bolt connection restraints

[ikkedus]

I'd like to analyse a 3d static structure with Designstar. The structure is hinged with 2 plates with a hole of 11mm. Through the holes are 10mm axles with a simple nylon bushing. I'm trying to find out what will fail under load, the beam or the hinges.
[tt]
----- | F
-o---++++++++++++++++++++++++++++
-----
>-----
[/tt]
- = plate
o = hinge
+ = beam
| = force
> = prescribed restraint, no translation in x-direction

I'd like to analyse this by putting a restraint on the holes, but i don't know which restraint i have to use. The structure can rotate around the bolts, but i don't know how to apply a restraint for that kind of condition. Maybe it is better to model the complete bolt and specify some kind of contact condition?

Thanks in advance,
Ikkedus

 

[vonlueke]

ikkedus: If your axle (or bolt) is modeled with a beam element, simply release the torsional dof (dof 4, also called Rx) at one end of the beam element. The beam element will now act as a hinge. Search your software Help to find where the beam "end releases" function is located, in order to release beam element dofs.

I'm assuming you know your structure must be stable in order to perform a typical statics analysis. Otherwise the structure stiffness matrix will be singular, causing the software to fail to find a solution.

 

[vonlueke]

If your axle passes through your clevis plate and connects to something else, which is resisting the applied torsional load, then releasing the axle beam element torsional dof won't work. In this case, without releasing any dofs on your axle beam element, you could spoke the axle beam element end node (located at your clevis plate midplane) to the inside of your 11 mm bushing hole using rod elements. (Rod elements have a ball joint at each end; i.e., only dofs 1, 2, 3.) You could perhaps set the stiffness of the rod elements to a relatively high value (e.g., increase E by a factor of 100), thereby creating an essentially rigid, point support for the axle, which also allows the axle to spin.

This approach gives you a conservative analysis of your axle because it neglects bending resistance of the nylon bushing, which might be relatively low, anyway, depending on the scenario. Not necessarily a bad approach.

But if you have a bolt (or axle or hinge) that terminates at a plate, indeed release the torsional dof at one end of the beam element, as described in my first post.

 

[DRDearth]

To All.. you are in luck... Check Machine Design magazine "FEA Update Column" I have a 3 part series on finding reactions to Bolts and Fasteners. Part I just came out in the Dec 9, 2004 issue of MD. Parts II & III come out in February and March. With the next 2 installments in this series readers can download a sample problem with reactions at bolts computed using conventional hand calculations, a Rigid Link Nastran model and a Flexible Nastran model.


David R. Dearth, P.E.
Applied Analysis & Technology

 

[ikkedus]

Hallo Everybody,

thanks for your answers.

this week i found a way to model the bolt and apply a surface contact condition that works quite well. The result matched the value of deformation that was found in a 'real life' test very well. At the moment i'm analyzing the other models with the same kind of hinge but different loading and plate thicknesses.