Elmer Integration (GSoC) - Activity Log

About the development of the FEM module/workbench.

Moderator: bernd

m42kus
Posts: 159
Joined: Tue Sep 02, 2014 9:41 am

Re: Elmer Integration (GSoC) - Activity Log

Postby m42kus » Wed Jul 24, 2019 1:15 pm

JamesMarine wrote:
Wed Jul 24, 2019 1:04 pm
I assume the major workload is on the solver input file, is that correct? Do you expect any other major workload?
For me most of the work is to figure out how elmer equations I haven't worked with yet work especially together with other equations like flow. The actual implementation is not that much work because I developed a very flexible framework for elmer.
JamesMarine wrote:
Wed Jul 24, 2019 1:04 pm
If multiphysic such as FSI is not included, what is the typical case that Elmer can do but Calculix can not do?

Sorry that I am a beginner and my question may not be quite professional.
Well my goal two years ago was to replicate what calculix can do. The reason I integrated elmer is because of the potential it has as a flexible multiphysics solver.
My GSoC Devlog: ceeli.github.io/gsoc
JamesMarine
Posts: 12
Joined: Tue May 21, 2019 2:07 pm

Re: Elmer Integration (GSoC) - Activity Log

Postby JamesMarine » Wed Jul 24, 2019 1:24 pm

m42kus wrote:
Wed Jul 24, 2019 1:15 pm
JamesMarine wrote:
Wed Jul 24, 2019 1:04 pm
I assume the major workload is on the solver input file, is that correct? Do you expect any other major workload?
For me most of the work is to figure out how elmer equations I haven't worked with yet work especially together with other equations like flow. The actual implementation is not that much work because I developed a very flexible framework for elmer.
JamesMarine wrote:
Wed Jul 24, 2019 1:04 pm
If multiphysic such as FSI is not included, what is the typical case that Elmer can do but Calculix can not do?

Sorry that I am a beginner and my question may not be quite professional.
Well my goal two years ago was to replicate what calculix can do. The reason I integrated elmer is because of the potential it has as a flexible multiphysics solver.
Do you have any plan to work further on the FSI fuctions of Elmer solver? I am not an expert in coding, but if you need, I'll see what can I do to assist.
m42kus
Posts: 159
Joined: Tue Sep 02, 2014 9:41 am

Re: Elmer Integration (GSoC) - Activity Log

Postby m42kus » Wed Jul 24, 2019 5:09 pm

JamesMarine wrote:
Wed Jul 24, 2019 1:24 pm
Do you have any plan to work further on the FSI fuctions of Elmer solver? I am not an expert in coding, but if you need, I'll see what can I do to assist.
If you could sort out the elmer stuff, I'd be willing to do the coding part :) What is your timeline regarding this feature? Do you need it soon?
My GSoC Devlog: ceeli.github.io/gsoc
HoWil
Posts: 1199
Joined: Sun Jun 14, 2015 7:31 pm
Location: Austria

Re: Elmer Integration (GSoC) - Activity Log

Postby HoWil » Fri Jul 26, 2019 6:17 am

JamesMarine wrote:
Wed Jul 24, 2019 1:24 pm
Do you have any plan to work further on the FSI fuctions of Elmer solver? I am not an expert in coding, but if you need, I'll see what can I do to assist.
Hi James, Hi Markus,
I would love to see this evolving and can promise my very limited time budget for testing.
Br,
Howil
JamesMarine
Posts: 12
Joined: Tue May 21, 2019 2:07 pm

Re: Elmer Integration (GSoC) - Activity Log

Postby JamesMarine » Fri Jul 26, 2019 11:37 am

m42kus wrote:
Wed Jul 24, 2019 5:09 pm
JamesMarine wrote:
Wed Jul 24, 2019 1:24 pm
Do you have any plan to work further on the FSI fuctions of Elmer solver? I am not an expert in coding, but if you need, I'll see what can I do to assist.
If you could sort out the elmer stuff, I'd be willing to do the coding part :) What is your timeline regarding this feature? Do you need it soon?
The timeline is in the next month(end of Aug), do you think this is possible to acheive? What kind of work do you need from myside, can you please kindly give some more detailed description? I can start from now.
JamesMarine
Posts: 12
Joined: Tue May 21, 2019 2:07 pm

