is usually to improve an existing layout or to create something new.
I have spent my career analyzing reserve strength / collapse / safety of existing structures, tunnels, dams, etc. This was always in your category 1. This is not about improving or creation. Admittedly, I may be the only FC user who is interested in this. The main reason I developed my own solver is that none of the FOSS packages can analyze beyond full (plastic) collapse. CCX just fails to converge near the peak. Elmer (as I read before) doesn’t even support plasticity.
uwestoehr wrote: Wed Aug 17, 2022 5:44 pm Then check it out. FreeCAD 0.20.1 + my patch collection: https://forum.freecadweb.org/viewtopic.php?f=18&t=71058
supports almost all stress features Elmer provides:
https://wiki.freecadweb.org/FEM_EquationElasticity
As I indicate above, it’s not just about producing stresses. In addition, as I demonstrated in several old posts, FEM elasticity raises a lot of debate about stress concentrations. Ever finer meshes give ever higher stresses.You really need plasticity to deal with this. The new acceptance criterion then becomes plastic strain (as I showed before). Again, none of the packages I know of have a plastic strain acceptance check. This, however, is a really easy add-on to FC I am planning to work on. IMHO, running CCX with plasticity + a post-processing check on plastic strain is the ideal answer for structural design. I don’t see how Elmer enters this workflow.
User avatar
NewJoker
Veteran
Posts: 3014
Joined: Sun Oct 11, 2020 7:49 pm

Re: [general] about the future of the FEM workbench

Post by NewJoker »

HarryvL wrote: Thu Aug 18, 2022 2:37 am The main reason I developed my own solver is that none of the FOSS packages can analyze beyond full (plastic) collapse. CCX just fails to converge near the peak. Elmer (as I read before) doesn’t even support plasticity.
Have you tried fixing those convergence issues in CalculiX ? Or are they unavoidable in your case ?

What about code_aster ? Maybe it would let you achieve more.
Jee-Bee
Veteran
Posts: 2566
Joined: Tue Jun 16, 2015 10:32 am
Location: Netherlands

Re: [general] about the future of the FEM workbench

Post by Jee-Bee »

NewJoker wrote: Wed Aug 17, 2022 9:18 pm Right, when it comes to non-mechanical analysis types (acoustic, electromagnetic, fluid flow). But it's the other way around when it comes to capabilities in mechanical analyses. As it was already stated in this thread, Elmer doesn't even support standard contact and plasticity which makes it useless in many scenarios.
What i meant was about the implementation in FC.

uwestoehr wrote: Thu Aug 18, 2022 1:13 am 4. in Elmer I go to menu Open and load the changed STEP
Ah that is the key. i thought(maybe more hoped) you was talking about in FC...
User avatar
HarryvL
Veteran
Posts: 1283
Joined: Sat Jan 06, 2018 7:38 pm
Location: Netherlands

Re: [general] about the future of the FEM workbench

Post by HarryvL »

NewJoker wrote: Thu Aug 18, 2022 7:12 am Have you tried fixing those convergence issues in CalculiX ? Or are they unavoidable in your case ?

What about code_aster ? Maybe it would let you achieve more.
I approached Guido Dhondt 4 years ago and asked him to implement an arc length control routine. He said it was on his TODO list, but it is still there. Without this you cannot overcome limit points. I tried code aster 6 years ago and gave up. I couldn’t get correct answers for the most basic problems. My code is quite simple and does the job.
User avatar
NewJoker
Veteran
Posts: 3014
Joined: Sun Oct 11, 2020 7:49 pm

Re: [general] about the future of the FEM workbench

Post by NewJoker »

HarryvL wrote: Thu Aug 18, 2022 8:49 am I approached Guido Dhondt 4 years ago and asked him to implement an arc length control routine. He said it was on his TODO list, but it is still there. Without this you cannot overcome limit points.
Right, Riks solution procedure is one of the most missing and commonly requested features in CalculiX. I've heard that it was planned for October 2020 but ccx 2.20 came out recently and still no sign of arc length solver. It's still on the preview list: http://dhondt.de/pre_calc.htm
Hopefully, it will be added in the near future.

Maybe the explicit solver (quasi-static analysis) would help you overcome those large nonlinearities. I don't know how well it's implemented in CalculiX but in Abaqus, it could be the way to go.

