Discussion Groups

Simulation Mechanical and Multiphysics

Reply
Contributor nav
Contributor
nav
Posts: 15
Registered: ‎09-27-2011

Large Bearing Model - Using Gap and Beam Elements?

718 Views, 6 Replies
10-20-2011 09:11 AM

I'm currently working on a project that uses a large (approx. 20' DIA) bearing and I've been trying to come up with an accurate process for modeling the bearing so I can obtain contact stresses between the ball bearings and inner and outer races.

My FEA background is somewhat limited so I began by creating much smaller bearing models to attempt to come up with a process.  Currently, I create the inner and outer races as solid parts and then uses 4 line segments (essentially an 'X' to simulate the four contact points of ball on the races) the large lines of the 'X' are set to be beam elements and then very small gap elements are placed at one end of the large lines to represent the gap between the bearings and races.

I then fix one face of the inner, set up surface forces on the outer race and run the simulation.

I always seem to get errors such as the following:

 

warning: your model may not be tied down
enough or you may have a change in
stiffness somewhere in your model
which is too abrupt. Check DOF: 25490

Warning: quasi-periodic solution achieved.

With following gaps active:

by activation tolerance 0.0000E+00

 

I'm under the assumption that these errors come from the boundary conditions (or lack there of) and also the stiffness that I've been using for the gap elements (1,000,000 lb/in).  

 

As stated before, I'm fairly new to FEA and have been trying to come up with a successful solution by doing my own research, but I haven't been have much luck.  Are beam elements what I want to used to represent the ball bearings? Or would truss elements be a better option?

 

Any input is much appreciated and if you need addtional information, please ask!

 

*Edit*

 

To create the wireframe/solid model I've been exporting each section (i.e. wireframe, then solid) and opening them seperatly in Autodesk Simulation. I then use the merge feature and adjust the mesh so that the nodes between the line elements and 3D mesh of the solids align. Not sure if this is the best way to combine the model.

 

 

 

 

Please use plain text.
Employee
INACTIVE_AstroJohn
Posts: 492
Registered: ‎03-25-2010

Re: Large Bearing Model - Using Gap and Beam Elements?

10-20-2011 10:34 AM in reply to: nav

Hi nav, welcome to the group.

 

I think you are correct that the lack of stability and the very large stiffness are contributing to the warning message. The following page in the Help should get you on your way: Help > Autodesk Simulation > Setting Up and Performing the Analysis > Set Up Analyses Part 3 > Perform Analyses - Run Simulation > Perform Linear Analyses > Perform Analyses with Gap Elements. Stabilizing the  model will probably help with the quasi-periodic solution, which means that the solver went in a "loop" where the same gap elements were in contact on different iterations. If we did not detect the periodic solution, the solver would truly get stuck in a loop where it always repeats which gaps are in contact.

 

About your question of beam versus truss elements, you must use beams if the members of the X are connected at the center; that is, 4 lines/elements make up each X. The joints of truss elements are pinned, so the X's would collapse if they were used.

 

Sincerely,
John Holtz, P.E.
Senior User Experience Designer, Simulation
Autodesk, Inc.

Current version of Mechanical & Multiphysics: 2013 SP1 (2013.01.00.0012 28-Jun-2012)
Please use plain text.
Contributor nav
Contributor
nav
Posts: 15
Registered: ‎09-27-2011

Re: Large Bearing Model - Using Gap and Beam Elements?

10-20-2011 10:59 AM in reply to: INACTIVE_AstroJohn

Thanks for the reply John, I'll read through that Help section and see if that helps to clear some of the warnings up.

I wanted to clarify on the beam element setup that are representing the bearings.  There are actually 2 beam elements, one facing (\) and the other (/).  They overlap, however where they overlap they haven't been split.  At the top ends of each (\) and (/) are very small line elements that have been set to Gap elements, compression with a gap.  These were included so that at two contact points, the beam is actually in contact with the races, and the other end is a gap element connecting the end of the beam to the races.

 

Please use plain text.
Employee
S.LI
Posts: 398
Registered: ‎05-06-2010

Re: Large Bearing Model - Using Gap and Beam Elements?

10-20-2011 11:20 AM in reply to: nav

The difference between beam and truss is there are rotation DOF in beam element.

The error shows that some DOFs are not constrained. Maybe you should check rotation DOFs of beam elements to make sure there is no free DOFs.

----------------------------------------------------------------------------------------------
If this response answers your concern, please mark it as "solved".
Please use plain text.
Contributor nav
Contributor
nav
Posts: 15
Registered: ‎09-27-2011

Re: Large Bearing Model - Using Gap and Beam Elements?

10-20-2011 11:38 AM in reply to: S.LI

Thank you S.Li for the response, I'm aware of the difference between a beam and truss element.

 

In my previous post, I explained (or tried to) how the bearings are represented in an 'X' configuration by using 2 beam elements, the large legs (\) and (/).  At the top of the legs, two small lines were drawn to be used as gap elements.  Having no previous experience with gap elements, do  you have any suggestions for how I should constrain them? 

 

Attached is a picture that hopefully better explains my setup.

Also, I'm assuming that since my inner race is the only portion that has a surface boundary condition (fixed), the use of weak springs to connect the inner and outer races alongside the gap elements would be helpful (if not necessary?).

Please use plain text.
Employee
S.LI
Posts: 398
Registered: ‎05-06-2010

Re: Large Bearing Model - Using Gap and Beam Elements?

10-20-2011 12:57 PM in reply to: nav

Do the two beam elememts share a common nodes at the cross place?

 

I don't think gap element has any rotation DOF.

if your situation is beam->gap->solid(brick/tet) or fix, I don't think the rotation DOF of beam is constrained.

 

 

----------------------------------------------------------------------------------------------
If this response answers your concern, please mark it as "solved".
Please use plain text.
Contributor nav
Contributor
nav
Posts: 15
Registered: ‎09-27-2011

Re: Large Bearing Model - Using Gap and Beam Elements?

10-20-2011 01:04 PM in reply to: S.LI

The two beams do not share a node where they cross.

I've attached another image showing the DOF constraints placed on the beam and gap elements.

 

Please use plain text.