Holes Tearing Out of Sheet Metal Box 1

[drawoh]

I don't know whether to put this under structural or Aerospace. I am trying to satisfy some aerospace engineers, so it goes here.

I am trying to work out the shear strength of some holes in a sheet metal box subjected to a catastrophic crash load. As per my diagram, I have analysed it as per the middle figure in which chunks are ripped out of the box. I can model this and do calculations. It seems to me that he box should fail as per the figure on the right, with the box splitting around the screws as they are torn out.

Does this sound reasonable? Is there literature on thin sheet splitting as per my left hand figure?

--
JHG

 

[rb1957]

there're a couple of failure scenarios at work ...

one is straight-forword bearing of the pins onto the plate, maybe equally distribute the load over the fasteners. for 9g crash this may not be a valid failure (in a crash you may allow bearing type failures, so long as the thing doesn't fall apart).

another is fastener shear strength.

the other is shear tear-out. look at how Bruhn analyzes shear tear-out for lugs and adapt for your geometry.

your other sketch is more like a tension failure, but typically fsy < fty so it shouldn't happen with typical material and geometries.

another consideration should be buckling of the plate ... this might control how the fasteners load up.

clear as mud ?

 

[drawoh]

rb1957,

This failure would be a 9g crash forward, causing the box to tilt forward. This screw pattern would be in the rear.

I worked out shearing of the screws. We are okay.

I looked in Handbook of Bolts and Bolted Joints by Bickford and Nassar, and there was a discussion of lugs tearing out. As shown, the hole in the lug was fairly close to the edge, and the failure mode looks reasonable to me, on their model. Strictly speaking, my model does not have an edge.

This is an off-the-shelf case. It would be so much simpler if I had designed it.

--
JHG

 

[IRstuff]

9g isn't exactly a catastrophic crash though. 25g to 40g would more like it.

TTFN
faq731-376

 

[drawoh]

IRstuff,

9g can be catastrophic if you are hit by a sheet metal box and its contents, minus the mounting brackets, of course.

--
JHG

 

[IRstuff]

I don't disagree with that, it's just that aircraft equipment aren't supposed to break free with ANY shock load less than 25g. So, why is the box only being designed for a 9g shock? For the platforms we deal with 12g is defined as a "hard landing," not a crash at all.

As for your figures, your RHS figure would be plausible if your fastener was a knive, but it's not. The LHS figure seems semi-plausible for a blunt body, but I would have expected some sort of crumpling or buckling of the material.

TTFN
faq731-376

 

[rb1957]

just covering the bases ...

9g is "all" we're required to cover ('cept if you're talking about 16g seats) as far as "items of mass in the cabin (that could impact passengers)".

it looked from your sketch as though there was a flange along the edge.

if the web "tore" from one fastener to the next, that could be a failure. you could start with an equal fastener shear load, calc how much the material between the holes can carry Fsy (or Fsu) *As. As = thk*p*2. an easy p is the spacing along the CL (conservative); more fussy is CL spacing + 2*0.2929*R ...

but not seeing the whole picture, i'd worry about where the load go once it gets into the web ?

 

[YoungTurk]

Catastrophic is as catastrophic does. Industry jargon would be an "event" or "emergency landing".

rb is stearing you on the right path. Determine the lowest allowables in the joint for shear and tension, then combine using Bruhn's equation or similar.

Nice sketch, but don't invent a new failure mode. Check shear bearing and tear out. As you drew it, there is no surface to tear out to, but you can call the adjacent hole an "edge" if you want to be conservative.

For the holes in a row, I would swag a tear out area of at least 2 times (Pitch - diameter) giving tearing strength in shear not less than (2 Fsu (Pitch - Diameter)).

Usually the harder value to get your arms around is the tension allowable for pull through. Bruhn and the McCombs supplement have fastener tension allowables for rivets that can give you a ballpark estimate for a bolt/screw of similar size. Also see the following paper:

http://www.lightmetalage.com/PDFs/LMA-2009-04-040_PullOver.pdf

 

[drawoh]

...

if the web "tore" from one fastener to the next, that could be a failure. you could start with an equal fastener shear load, calc how much the material between the holes can carry Fsy (or Fsu) *As. As = thk*p*2. an easy p is the spacing along the CL (conservative); more fussy is CL spacing + 2*0.2929*R ...

This is my middle sketch, and that is what I have analysed for. I am must wondering if there is a better failure mode. YoungTurk is correct. The fold at the bottom is the bottom of the box. The screw and nut would have to punch through that.

--
JHG

 

[rb1957]

can we step back a little ... you've got a "massive" load coming into the web. if the web failed by tear-out at one fastener assuming average load then this'd lead to a cascade failure. in reality this'd work with the web at the highest loaded fastener yielding, off-loading the fastener to the other fasteners; eventaully all would be yielded and then fail. so now we've got the thing providing the load to the web unattached at this location. now there are two scenarios ... either this thing is whistling down the cabin looking for a target or it remains in place with one less attachment. the point is local failure doesn't necessarily mean "death and destruction" (and jail time for the engineers).

now, why is this web so important ? how does the "massive" load get out of the web ? is shear tear-out the critical failure mechanism ? (it's just that we seem to be flogging this horse).

will the web fail in shear ? fsu*2*t*p (p = the distance between hole walls = pitch-diameter)

can the web support the load applied (in compression) ? inter-rivet buckling is a good enough model.

where does the load go next ? my concern is it goes out the edge flange, and flange bending is possibly the more critical failure mode.

 

[drawoh]

rb1957,

I am only trying to figure out the one mode at the moment.

I have analysed for a whole bunch of stuff, including the hex nuts ripping directly out through the wall. There is a lot of reinforcement around the screw holes. M6 nuts would have to rip through about 6mm of wall, moulding and gusset The moulding and gusset provide some support against sideways ripping, but not as much, I think. Ignoring them makes for a simpler and more conservative analysis.

--
JHG

 

[rb1957]

"fill y'r boots"

 

[drawoh]

rb1957,

Thanks.

By the way, I did consider the screws ripping sideways out the nearby edge. This problem is covered by Bickford and Nassar, and probably Bruhn, so I would not have posted the question. Also, there is a foot and a screw pattern at each corner of the box. Two of these would have had to tear all the way through the box. There is also a moulding all around the edges, but I do not know how it is attached. It might provide considerable strength.

--
JHG