fatigue of complex riveted joint - bypass load

[ahkrit]

Hello.

i would like to know how the bypass load in a complex riveted joint could be calculated? The fasteners are modelled with bush elements and the neighbouring shell elements show different von Mises stresses. The load propogation also not clear as the joint is complex and has 3D load case ( each end has different loads and moments). So the basic calculations like in Niu`s book do not really help in that case.
How could a fatigue calculation work in that case?

Thanks in advance!

 

[rb1957]

is bending truly transferred through fasteners (as a moment) ? sure, bending (due to shear) is a component of fastener design ... haven't really seen it in hole fatigue.
if you really want to do this, then I think you need to create your own analysis (and test to validate).

for myself, I'd use the largest principal stress (the larger of the two surfaces). see the attached pic. You have the pinload, yes? for the four elements around the fastener two (A and C) should have a higher principal stress than the other two (B and D) use the higher as the "remote" (reference) stress and the smaller one as the "by-pass" stress. This is not perfect by any means, but is a good way to use the data you have. (Perfect would be to use a canned code, like ncode, to do fatigue analysis.) Maybe collect data for all fasteners, but it should be obvious which fastener is critical (it should be one of the perimeter fasteners) so one calc to characterise the joint.

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

 

[ahkrit]

Thanks for the info. Well if it is "not perfect by any means" why would you suggest to do it that way?
On the other hand then calculating all the other factors depending on simple geometry (like plate width W in Nius book) i have no idea how it could implemented for bended plate with complex geometry... any idea on that?

Bending load which is transfered due to shear through fasteners, why should it not be included in fatigue? Especially if this connection has quite high two axial bending which varies in time?

I was trying to look for what ncode is (or canned code). is it usable for different complex geometries?

I am getting really fed up that everywhere only the basic one axial tension cases presented for fatigue. like that is the only possible load in the univerese and one gets lost if things get more complex... but please give me some advice! would be much appriciated!

 

[SWComposites]

ok, a few questions:
what is the material for your plates?
what is the application for this analysis (you ref Niu, so is this for aircraft?)
do you need bypass loads for static or fatigue analysis?
in what detailed analysis method are you going to use the bypass loads?

figuring out bypass loads for a general multi-axially loaded, multiple fastener joint is quite complicated, either by hand or with FEM results. probably too complicated to explain on this forum, but if you have a specific joint configuration, with dimensions and loads, then we might be able to do some hand calc estimates to guide you.

 

[SWComposites]

oh, and do you need "gross bypass load", which is the total load going by the side of the hole, or
"net bypass load", which is the load going by the side of the hole which is not reacted in bearing?
note: for tension loaded joints, gross bypass = net bypass load + bearing load
for compression loaded joints, gross bypass = net bypass load

 

[ahkrit]

@SW composites

what would you suggest if you are modeling a joint where each connected beam has two axial bending and torsion, thus the distribution looks something like on the picture above posted?
So all the fastener forces look in different directions. I would say net bypass load is neccessary, then i have everything as fastener loads i have already,or not?

but once again. in that case what are the w and r/c values for stress concentration factors (Ktg and Ktb) if the direction of fastener load varies with time?
can this calculation method used anyway if the fastener load direction variies in time?

 

[rb1957]

"I am getting really fed up that everywhere only the basic one axial tension cases presented for fatigue." ... well, that is, to a very large degree, what happens.

The fastener fatigue is dependent primarily on fastener tension stresses. Hole fatigue is primarily due to bearing (due to the fastener shear) and the by-passing tension stress in the plates. Yes, there is bending due to shear, but this is, in typical joints, not critical. Yes, there is some bending under the head/nut, but this, in typical joints, not critical.

"I was trying to look for what ncode is" ... try googling "ncode" ... it came up as the first hit ! Ncode is fatigue analysis s/ware. I don't know what codes it works with (nor what you have).

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

 

[ahkrit]

The problem is, that in my case the direction and magnitude of the loads on the fastener (and the bypass load too) changes with time.
Bending moment also occure on the beams which causes the kind of "circular" reaction pattern on the fasteners as posted above. How could one proceed with fatigue in that case?

With ncode my problem is as soon as i cannot see the step by step calculation, (which also includes depending on the calculation method (nominal or local stress values or i dont know what..) what values at which point should be evaluated with a given discretisation)
then how could i do that with a full new software? i am using abaqus btw.

 

[rb1957]

Then I'd suggest adopting a conservative direction, say the peak happens at the same location on the hole, circumferential would be my suggestion (to align with cabin pressure, if this is on a fuselage) or spanwise (to align with normal bending, if on a wing).

Else, if you have way too much budget and time, exact stress at the hole in say 10 deg intervals for each load case and run these individual calcs.

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

 

[Ng2020]

Your problem is not unique ahkrit
A typical approach is to assume that for each flight segment, the loading direction does not change, you should be able to verify that this approach is conservative, and indeed it becomes essentially a 1d problem. You might do this by identifying the flight segment which produces the highest bypass tensile stress (and contributes the most damage). If it helps, create a free body of the critical hole for this load case. Now assume loads for all the other flight segments act in the direction of this critical case, and therefore contribute to damage at the same location on the edge of the hole.

 

[ahkrit]

@Ng2020
i think so as we that my problem.is not unique, still have not found any usefull information on the calculation procedure.

I am working on floating structures, that is why the load can vary easily bith in direction and magnitude.


Can you offer anyone who could provide individual courses/technical supervision?

 

[rb1957]

"floating" meaning on water ? I'm sure there's a tonne of guidance for that specialised field ... trouble is it's hard to find the rock (that this info is under).

But generally, for a complex joint such as yours its either ...

1) figure out a conservative loading, an envelope that covers all, or

2) do an incredibly detail calc (in an attempt to capture the "truth") and spend years analyzing the joint.



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