Thanks for this information.
I don’t have much experience in Freecad but we need to have shell mesh if we want to run some crash analysis with Radioss.
Is it possible with Freecad to generate shell mesh ?
Hey so, I’ve got professional contacts with the radioss Altair team and they’re very good sports.
I personally think that if we pushed to get Radioss supported in FreeCAD they’d be completely receptive to requests for aid.
On another note, I’m also struggling to compile it at the moment, or to simply even run a test case.
Consider this a thread bump saying that y’all ain’t alone in your struggles for now.
That would be great. Open Radioss needs a GUI and FreeCAD can be the best choice for it. However, we have to clarify the issue with no Windows version first.
Okay, I’ve got it working but it wasn’t particularly easy.
It seems like the default mode of output that they’re going for in the future is going to be their h3d file, which is a proprietary altair file format.
There’s another output file called a .sty file that can also be requested with the second runfile (the engine run) and I’m testing it with the birdstrike model now.
This solver is faster than anything I’ve ever seen before. It’s incredible.
For those who haven’t seen it yet, they replied with some good news about Windows implementation:
We are working on this topic.
As a short term workaround, you can use the WSL (e.g. Ubuntu, which works well on my windows10 machine).
It’s desribed in detail in the tutorial: “Getting Started With OpenRadioss and GitHub - Tutorial” on the OpenRadioss confluence page.
I wonder how it compares with LS-Dyna or Abaqus Explicit. I use the latter frequently, often for impact analyses.
If I am not wrong, here is the story of Radioss (this is approximate, I do not say that it happened exactly like that, but it is close to) :
In the 70s, Dyna3D (dynamic explicit FEM code co-developped by Berkeley Univertsity and Lawrence Livermore National Laboratory) was (or has been) open source (do not know which licence term, see below what happends). A french company (E.S.I.) has used this code to released a commercial software called PAM-Crash in the 80s. Then, few years latter (mid-end 80s), a worker from E.S.I. deciced to leave and created another french company called MECALOG (since then bougth by Altair) and also to released a commercial software called Radioss (probably based on PAM-Crash). At the end of the 80s, one of the creator of Dyna3D (John Hallquist) created a company (LSTC) to released a commercial software called LS-Dyna (based on Dyna3D).
All in one → LS-Dyna (LSTC), Radioss (Altair) and PAM-Crash (E.S.I.) are all based on Dyna3D and probably very close each other, then, to answer your question : Radioss is probably as fast as LS-Dyna. But, is OpenRadioss is exactly the same solver than Radioss … ?
Anyway, Radioss is a really solid software largely used in competitive industrial domain, this is a great news for engineers, and for open-source eco-system !
basically it’s unix’s cmd, I know the git installer on windows comes with a bash shell so it’s certainly possible to have one on windows, I think this comes from mingw, however the script itself might make calls to commands that are not available on windows. Ideally the script should be translated to powershell and use that while running on windows.
Another thing to consider, does this compile on macOS?
I just got off a call with the radioss guys.
They’re all in on helping us.
They just finished a few things; one that lets gmsh make meshes for openradioss, another that makes vtk files, and yet another that makes csvs that can be read by anything.
Soo yeah they’re basically all in on supporting paraview, us, gmsh, and everybody else.
These guys are out to take over the world lol.
We should help them.