radius used for stress relief

huangyhg

radius' used for stress relief
please refer to the following picture...
fig 1a and b is a typical example to reduce stress at its highest point (a sharp pronouced corner) by introducing a radius or relief.
my question is if you have a pin canilevered through a hole (fig 2a), and you round the edge of the hole (fig2b) i dont see this having any affect in reducing stress.
i basically see the cross-section where it should fail shifted back tangent to the end of the radius. (see red line in fig's 2).
am i wrong? please share your thoughts because our fea calculations state otherwise. the stress was significantly reduced from 460mpa to 300mpa from fig 2a to 2b respectively.
the only way i see this helping is if the radius was large enough, it might provide some support after small deflection.
thanks,
j
quote:
the only way i see this helping is if the radius was large enough, it might provide some support after small deflection.
that was my first thought as well.

i pulled out my old "kent's mechanical engineer's handbook - 12th edition" to see what they have to say. here are excerpts from pages 8-45 and 8-46:
quote:
the effect of the localized increase in stress, such as that caused by a notch, fillet, hole, or similar stress raiser, depends mainly on the type of loading, the geometry of the part, and the material...  for ductile materials it is not customary to apply, stress correction factors to   
i have something very similar to 1a and 1b
the safety factor reduces when i add a 1mm radius to a sharp angle.
it's about the same when the radius reaches about 5 mm.
how can that be. i'm using inventor fea.
regards
warren.

any thoughts about the fea is not giving you the right answer? it has been years since i did any fea analysis, but the computer loved giving me infinitely large stresses, which i was equaly happy to ignore most of the time.
in real life there might be some plastification in the corner and stress redistribution, with the conssequence that stresses are much less acute than those given by the computer model.
in my humble oppinion, the difference between the two models is a moot point.
slideruleera strongly hints at my thoughts:
- at the "macro" scale 1b is an improvement and 2b is deterimental, shifting the support point exactly as the josh2008 stated
- at the "micro" scale the effects may be very important, but this should only ever matter to someone working at very small scale with correspondingly finite tolerances.
cheers,
ys
b.eng (carleton)
working in new zealand, thinking of my snow covered home...