Cylindrical Symmetry - Wheel Rail contact - Ansys Workbench 13

[JSN86]

Hello Everyone

I have to do the analysis of the wheel-rail contact, and I'm having some trouble with the symmetry aspect of it.

Consider the wheel-rail assembly as in the attachment. As you can see, the wheel is reduced to 20º or so, of its cross-section. To do a simple static analysis, I assume, I need to have some cylindrical symmetry on the wheel, however after I apply the periodic or the cyclical tree tool(sorry couldn't think of a better term) when I try to mesh the wheel part, the mesher fails.

I'm using the sweep method to all the parts with all tri elements.

The cylindrical (periodic or cyclical) symmetry is defined with a coordinate system centered roughly 100mm above the top of the wheel, and with the directions of X and Y and, rotating counter-clockwise around Y axis.

Note that the coordinate system you see in the picture is the global coordinate system, and not the one I created.

I searched the forum for similar problems, and one of the search results directed me to a thread, that suggested the use of frictionless support instead of using the symmetry condition, but somehow that feels like cheating, and doesn't give me the impression that it's done correctly.

I await your answer.

 

[flash3780]

I'd recommend that you take a look at the meshing tutorial which is available in the ANSYS customer portal to get you started with meshing. More difficult geometries often need to be chopped up and glued together to mesh them efficiently. However, your particular model seems like it could be brick meshed with minimal cutting and gluing. Note that parts need to be joined with the "Form New Part" command in Ansys Design Modeler to form a continuous mesh when you enter the mesher (rather than joining meshes together with bonded contacts, which is inefficient and less accurate).

I wonder what you're trying to learn from your model specifically? If you're interested in the stress distribution within the wheel, perhaps the model doesn't accurately represent the geometry. If you're interested in the contact stress between the wheel and the rail, perhaps your model is much larger than necessary and could be paired down significantly. How much does the rail impact your problem? Is the entire height of the rail required? Finally, if it is a contact stress you're after... have you considered looking at an analytical Hertzian stress solution first? A Hertzian solution will give you a very good approximation of the stresses between the wheel and the rail to compare your model to.

 

[kundalin1]

Regardless of the meshing issue, I don't think you can apply cyclic symmetry to that model. I mean, if you apply cyclic symmetry the rail and the contact would also be repeated every 20 degrees.
When you are dealing with solid elements (no rotation degrees of freedom) the frictionless support is equivalent to a plane of symmetry condition, that is what I would do (two frictionless supports). The alternative is to add two plane of symmetry conditions, you will have to create two new custom coordinate systems to define the normal to the planes of symmetry.
If you need accuracy in the contact pressures you will have to make a submodel closer to the contact region (or go the analytical way as flash3780 suggested)

 

[JSN86]

Thank you.