HarryvL wrote: Thu Aug 18, 2022 8:49 am I tried code aster 6 years ago and gave up. I couldn’t get correct answers for the most basic problems. My code is quite simple and does the job.
Code_aster has an arc length solver and seems to be quite powerful but it's very unintuitive and likely requires a lot of work to properly set advanced analyses.
User avatar
HarryvL
Veteran
Posts: 1283
Joined: Sat Jan 06, 2018 7:38 pm
Location: Netherlands

Re: [general] about the future of the FEM workbench

Post by HarryvL »

NewJoker wrote: Thu Aug 18, 2022 9:08 am … still no sign of arc length solver. It's still on the preview list: http://dhondt.de/pre_calc.htm
I decided not to wait and wrote my own solver with arc length control. It is robust and never fails to capture limit points.

Anyway, the only reason for bringing this up was to show that there are use cases of FC FEM that go beyond flow analysis and/or elastic design.
thschrader
Veteran
Posts: 3129
Joined: Sat May 20, 2017 12:06 pm
Location: Germany

Re: [general] about the future of the FEM workbench

Post by thschrader »

HarryvL wrote: Thu Aug 18, 2022 2:37 am ...
As I indicate above, it’s not just about producing stresses. In addition, as I demonstrated in several old posts, FEM elasticity raises a lot of debate about stress concentrations. Ever finer meshes give ever higher stresses.You really need plasticity to deal with this. The new acceptance criterion then becomes plastic strain (as I showed before). Again, none of the packages I know of have a plastic strain acceptance check. This, however, is a really easy add-on to FC I am planning to work on. IMHO, running CCX with plasticity + a post-processing check on plastic strain is the ideal answer for structural design. I don’t see how Elmer enters this workflow.
This exactly what Ideastatica does.
https://www.ideastatica.com/
You can run 2 simulations:
1. Nonlinear stress/strain with given load
2. Design restistance analysis
With option 2 the solver triggers up the load until the max accepted plastic strain is reached, here 5%.
See pdf example. What amazes me is that you can plot the resistance-curve, load vs plastic strain.
The major backdraw of ideastatica is the "CAD"-funcionality. For more complex parts
you need hours...FC is 10-times faster.
ideastatica_report.pdf
(313.4 KiB) Downloaded 34 times
flange_resistance.JPG
flange_resistance.JPG (27.48 KiB) Viewed 1469 times

@uwestoehr:
Please see this example, plastic analysis of a frame.
frame_plastic.FCStd
(13.13 KiB) Downloaded 27 times
In this case the plastic cutoff criterion (lets say 5% strain) wont work,
the system collapses at load-factor 21 (3 flow hinges ==> collapse), but plastic strain is below 1%.
I will adapt this for elmer tomorrow. For a case like this I can check the results with my
commercial statics software.

Your 2D-cfd model:
How can you compute forces on the cylinder at the inlet (integration of the pressure)?
In cfdof-wb there is a tool, and you can use paraview with various filters.
However
Schönen Feierabend
User avatar
jnxd
Posts: 951
Joined: Mon Mar 30, 2015 2:30 pm
Contact:

Re: [general] about the future of the FEM workbench

Post by jnxd »

I haven't gone through the whole thread, but it appears Elmer vs. Calculix appears to be a major theme. Having worked (and more importantly seen better people work) with Abaqus (with which ccx has commonalities) in the past here's my two cents.

Firstly, Abaqus has the attention of many solid mechanics researchers because of its accuracy with relatively coarse mesh. Apparently it uses some magic behind the scenes even with linear elements (something related to making them higher order p-type and giving solutions at mesh vertices).

There are also the UMAT (user defined material) routines that researchers have developed over, sometimes, decades. Translating them to any other software will take up significant effort. I suppose if someone wants to use them in an open source alternative calculix is the way to go. Disclaimer: not sure if it is actually supported, but at the very least it is a reason why Abaqus continues to be important.

Lastly, at least in research fields I have seen people directly working with INP files rather than through GUI. Sometimes even the official release of Abaqus does not have the options desired in their GUI (I suppose its implemented later). I believe there was an option to add a few lines to input file first before sending it off to the solver.
My latest (or last) project: B-spline Construction Project.
Post Reply