Re: Elmer Integration (GSoC) - Activity Log

Postby JamesMarine » Fri Jul 26, 2019 11:39 am

HoWil wrote:
Fri Jul 26, 2019 6:17 am
JamesMarine wrote:
Wed Jul 24, 2019 1:24 pm
Do you have any plan to work further on the FSI fuctions of Elmer solver? I am not an expert in coding, but if you need, I'll see what can I do to assist.
Hi James, Hi Markus,
I would love to see this evolving and can promise my very limited time budget for testing.
Br,
Howil
That's great! FreeCAD forum is indeed a great place, we can work together for this and look forward to a better FreeCAD day by day.
HoWil
Posts: 1199
Joined: Sun Jun 14, 2015 7:31 pm
Location: Austria

Re: Elmer Integration (GSoC) - Activity Log

Postby HoWil » Tue Jul 30, 2019 7:28 pm

JamesMarine wrote:
Fri Jul 26, 2019 11:37 am
m42kus wrote:
Wed Jul 24, 2019 5:09 pm
JamesMarine wrote:
Wed Jul 24, 2019 1:24 pm
Do you have any plan to work further on the FSI fuctions of Elmer solver? I am not an expert in coding, but if you need, I'll see what can I do to assist.
If you could sort out the elmer stuff, I'd be willing to do the coding part :) What is your timeline regarding this feature? Do you need it soon?
The timeline is in the next month(end of Aug), do you think this is possible to acheive? What kind of work do you need from myside, can you please kindly give some more detailed description? I can start from now.
To start with it would be good to have a simple and working model, a sort of benchmark which shows all needed elmer-related settings.
There should be a 2D example in the documentation. Since Elmer in Freecad works only for 2D models (due to some mesh related thing I cannot remember) you could start with building a working 3D simulation based on a exported mesh created with FC. In the end we need a working .sif file and e.g. a .unv mesh file to play with.
m42kus
Posts: 159
Joined: Tue Sep 02, 2014 9:41 am

Re: Elmer Integration (GSoC) - Activity Log

Postby m42kus » Wed Jul 31, 2019 11:31 am

JamesMarine wrote:
Fri Jul 26, 2019 11:37 am
The timeline is in the next month(end of Aug), do you think this is possible to acheive? What kind of work do you need from myside, can you please kindly give some more detailed description? I can start from now.
Hey, sorry I totally missed you post :oops:

Yes it should be doable to have something working till the end of August although it will probably not be perfect yet ;) Just like HoWil said first a working 3D example with a commented elmer input file with descriptions of the essential options would be a good start. Most important are options that should be configurable by the user via the FreeCAD GUI. Also if you could play around a little with solver settings to find out a configuration with which the model converges best/fastest - that would be very helpful :)
My GSoC Devlog: ceeli.github.io/gsoc
julieng
Posts: 92
Joined: Sun Nov 25, 2018 8:57 pm

Re: Elmer Integration (GSoC) - Activity Log

Postby julieng » Tue Nov 10, 2020 3:51 pm

Hello,

My problem is that I cannot run Elmer with FreeCAD, some years ago I succeed with the 0.17 FreeCAD version.

I have installed the last version of FreeCAD 0.19 with windows 10.

I have installed the last version of Elmer. I have specified the Elmer path to the .exe in the preference option of FreeCAD
C:/Program Files/Elmer 8.4-Release/bin/ElmerGrid.exe and
C:/Program Files/Elmer 8.4-Release/bin/ElmerSolver.exe

I have done an elastic calculation of a loaded beam with calculix without issues.
I want to add the ElmerSolver -> Elasticity node. I can write and execute the case, the simulation doesn't converge and there is no results in the ElmerResult window.
I don't find on the web a tuto showing how to run an Elmer simulation with FreeCAD, only with calculix.

I share the case, someone can tell me what is wrong please ?

Julien
Attachments
Beam_FC019_Elmer.FCStd
(184.89 KiB) Downloaded 6 times
User avatar
bernd
Posts: 10975
Joined: Sun Sep 08, 2013 8:07 pm
Location: Zürich, Switzerland

Re: Elmer Integration (GSoC) - Activity Log

Postby bernd » Tue Nov 10, 2020 4:20 pm

try playing with equation pref. I had problems with iterativ solvers on simple problems like this. Thus I use direct solver for elmer in most cases. But I do not get any reasonable results either whereas ccx gives resonable results.

Any elmer guy around ... ?