Mechanically Fasyened Joint Analysis

[bostitch1963]

Hello,

I amm analyzing a multi-shear plane joint consisting of four aluminum sheets fastened tpgether using solid aluminum rivets.

For the static analysis I assumed bearing yield would result in all rivets being loaded equally.

For fatigue I created spring models (one using the NACA method and the other using Huth). Out of curiousity I loaded the model to ultimate to see if in fact the end rivets were bearing out - thus allowing for load redistribution.

What I discovered is that all the bearing margins are relatively high (MS > +0.50), but the shear margins are negative (MS < -0.20) for the end fasteners.

Knife edge won't allow for larger fasteners, and program mgmt will not allow higher strength rivets (due to manufaturing difficulties)

Thus, my dilema. I believe that ignoring the negative shear margins from the spring models is unconservative.

What are your thoughts on this?

Thank you

 

[bostitch1963]

Dang. I should flag my own post as inappropriate - all those typos are embarrassing. Can we edit our own posts?

 

[KENAT]

You can flag it as innapropriate and ask them to edit.

KENAT, probably the least qualified checker you'll ever meet...

 

[40818]

Depending upon your loading scenario will give a different load distribution within your fasteners.
Take an aircraft emergency landing scenarion (9g fwd), it is assumed that all the fasteners will take an equal share of the load due to strain compatability of the joints, this is the basis of the simple bolt group analysis (Sum of R^2). If the loading is not taken to match this criteria, such as limit loading then the strain matching of individual bolt joints will not occur, and you will develop higher elastic stresses within the material adjacent to each bolt. This is the way FE distributes the loading in a linear static model.
One ongoing probelm with derived fastener spring stiffnesses is that amongst other factors, they are a function of type and material of fastener, hole condition, fit and preload and geometry and material of the plates. The best way to get your constants is to test.
if as you say, your getting a decent margin on bearing, but low margins on shear means that its a crap design. We should never really design a joint to be shear critical, unless it has a very specific reason for being this way (fuse pins being one example).
Play tunes with fastener sizes and pitches if possible, or reduce the end fastener sizes to increase there flexibility, or look back at the loading deivation to see whether there is any over-conservative assumptions in there.

 

[rb1957]

personally i wouldn't use a fastener stiffness model for static strength calcs. i'd assume load redistribution withini the joint so that each rivet would be allowed to carry either it's maximum shear allowable or the local bearing allowable, whichever is smaller. the complexity in your joint is how the 4 sheets distribute the load in a rational manner ... you could you fastener stiffness models to start this, then allow some fasteners to yield (presumably the outer ones, near an edge), redistributing load to the less loaded fasteners (presumably the inner ones, away from any edge).

 

[RPstress]

If they're down in shear they'll yield in shear rather than yielding in bearing. In metal it's notmal to assume reasonable at ultimate redistribution with maybe 10% on top.