To: "'seaint(--nospam--at)seaint.org'" <seaint(--nospam--at)seaint.org>
Subject: RE: QUERY: Fatigue
From: Paul Pilarski <PPilarski(--nospam--at)HNTB.com>
Date: Mon, 19 Mar 2001 10:49:34 -0600
I confer with Mr. Hariharan's comments on fatigue. My fatigue
background came from MS research on ship hull fatigue, including prediction
and half-scale experimentation of long crack propagation.
There is nothing wrong with reasserting your opinion if you feel
another's may be misleading. Case in point, using a threshold limit (I'm
also more familiar with this term than the term, endurance limit) of 16ksi
is more rational for the case discussed than taking a theoretical value of
1/2 Fu unless one really wishes to gamble with the variables of fatigue
scatter. Realistically, the variables surrounding fatigue and crack
generation create such scatter that to use anything other than the lower
bound is not good practice.
Building fatigue is low-cycle fatigue which may be assessed with
proper details and stress concentration identification. I believe it is a
seperate animal that does constitute a different learning curve over the
offshore industries and ship building industries.
From: Christopher Wright [mailto:chrisw(--nospam--at)skypoint.com]
Sent: Saturday, March 17, 2001 10:05 AM
Subject: Re: QUERY: Fatigue
>Sharpshooter! (in jest!)
I'm sure there's another term for someone who stands off and potshots
other people's opinions, but I doubt it's as graphic. And again, I'm
reluctant to give even the appearace of such, since I've been on the
receiving end of a few. Your comments are well taken.
>E. "You have taken care of fatigue" was intended to mean that the fatigue
>requirements have been met. The stress limitation of 16 ksi for the type of
>weld as per AISC for the number of cycles, and 0.5 Fy (endurance limit) in
>the base metal at other points.
My reply was meant to caution that there's nothing magic about 16 ksi.
There are plenty of welding details which might be considered for a
fabbed T-section where the 16 ksi limit would be iffy at best. I don't
have any particular problem with using 50% of the yield strength as a
fatigue limit, since it's a conservative estimate for situations without
great stress concentration. Given the wide variety and quality of welded
details in building construction, using1/2 Fy as the fatigue limit would
range from optimistic to foolhardy as a rule of thumb. As an example the
base metal at stitch welded connections would have a stress category of E
or F in many instances and have an allowable well below 16 ksi.
>There is a lot more to
>Fatigue than what has been discussed, but that may not be of immediate
>interest to the forum.
It probably should be. I suspect building design is hell bent on learning
fatigue and fracture in weldments the hard way. Pressure vessels (where I
learned about fatigue) and offshore structures and ship building before
us all seemed to have their own separate learning curves. Building design
has been living with structures where dead load pretty much dominates
things, and fatigue is still pretty much an afterthought. I don't think
the emphasis on limit design and 'prequalified' weld joints is helping
much. The moment connection failures in the Northridge earthquake were a
wake-up call in the area of weld quality--so we'll see if it takes.
Christopher Wright P.E. |"They couldn't hit an elephant from
chrisw(--nospam--at)skypoint.com | this distance" (last words of Gen.
___________________________| John Sedgwick, Spotsylvania 1864)