Ansys Mechanical / Workbench Spherical Constraint

[BarryWatson]

I'm analysing a structure which has a cylinder and both the rod ends have spherical constraints. Free rotation x,y,z but can react translational loads in x,y,z. These spherical constraints are not grounded and are within the structure itself. When I use a body to body spherical contact the analysis does not run. Further reading tells me this is because solid elements need to be further constrained or the spherical contact surface has can infinetly rotate. However as it is body to body I can't put any global constraints at the contact surface.

Does anyone have any suggestions?

 

[rickfischer51]

Do you have weak springs activated?

Rick Fischer
Principal Engineer
Argonne National Laboratory

 

[BarryWatson]

What would you suggest regarding weak springs? Should I also have large deflections on?

I have managed to set up a model using contacts, but it is very expensive with regards to running time.

 

[rickfischer51]

What so you mean by "the analysis does not run?" Does it fail to run with an error message or start but fail due to non-convergence? What warnings and error messages are generated?

Generally, a model in a static structural analysis needs to be tied down somewhere of you get solution problems, typically a deflection limit exceeded or negative pivot ratios. Presumably the thing you are modeling is tied down somehow, so start with that. You can do this with a zero deflection, a fixed boundary condition, or a fixed joint. I would turn on both weak springs and large deflections. Large deflections is especially important of something is rotating.

I've found the easiest way to hose up a model with joints is to get the coordinate system wrong. Make sure the origin is at the center of the sphere and that it makes physical sense with respect to your model.

Rick Fischer
Principal Engineer
Argonne National Laboratory

 

[BarryWatson]

Thanks for your help. Prior to your post I turned on large deflections, weak springs and indeed the local axis of the spherical is central to the local axis of the spherical constraint. I'm awaiting results...

Globally the boundary conditions are correct. The problem is with the body to body contact being underconstrained, this was the problem with the first few runs as stated in the error message. Hopefully this latest run will work and the local and global reactions should prove this.

Thanks for your input.

 

[BarryWatson]

After a few ways of modelling this I managed to get accurate results. However, as first anticipated I was not able to use the body - body to spherical constraint. Instead I set up a local axis at the spherical node, transformed it to the cylinder line of action and used a remote displacement for the correct rotation and translation, taking the physical cylinder out the model. It yielded reaction forces as expected. Thanks for your help.