Integration of tochnog solver in FreeCAD FEM

Hi,

I connected the solver ‘Tochnog Professional’ to Freecad. See https://www.youtube.com/watch?v=YLyowfy3N7Y for a simple test (or search in YouTube for ‘Tochnog Professional Freecad’). The video shows how to replace the Calculix executable with Tochnog.

Please realise the following:

  • Tochnog Professional is commercial (dongle locked)
  • As trial it can be used freely up to 100 elements with all available options (plasticity laws, groundwater flow, etc).
  • For the trial version no dongle is needed
  • To skip license checking (which takes time) use ‘tochnog_trial -yes’ in the Tochnog input file (eg in tochnog.dat) !
  • Take care that the tochnog executable in the src directory of the distribution can be found by your operating system (adjust the PATH environment symbol, or put the executable somewhere where the OS can find it)
  • Tochnog runs on Linux and MS Windows, both on 32 bit and 64 bit
  • You can download from http://www.feat.nl
  • Tochnog is mostly used in geotechnics (things in the soil: landslides, driving piles, slope stability, etc.) or civil engineering (masonry, earthquake response, etc).
  • Tochnog reads the abaqus inp file (anyway most of the bulk data like elements, nodes, etc; so not complete inp files)
  • Tochnog writes at output a frd file.
  • Reading the inp file and writing the frd file allows easy replacing Calculix by Tochnog (see the video)
  • All other options and background theory can be found in the Users manual.

Dennis

nice to see it supported.

BTW max displacement of the example beam should be 87 mm not 174 mm …

:smiley:
…and the solver needs 23 sec to calculate instead of 1 sec.
runtime.JPG

Seems to be a commercial site.
https://www.feat.nl
I downloaded the solver and did a run with the beam-example,
but I got an error.
solverTochnog.JPG

bernd: thanks, I missed that. Still found an error in conversion of the inp file, I will correct.

thschrader time_current=1: that is the end time in the calculation (not the cpu time).

thschrader error on running: the trial is limited to 100 elements, so that’s why you get the error, I still need find a better way in freecad to catch that error, and give a clear message. Any idea?

thschrader commercial: yes it is commercial (like you see when you download directly above the download, and other license info on our side). But since it has some other options (material laws etc), it still may be of interest to people using up to 100 elements. I will adjust that I my first post above, to prevent people not interested in that from looking further.

thanks, other ideas welcome.

No Sir, I have no idea.
In the startup.pdf of the unpacked tochnog.zip is the info, that you need a dongle
to activate the program. Maybe there is an additional ccx on the dongle
(or something like a key) which runs better. They take a bunch of freeware-products
an sell it as a commercial solution… :wink:
regards Thomas
unpacked.JPG

@thschrader:

In the startup.pdf of the unpacked tochnog.zip is the info, that you need a dongle
to activate the program.
Correct, if you want to use more than 100 elements. Up to 100 elements no dongle is needed.

Maybe there is an additional ccx on the dongle(or something like a key) which runs better.
No. Like I say above: up to 100 elements the executable can be directly used.

They take a bunch of freeware-products an sell it as a commercial solution..
?? We don’t sell freeware products. Why do you say that?
If someone wants to use up to 100 elements for free, please do. If not interested, please don’t. I clarified the first post. At some uni’s they use the Tochnog trial with some specific options (up to 100 elements), maybe nice if they can do it i.c.w Freecad.


@Bernd:

If you find this post is not ok (thschrader seems to think so), please feel free to delete it.

The error in reading the inp file is corrected, thanks again

Of course it is ok and thanks for letting us know about it, people here however most of the time strongly prefer and support opensource and cross platform solutions :unamused:

@saso

Of course it is ok and thanks for letting us know about it, people here however most of the time strongly prefer and support opensource and cross platform solutions

Thanks. Yes, of course nice to use opensource. We have made much gpl source available. But making a living is also nice, so this ‘tochnog’ version is commercial (but still a 100 element free version may be nice for students trying ‘strange’ materials laws and so).

Many free- and non-free programs available these days. And they use each other and interface with each other. Not a problem in that I think.

