Is it possible to use FEA on a bolted structure?
If not can I run the FEA on the structure as a welded frame & get loadings at the connection points to manually calculate bolted connections?
sure you can, have you tried? where are you getting stuck?
Not tried yet. Unfortunatly the only copies of Inventor Simulation are with the two people who almost never use inventor.
I was told when working for a previous employer that it wasn't possible to anylize a bolted structure, if it is possible then we may start using it.
Hi Steve,
Although we don't have a devoted bolted connection feature / load or contact type, there are ways you can get close.
The most straightforward would be to split the faces of the flange or plate with the "footprint" of either the washer or bolt head. Exclude the bolted connection from the Simulation. Then create a manual contact using those split faces. You can use Bonded or spring type contact.
This gets more tricky, but you could also cut a section out of the bolt shaft itself, so there is a gap. Then, create a pre-tension in the bolt by adding a force to each half of the bolt (on the newly created faces by the cut) pointing towards each other. Be sure to make the cut wide enough so that it lies outside the default distance for automatic contacts or else a bonded contact would be made, and the forces wouldn't have any effect.
If the bolts were placed from Content Center, you cannot modify them with the cut, unless you chose to place them "As Custom".
Please let us know if you have any additional questions, comments or suggestions.
Thanks, -Hugh
Hi Sushil,
Attached is the screen capture of the bonded contact and split-face method.
Cheers, -Hugh
Great analysis, well done! This process of dividing up the faces and then applying appropriate loads has lots of other applications as well. Simple. Thanks for doing this.
can you please help me to do simulation on bolted connection , in my case , to lift a metal frame by using hollow section , this hollow sections connected to the metal frame through bolt connection, i m stuck with the simulation, can you please advice me on this?
Hi Hugh,
I am working on an assembly that requires analysis of a few bolted joints and I'm used your first method i.e. BONDED contacts between washer surfaces for my simulation (Highlighted red on the attachment). In addition I have also applied a SEPERATION contact between BLUE bracket & GREEN bulkhead. The GREEN bulkhead is constrained with a FIXED constraint. Then the load is applied to the simplified bearing assembly shown in GREY. To calculate bolt stresses I need to find out the force components acting on each red surface. Can you please advise me the best way to do this.
(One way I figured out was to apply a FIXED constraint to each of the surfaces highlighted in red. Then you can get X,Y,Z reaction force components acting on that surface. As the distance between RED surface and the GREEN part is relatively small adding FIXED constraints again to RED surfaces doesn't affect results significantly. But if the RED surfaces are further away from the GREEN bulkhead adding FIXED constraints to RED surfaces, as per my belief distort results. So if there is a way of finding the force components acting on the RED surfaces without adding FIXED constraints to each RED surface that would be ideal)
If anyone else think that my method is flawed your opinions or suggestions are welcome.
Thanks & Regards,
Sachith
Hugh,
Thank you for the comparison on modeling techniques. It has been YEARS since you've posted your comments, however I just came across your comparison yesterday (Literature really has the capability to speak to the future).
I have a few question on the difference in results that I'm observing in your example.
1.) On the example that splits the bolt, the browser shows 3 loads and the image shows 1 load. Are the other 2 loads active? I ask because the there is evidence of a torsion in the visual displacement of the "split-bolt" model. (Difference in loading might help explain the other questions I have)
2.) I would expect evidence of approx. 9MPa on the "split-bolt" plate towards the center of the plate similar to the bending stress at the same location of the "contact-method" unless it's simply out of scale.
3.) The resulting "split-bolt" maximum stress is almost four times (4x) the maximum stress of the "contact-method" results.
Any and all enlightenment to the differences in outputs would be sincerly appreciated.
Thanks!
-Ben