My model is a bit different but the scenario is the same. I have tried many different ways to mesh the two parts but cannot produce any successful output. From what I have read it seems to e related to creating a mesh with correct convergence. As you can see in my model the rail is cut in 3 sections so I can fix the end sections to simulate a stitched weld along the rail. I have tried basic mesh, Tertrahedral gmesh and also extrude mesh on a simpler model. Simulation usually fails.
what is the correct approach to this scenario ( cylinder vs flat) ? are there any examples I can learn from. I have spent 2 days on this and clearly I am missing something.
I will check the file a bit later. For such simple geometries, it’s best to use a hex mesh. However, typically there’s a significant refinement in the contact area and much coarser mesh away from it.
However, it’s also highly recommended to make use of symmetry whenever possible (if geometry, BCs and loads allow it). It not only speeds up the analysis but also make it much easier to constrain the model. With proper symmetry BCs you may eliminate all or almost all rigid body motions. In the case of contact problems, it’s best to use prescribed displacement instead of force load as this aids convergence: https://youtu.be/-a5TX1-9rJ4?si=xuZlRaIgKrkBWtOm
Thank you. I now understand using prescribed displacement instead of force. I have a couple of questions about the mesh set up. How do you choose the minimum element size and why is mesh refinement not used in your modified model?
The model I shared was just to show you how to make the analysis work and propose some simplified setup utilizing symmetry. I left mesh refinement for you (also because the file would be larger). It may require several iterations - you should check if the results change significantly when you refine the mesh. There are no specific rules for element size, it’s usually selected for a given case based on experience (and sometimes engineering codes or literature), then refined according to a mesh convergence study. You can find some general tips here: FreeCAD FEM - Meshing basics and here: FreeCAD FEM - Mesh convergence studies.
For train wheels on rails, there are multiple Hertzian and non-Hertzian models described in many research papers and handbooks (such as “Wheel-rail interface handbook”). However, there are also guidelines for contact between crane/gantry wheels and rails.
For monorail beams I think that the Hertzian contact is not so important because of the bending on the flanges which leads to yielding in the area near the flange-web connection
Yeah, but it might be interesting to fix the flange and refine the mesh in the contact region based on the Hertz theory solution. Plus I’m still referring to train wheels and rails as in the Russian tutorial referenced by the OP. It’s a very interesting but quite complex problem actually. But crane rails may also involve such calculations:
It’s a classic diffusion of the wheel load on the web. On the latest version of prEN1993-6 (next publication) there are also formulae to take into account supplemetary local stresses due to wheel load eccentricity and lateral forces due to skewing or acceleration/deceleration (bracking) of the crane bridge.