Greetings, Dennis

I’m fine with this. I changed the topic title in your first post to clarify what is this topic about. It would have been helpful if you would have been introduce the technog solver in your first post, Licence, OS on which the runs at. Where to get the solver from? How is the integration done? As far as I can tell by converting the inp file, or does technog uses Abaqus inp as well? ETC. You can still do by Editing the first post and add an EDIT inside.

cheers bernd

added it to the solver supported. https://forum.freecadweb.org/viewtopic.php?f=18&t=26326

Depends on 100 volumes and 100 shells and 100 edges or !00 volumes or 100 shells or 100 edges

You can count them easily by python:

load the Fem 3D example:

obj = App.ActiveDocument.getObject("Box_Mesh")
obj.FemMesh.EdgeCount
obj.FemMesh.VolumeCount
obj.FemMesh.FaceCount



>>> obj.FemMesh.EdgeCount
52L
>>> obj.FemMesh.VolumeCount
79L
>>> obj.FemMesh.FaceCount
96L
>>>

since this is a volume only analysis, only the volumes are written into inp file. Means the cantilever should run.

There are some possibilities. One would be to extend the prechecks: https://github.com/FreeCAD/FreeCAD/blob/f3098d5acbc22480acbb7f27ac2a2b7927ae67c3/src/Mod/Fem/femtools/ccxtools.py#L271-L471 especially here https://github.com/FreeCAD/FreeCAD/blob/f3098d5acbc22480acbb7f27ac2a2b7927ae67c3/src/Mod/Fem/femtools/ccxtools.py#L301-L310

100 are really not many :open_mouth: :mrgreen: and 10000 are a lot of bucks …

@Bernd

since this is a volume only analysis, only the volumes are written into inp file. Means the cantilever should run.
Ah, you are correct, I didn’t see that element numbers start high in the inp file. I will check what is going on.

I changed that opening post. I you miss further info tell me.

Thank you.

I have no problems with commercial software. But before ordering I would prefer to
test the full functionality of the program for a limited time, lets say 30 days.
example:
https://www.dlubal.com/en-US/downloads-and-information/free-trial-versions/download-trial-versions?bitversion=64-bit-version
With only 100 elements it is impossible to validate even simple problems.
Regards Thomas

I have no problems with commercial software.
That you have no problems with commercial software I believe. I had the impression that your posts suggested something else (claiming ‘us selling free software’).


But before ordering I would prefer to
test the full functionality of the program for a limited time, lets say 30 days.
Yes, I understand.

But we do it in another way: we typically define with a potential customer the calculations of interest for them, do the calculations for them, help them understand how it is done, etc. And already that is done as paid project (they give money for this project, and if they are happy they can buy afterwards a license). This is because most of the time we do 'strange, heavy nonlinear calculations, which they cannot do with other programs ’ which needs effort of us and the customer, so this method works best for us.

But, again, the 100 element limited version may be of interest to some.


About that error you get running the example, here comes my guess:

  • your windows does not find the tochnog.exe executable; you need to adjust your PATH env variable or put the tochnog.exe somewhere where windows can find it (I will put an extra remark in the opening post in a moment)
  • if it is not that, you antivirus program may block running?
  • or something else, we need to look better then

thanks, kind, Dennis

If FC integrates this solver (more than just dropping it in over CalculiX) then the UI should transparently note that this solver is ‘limited’

@dennis: the forum has tags for citation of other people. Just use the third icon from left just beside the underline icon. This is much better readable than if you gone use just an “>” character

@dennis2: thanks for adding the additional information to the first post. It is exactly what was missing.

@dennis3: What is your solver most used for? Any customer who uses it in structural engineering or geological fields ?


cheers bernd

Hi,

@bernd: Mostly used for geotechnics (things in soil: landslides, driving piles in soil, slope stability, etc.) and civil (masonry, earthquake response, etc). I will add that to the first post.

@kunda: I agree. But no integration needed at the moment I think, later we will see.

Thanks, Dennis

Very interesting! What do you normaly use for pre- and postprocessing?

Do you use for structural earthquake response analysis shell or volume elements?

Pre:
GID Cimne, GMSH, Netgen, reading (converting) inp files, simple build-in meshing in Tochnog itself, Mecway, PrePoMax, optionally mesh and groundwater pressures from FEFLOW (but Tochnog can itself also calculate groundwater pressures), …

Post:
GID Cimne, GMSH, Tecplot, PrePoMax, Paraview (I may look into development of a baby-Paraview in Freecad once, I do not know now), Mecway, …

No shell elements, only volume (but with special postprocessing to calculate shell moments and forces from volume element nodal forces, that is more accurate than from stresses).

I found the quote icon :smiley: