semi rigid connections 3

[engroma]

Hi all

I have to model bolds which connect two parts of an assembly and I am thinking of using rigid elements or modelling the bolts with a beam and rigid elements, therefore I would get rigid joints.
I would like to have a semi rigid connection instead of a rigid one.
How can I model a semi rigid connections between two or more parts using FEMAP?

Thanks
ingroma

 

[GregLocock]

My advice is never to use rigids. Chunk out your structure as a series of beam or plate elements, or write out a stiffness matrix.

However if you are not really interested in local stresses then rigids may be OK.

Cheers

Greg Locock

SIGlease see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[rb1957]

well how are the two pieces joined together in real life ?

are they discretely bolted together ? welded ?

 

[engroma]

the pieces are bolted together using fastener or bolts, I am interested in the local analysis because the parts are manufactured using composites.

thanks

 

[GBor]

While I agree with Greg, using a beam with the cross-sectional properties of your bolt and attaching each end the of the beam to holes in your structure with rigids is a common practice.

Instead of using rigids, you may want to use springs (which can also be beams using a stiffness of AE/L -- area * modulus / length).

You may want to search for a paper called, "Method of Modeling Bolts" by Jerome Montgomery and "Influence of Fastener Flexibility on the Prediction of Load Transfer and Fatigue Life for Multiple Row Joints" by Huth.

Garland E. Borowski, PE
Engineering Manager
Star Aviation

 

[spongebob007]

One thing I have done is to partition the surface on the bolt hole with a circle that is the diameter of the bolt head. I attach the rigid elements to all the nodes on this partitioned surface in order to spread the load out in a manner that is more like how the bolt head might actually behave.

You might also find that you have no choice but to explicitly model the bolted joint with 3D elements and run a contact analysis.

FEA is somewhat of an art, and it is often not a one shot deal. Try modeling your structure a couple of different ways and then look at what the results are.