Fatigue - Kt Factors for chem-milled steps

[SethMcKinley]

Hi folks,

I am doing some durability (fatigue / DTA) analysis on a lower wing skin replacement mod. for a 25000+ lb ex military aircraft.

Basically the mod. involves replacing the original skin panel (which consisted of a skin, and an internal doubler and tripler fastened together with spotwelds - a corrosion nightmare) with a single piece skin of identical geometry as the original skin panel - i.e. the chem-milling is used to cut the profile and depth of the original doubler and tripler.

I have some decent fatigue data for un-notched chem-milled specimens (i.e. the specimens were chem-milled evenly on both sides)
However, I don't have any Kt notch data specifically for chem-milled steps (and my regulator won't accept Peterson's machined notch data).
Does anyone know were I could get Kt values for chem-milled steps?? I would appreciate input from anyone that has experience qualifying a chem-milled skin.

I use Nasgro and Afgrow for DTA. My approach to analyzing crack growth along the chem-milled step is pretty rudimentary:
1. I use the surface crack model and the thinnest gauge at the step.
2. I factor up the stress spectrum based on my assumed Kt factor (which as I mentioned above, my regulator has not bought into). I realize this is conservative as it applies the Kt to the entire thickness, so I will probably refine this to get a better life.

I would appreciate any input on how best to approach this problem, and how it is typically handled in the industry.

Thanks for your response.

 

[MikeHalloran]

I wouldn't buy it either.

A long time ago, my then employer tried electolytically assisted grinding of (hardened steel) axle shafts. Productivity gains were amazing.

Unfortunately, the fatigue life went essentially to zero. I.e., the part's life wasn't reduced by any computable factor; it just lost _all of its fatigue resistance.

How about some nice big vacuum chucks and CNC mills?










Mike Halloran
Pembroke Pines, FL, USA

 

[israelkk]

MikeHalloran

I suspect the problem you had with the shaft was due to hydrogen embrittlement. Chemical etching or even EDM produces hydrogen that penetrate the metal and causes hydrogen embrittlement. This is detrimental for hardened parts (above 35-40RC). The etching is best done before heat treating.

As I understand SethMcKinley's skin is not hardened so it should be free of hydrogen embrittlement.

 

[rb1957]

back to the OP ...
i think this is a significant improvement in the design. I think there are enough process spec's out there to define a good procedure (but that'll be for the M&P to do).

i'd be a little hesitant about trying to define a Kt for the step, 'cause there's a bunch of other structure around the step that's helping. Rooke and Cartwright have a SIF solution for a crack approaching a step.

I'd suggest a simpler approach.
For long crack lengths (cracks wider than the step) lump the step area as a central stiffener (easy SIF solution) and you're left with a single thickness sheet.
For residaul strength, I'd neglect the step area completely ... in my experience integral pads aren't very effective (bonded/rivetted straps are very effective). I'd model the stringer and a constant thickness sheet.
For very short cracks (like at a rivet hole in the step) I'd apply the constant skin thickness stress (ie, assume the pad is ineffective in reducing the stress), and I'd only consider the pad thickness if I was growing a part thru thickness flaw. This comment applies to pads that are across the principal stress direction (like rib or frame pads); stringer pads (aligned to the principal stress) are effective.

any help ?

good luck

 

[Kwan]

Food for thought,

It would seem to me that you would be more critical at the fastener location due to load transfer(which I'm assuming is in the pad). For example, the skin connected to the spar will transfer your shear. This should have a higher Kt than the chem mill step.