Fastener modelling

[juancervi]

Hello

How can I represent fasteners (Hi lock type) in a structure represented by solid (CTETRA elements)?. The model takes into account the holes for the fasteners.

I have used Beam elements connected by rigids (RBE2) to the nodes in the hole surfaces, but it gives very high stress results in the border elements.

One of the the objectives of this model is to evaluate the stresses near the hole border to predict the life of the structure.

I Have attached an image of one side of the joint, showing the RBE2 spiders, the beam element and the solids.

Thanks

 

[rb1957]

first, using RBE spider mesh to distribute is reasonably good (and common) practice. joining the sipiders is a preference ... you're using beams, fine; i use CELAS.

"One of the the objectives of this model is to evaluate the stresses near the hole border to predict the life of the structure." good luck trying to do this with any model! with your model you're doomed ... only eight elements around the hole. looks like you're using TET10s (TET4s would doom you no matter the mesh fineness). if you're trying to do F/DT, it's better to extract the fastener load and the by-passing stress from the model, and then do a fatigue calc with these. another reason you're doomed ... how are you accounting for the interference stress created by the Hi-Lok ?


another day in paradise, or is paradise one day closer ?

 

[juancervi]

The fastener is suposed to be installed with transition fit, according to NAS 618 (this leads to no interference).

I Agree to the idea of a more refined mesh in the hole border, but if i am using RBE2 elements i think that it`s not necessary.

What do you mean to "by-passing stress from the model"?

Thanks for your answer

 

[rb1957]

the stress in the plate is due to two sources ...
1) the pinload, reacted by a remote tension stress, and
2) the bypassing stress, like tension around the hole, yes?

in a lug you'd have only 1 direction of loading. in a plate you could have two (x- and y-).

eight TET4s would be a coarse mesh at a hole. eight TET10s is better (obviously) but if you really want to know the stress at the hole i think you need something like 32 or 64 (or more !)

for me Hi-Loks install with some interference, more than the none you're modelling. i think this is an important component of stress in the bore, if that's what you're interested in.

another day in paradise, or is paradise one day closer ?