How can I simulate bearings that allow the rotation only around the X-axis, as shown in the picture above?
How can I correctly find the stresses and displacement of a pressfit done by thermal expansion? I usually define tie constraint and then define the thermal field. Do I need to set the option ‘‘surface interaction’’ and ‘‘contact pairs’’ or make other things?
I usually make a normal distribution of the output values of the contact surfaces in order to find the stresses and the ‘‘interference fit’’.
To control the rotation of solids you have to use rigid body constraints. You can prescribe rotational BCs to their reference points.
Here’s what is advised in “Building Better Products with Finite Element Analysis” by V. Adams regarding traditional press fit simulations with thermal expansion:
[…] use an initial clearance between parts and apply a thermal expansion coefficient to the inner part. Combining this α with an ambient temperature calculated to expand the part to its preassembled size would allow the press-fit to engage more naturally.
Basically, you have to model proper initial clearance instead of penetration and run a static analysis with temperature field and thermal expansion coefficient calculated in such a way that a proper deformation is achieved. It might be also necessary to use orthotropic material so that it expands only in the desired direction.
This is usually achieved with contact but reference points of rigid body constraint can move and thus they can be used to simulate complex motion with properly defined BCs utilizing amplitude definitions.
Thank you.
So, assuming a correct initial clearance between the solids and definining a static step with a temperature field, how can I set the interaction of the two surfaces that will be in contact? I think that now a constraint tie shouldn’t be the correct way, but a sort of linear surface interaction should, isn’t it?
You should use contact (without adjustment), not tie constraint for this. Initially, the surfaces will be disconnected and then contact should be established once the model deforms.
I did a simple model to verify the bearing simulation. I defined a rotation on one edge of 90° with a rigid body BC. Then, I defined a rigid body constraint and imposed a fixed U1,U2,U3 to the reference point in order to simulate the bearing.
So, I would expect the same rotation of the two defined reference points. But it’s not what expected. It says ‘‘failed with results’’ but with strange behavior. Bearing_sim.pmx (1.8 MB)
The question is what exactly you want to simulate in this case. Perhaps it would be better (and easier) to just apply a centrifugal load if you want to calculate the stresses due to rotation. To simulate free rotation you could also use a dynamic analysis. But it all depends on the goals and expected outcomes.
What I wanted do with this simulation was to verify if the bearing works as expected. So I applied a rotation of 90° on one edge; so I expected to have the same rotation on the ‘‘bearing’’ edge. I didn’t mean to analyse any stress, only a correct set up of the bearing simulation during a generic rotation around X-axis
