Rivets modelling to connect 2D to 3D mesh

[flyforever85]

Hi all,
I have a 3D mesh connect with rivets to a 2D (midsurface) mesh. I'm not sure how to model rivets in this case. If the case was 2D-2D I would create a RBE2 web for each hole and connect each central node to a CBAR. How this translates into 3D-2D?

 

[rb1957]

The issue is, as I'm somewhat sure you're aware, that nodes on 3D elements have only X,Y,Z force freedoms (ie no rotational freedoms; whereas 2D elements have 6 dof nodes.

If you've modelled the solid anyway 1/2 decent, then you'll have several nodes though the thickness, at least nodes on the two faces. So a RBE spider to the mid-thickness point would be the way to go.

That said, I wouldn't normally model rivet holes. I'd model the rivet as a node if I really wanted to, or else extract rivet loads from a more "gross" load ... the shear along this line of rivets is X lbs, and there are Y rivets, so rivet load is X/Y.

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

 

[FEA way]

Which software do you use ? I would use beam (1D) element for the rivet itself and connect it to the rest of the model with rigid links (spiders). One link attached to the surface of the hole in solid part and the other link attached to the edge of the hole in shell part.

 

[rb1957]

the usual approach ... a beam element or a CBUSH (that'd be my preference if I was modelling a fastener). We do get hung up on the difference of 1/2 a thickness (using mid-plane surfaces for skins). It really doesn't matter (IMHO, IMAO?) if you model skins as the OML, the IML, or the mid-plane ... in 99% of models. We all get trapped by taking a 3D machining STP file into our FEA of choice and then we have aero loads from the OML and now we have 0.1" gaps and it's just a bytch ...

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

 

[flyforever85]

Thank you all for your answers.

I need to model the hole because during the machining of that component that hole was drilled larger than specification. Since the component is considered critical and not simple to install, we are checking if and how the stress distribution changes. We are considering to create a web of RBE3 rather than RBE2

 

[rb1957]

ok, well, you know the situation better than we do. Must be a very sensitive part if an oversized hole causes this much response.

Have you considered an oversized fastener ?

Have you considered bushing the hole ? An interference fit bush usually works well (with Al parts).

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

 

[flyforever85]

We are assessing the problem atm. A bulky reinforcement is attached to longerod (the 2D part of the analysis) and we are worried for it of course, more than for the reinforcement. I assume bushing a hole of a 0.81 mm plate is not a possiblity.

 

[rb1957]

0.81mm = 0.032" ... so I guess we're dealing with s/m parts, Al, possibly steel, but probably not composite.

bushing thin s/m is difficult.

can you use an oversized fastener ? This should make most things about the s/m piece "better" ... so long as you have edge distance.

how much is the oversize ??

is this one hole is a line of several ? will the joint work with one fastener ineffective ? (ie install B/P fastener (with clearence) into the oversized hole ... not pretty but is it "wrong" ?

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

 

[SWComposites]

Attaching to a modelled hole with either RBE2 or RBE3 will give nonsense/rubbish/wrong stresses at the hole edge.

If you have an oversized hole, one would typically take the baseline joint loads and assess the oversize with hand calcs.