Stress Concen. of Stepped Shaft in Tension, r/d=0 2

[kevindurette]

We had some parts made, and the supplier didn't follow the print. We're trying to decide whether we can still use them to get out of a bind.

We have a stepped shaft loaded in tension where there is no fillet radius. It's made of induction-hardened chrome-plated steel. (I can give more specs on the material if necessary.) I don't believe the stub end is chrome-plated or case hardened.

D = rod diameter = 3 inches (and 24" overall length)
d = diameter of end stub = 1.25 inches (let's say about 4" long)
r = fillet radius =~ 0 to the naked eye
sigma yield = 100 ksi minimum

In the literature, the Kt stress concentration factor seems to scale inversely with r/d, but common sense only tells me that this value cannot be infinite. If that were true, the end stub would have already fallen off under its own weight, or it would have at least been torn off when I picked up the shaft by it. I'm willing to bet elastic and plastic deformation limits this value.

What maximum Kt factor should I use? (This will be before a safety factor is also applied.)

Thanks.

Durette

 

[GregLocock]

Test one

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[rb1957]

are you worried about fatigue life or static strength ?

if static strength, then Kt Shouldn't be a problem, there'll be local yielding at the shoulder, but 'cause you're using a ductile steel, yield = 100 ksi, i doubt it'll be a problem.

if fatgiue, i doubt you can make it work, out of the box. if you have to use them, shotpeen the shoulder. maybe undercut the shoulder ... grind a semi-circular grove into the face of the shoulder, as big a radius as you can.

have you paid for the pieces ? going after the manufacturers (for not following the drawings) ? costs of loss of business ?? this goes if you use the parts; but then this thread is at least "due diligence".

good luck !

 

[corus]

The value of Kt is infinite for a zero radius hole, by commonsense alone. It doesn't fall off when you pick it up though as the 'infinite' stress is limited to a small localised region. If you continue to pick it up and put it down though, it will eventually fall off as cracks develop and grow.

ex-corus (semi-detached)

 

[kevindurette]

If a manager is forcing me to "rate" the product, what value should I use to be realistic while still being conservative?

 

[rb1957]

still is your concern static or fatigue ...

what is the nominal stress (at the root of the shoulder) ?

is it predominately bending or axial ?

 

[flash3780]

Your stepped shaft does, in fact, have a fillet radius (albeit small). You could reduce the stress by adding an undercut if it is a load-bearing part.

 

[TVP]

Kevin,

I work with these kinds of components every day. As the others have mentioned, it is critical to understand if there are cyclic loads, as this type of part will almost certainly fail prematurely due to fatigue if there is no radius at that transition. Also, the concentricity and run out of the thread to the rest of the shaft will affect the stress in that section, so I recommend inspecting these values as well.

If you cannot introduce a radius into the face/shoulder of the rod as mentioned by rb1957, then I would not use the parts for cyclic loading applications. Another option to consider is roller burnishing the transition, which will induce subsurface compressive stresses and thus delay fatigue crack initiation and growth.

 

[kevindurette]

Thank you everyone for the help everyone. I put in my recommendation, and the issue is out of my hands now...