dianne_1234 wrote:
> On Sun, 14 Mar 2004 18:35:06 GMT, jim beam <
[email protected]>
> replied to Jobst Brandt thusly:
>
>
>>you'll find that a microhardness test at that deformation
>>point yields a harder material than the region that's not
>>yielded.
>
>
> Have you really done this? Can you actually get a hardness
> tester in there? Is hardness a reliable measure of fatigue
> strength in aluminum anyway?
no, you wouldn't typically do a hardness test on a curved
surface. what i'm trying to say is that if you yield a
material like this, it does not automatically make it
weaker. hardness /is/ related to strength.
in ascii
S ~
| ~ ~ / 2 ~ / / 1 / / / / / / /
______________________ s
if the material starts yielding at point 1 and plastically
yields to point 2, any reapplication of stress at that
location will proceed immediately up to point 2, it does not
go through point 1 again.
no, hardness is not a measure of fatigue strength - but it
is a measure of work hardening and therefore "strength",
which can then affect fatigue.
>
>
>>a manufacturer being squirrelly is no basis for a fatigue
>>crack growth hypothesis.
>
>
> Right.
>
> But most rebuilds are in the direction of easier tear-out:
> from, say, 3x to, say, radial. And this may orient any
> damage from the original pattern at roughly right angles
> to the new direction of load.
i should have been more specific - i'm talking about a
rebuild with the same cross pattern, not radial. radial is
always much more stressful for a hub, whether it be new or
a rebuild.
>
>
>>if a crack has initiated in one orientation, perpendicular
>>to applied stress,
>
>
> Yes, usually a great textbook example for the case of pure
> tensile fatigue.
>
> But is this really a frequent load case in hub flange
> failures? In my experience, most hub flange failures are
> from one hole to the next, until a whole crescent piece of
> the flange separates from the rest of the shell.
like this:
http://technology.open.ac.uk/materials/mem/mem-
ccf1.html
they go in pairs, but if you re-orient to the next flange
pair along, [if you see what i mean] you should be able to
dodge the fatigue bullet longer than you would rebuilding
exactly as before.
>
>
>>and the hub is rebuilt using that same orientation, that
>>same crack is going to continue to grow. how can you argue
>>against that?
>
>
> One counter argument might be that as the crack grows, the
> new geometry of the remaining good material changes the
> stress state because the features are different now (what
> was once connected is no longer, so the load path may
> change).
if you went from 3x to radial, yes, you'd still get
cracking because the load paths will share every other
flange pair and continue cracking, but from one 3x to
another 3x oriented perpendicular to how it was before,
pretty unlikely.
>
>
>>if it's rebuilt at 90 degrees to the original, i.e. those
>>old cracks are now axial with newly applied stress, those
>>cracks are not oriented for growth, so any fatigue process
>>has to be newly initiated does it not?
>
>
> Seems so, but I wonder if an existing crack has enough
> stress in shear to grow even with the new orientation of
> the tension?
>
>
you really don't see fatigue cracking in a true shear mode.
something like a coil spring has shear loads, but cracks
open perpendicular to stress, not along a shear axis.