I am simulating a bar of SS crashing into a concrete wall at a speed of 1600 m/s, i already did this in a different solver and tried the same in this but the results are different. The conditions and parameters are all the same as that of what i did in ansys, but results are different what could be the problem cause the results arent exactly close.
Have you seen this tutorial ? https://youtu.be/-X0Shj7UXE0?si=5W5m5Sa9X32d-8Rn
Explicit dynamics procedure in CalculiX is very limited and problematic. It might be best to try OpenRadioss instead unless the problem under consideration is really simple.
Of course, it’s hard to say what’s wrong in your case without seeing the model (at least the screenshots, but it’s best to just share the .pmx file).
Thank you for the response. I used this exact video as a reference to model the simulation; the only difference is that my model was created in 3D. I am aware that explicit dynamics in CalculiX (CCX) is quite limited.
I am currently exploring multiple preprocessors and solvers, one of which is PrePoMax. The community support, documentation, and availability of tutorials for PrePoMax are significantly better, which is why I chose to use it. In contrast, I was unable to find comprehensive tutorials for OpenRadioss, a reliable preprocessor, or sufficiently detailed documentation, which made it more difficult to adopt.
I would also like to thank you for your contributions in creating PrePoMax tutorial videos—they were very helpful and informative.
The image shows the simulation I performed in ANSYS Explicit. I used the exact same material properties, time steps, boundary conditions, and velocity. Initially, I suspected that the discrepancy might be due to meshing differences: in ANSYS, the mesh consisted of simple hexahedral (cube) elements, whereas in PrePoMax the mesh initially consisted mainly of triangular elements. Since triangular elements are generally stiffer than quadrilateral elements, I switched to a quad-dominated mesh in PrePoMax; however, the results remained unchanged. I then tried a tetrahedral mesh, but for some reason the results were even worse.
I am a new user so I unfortunately do not have the required permissions or access to upload the .pmx file.
Actually, there’s a converter from CalculiX .inp files to OpenRadioss, maybe you could try it: Tools/input_converters/inp2rad at main · OpenRadioss/Tools · GitHub
There are also some add-ons for preprocessing in FreeCAD: https://forum.freecad.org/viewtopic.php?t=101949
Thanks, I’m really glad you like them.
I see that you also used element deletion in Ansys. Unfortunately, CalculiX doesn’t have it, although damage initiation was added in the newest release. It looks like the main difference between your models - with CalculiX, the mesh can only deform and eventually distort excessively while Ansys or Abaqus can delete failed elements.
In PrePoMax, you can generate good hex meshes (they are pretty much always much better than tet meshes) using Gmsh algorithms: revolution, extrusion, sweep, transfinite and offset of quad/quad-dominated shell mesh.
Also, keep in mind that it’s best to use symmetry and model e.g. 1/4 of the model whenever possible. 2D models can also be helpful, although they sometimes cause issues in CalculiX because its 2D elements are extruded to solids. For example, rigid body constraints cannot be used with 2D elements in explicit dynamics.
You should be able to do it now.
1 Like
Thanks a lot for your response it was extremely helpful. The files unfortunately too big to upload.
From my understanding now is that I should use either PrePoMax or FreeCAD and export the files to .inp format and use the python converter to .rad.
Maybe you could use some hosting website, but perhaps the file won’t be necessary considering the rest of our discussion.
I would try the inp converter first. Apart from incomparable robustness, OpenRadioss will let you achieve what you obtained in Ansys - element deletion and thus perforation of the plate.
Hey @FEAnalyst ,
Thanks again for all your help.
I initially tried using PrePoMax with OpenRadioss, but the workflow involved too many conversion steps. The model had to be exported to a .inp file, then converted to .rad using a Python script, followed by running it in OpenRadioss—which generated multiple warnings. After that, additional output converters were required to obtain .vtk files for post-processing.
Because of these multiple conversions, I suspect there was some degree of data loss, which likely affected the accuracy of the results. Due to this, I eventually switched to LS-PrePost, since its keyword deck can be used directly as input in OpenRadioss, resulting in a much cleaner and more reliable workflow.
While CalculiX (CCX) may not be ideal for explicit simulations such as crash and impact analyses, it works very well for structural and thermal problems. PrePoMax is also an excellent, beginner-friendly GUI that makes both pre- and post-processing much easier.
Overall, your guidance was extremely helpful and played a big role in moving my project forward—thank you for that.
hi, conversion to LS Dyna format is officially supported in CalculiX GraphiX by send commands. Element and surface group definition is supported; later material and boundary condition is needed to define by user input decks.
It could be useful when PrePoMax have this feature also, with additional enhancement in material and boundary condition when exporting.
Export from PrePoMax to .inp is “natural” since this format is always used by CalculiX when submitting analyses with it. So the only problematic step is conversion from .inp to .rad - this may indeed fail to handle some features. Then output conversion to .vtk shouldn’t be a problem either as using ParaView for postprocessing is common for many simulation tools that don’t have native postprocessors.
However, maybe the workflow with FreeCAD and its new add-on workbench would be better for you. It would just require preparing the FE model from scratch in FreeCAD.