Best practices to simulate shaft-Bearing assemblies

Julio_21000

Dear,

I would like to simulate an assembly made out of one shaft and two bearings. For these two bearings I have radial, axial and tilting stiffness values out of another calculation. Therefore I think about setting these stiffness values using springs. But I am afraid of getting an unstable assembly that would lead to large displacements.

Is there any other method within Simlab to simulate bearings?

Thank you.

Best regards

Ioannis Makris_22013

Hello Julio,

In most simple cases bearings are not modeled. Instead, the bearing assembly is used for creating the appropriate constraints (fixed bearing and floating bearing).

If you want to go a step further you can create the bearing bodies, assign the appropriate material and observe what happens. In this case rigidity and stiffness is relevant to the bearing geometry and material.

If you want to create a model in which the geometry of the bearings is not present, yet the stiffness to certain DOFs is, then the spring option is a nice alternative.

The thing is, you cannot know if your model is unstable before running the Non-Linear Solution and observe what is happening at every increment.

My suggestion would be: Start with Linear Static, then go to Non-Linear Static with fixed and rigid constraints by adding non-linearities one by one (Start with Large Displacement, material and then contact). When your Non-Linear model converges to expected values you can experiment with other boundary conditions and different elements (Springs etc.).

I hope this helps.

Kindly,

Ioannis Makris

AlessioLibrandi

Hello Julio,

one additional note. 

In order to create a spring for such cases, you could use the panel analysis/connectors/bush.
Here you can define not only the stiffness in the three directions, but also the different damping coefficients which would help you to increase the stability (see attached picture).

Alessio

Julio_21000

Thank you both for your kind answers,

If I am right, bush connectors do not allow body-ground relationship, and therefore is not usable for my particular simulation. But will have them in mind for future setups.

I have run the simulation (non linear, because I have some press-fits contacts on the shaft) using just springs, and as expected, it has given me back no results (stresses  and displacements = 0)

I am afraid of using fixed constrains and of getting because of them, a non realistic deformation/stresses on the shaft. Any advice here?

On the other side, ideal would be to model the bearings´ bodies, but in order to reproduce their stiffness right, a detailed geometry knowledge is mandatory. Also internal geometry details. Most of the cases this is keep in secret by the manufacturers. I have seen that Simlab has a built in tool to model bearings:

The correct way to use this tool would be the last option on the list, but as told, most of the geometry parameters are not known. I also have 3D models of those bearings, but I know that the internal geometry of them is not 100% correct. For instance number of rolling elements in the 3D vs actual bearing is not right. Additionally we get form the manufacturer *.xml files which contains relevant data (geometry and stiffness values) for our standard calculations. I have attached one as an example.

Could you please advice me about the best practices or method to model bearings? Does Altair have any bearings´ database?

Thank you.

Julio