Thankyou for your time Christopher. Below I've attempted to clarify a few
points you have raised for your benefit and those interested others on the
> -----Original Message-----
> From: Christopher Wright [mailto:chrisw(--nospam--at)skypoint.com]
> Sent: Wednesday, 23 February 2000 12:04 PM
> To: yenem(--nospam--at)iinet.net.au
> Subject: Re: Urgent help required regarding combined stresses above yield
> Your methodology is a little confusing. First, if you're figuring the
> stress you should be using the elastic section modulus.
If the section is yielding under the actual modelled loads, then plastic
design is appropriate isn't it? I thought the elastic section modulus is
only valid where the extreme fibre stress is at or below yield.
Second it's not
> clear how the stub is loaded. You need a free body diagram because it
> certainly isn't obvious where all this loading is coming from.
The free body diagram would consist of a shear force at the base of the stub
(ie the horizontal reaction at the support), an opposing shear force at the
top of the stub but offset out of the plane creating torsion in addition to
bending stresses, and a tension force due to the stub being in tension (the
vertical reaction at the support).
> certain of the loading (actual loading including everything at the time
> of failure, not just service loading.
Yep. Failure occurred once the stacker was erected. No service loads had a
chance to be applied.
> You're right that the detail isn't a good one, but the stress you should
> start thinking about a failure mechanism. If the item fail the first time
> it was used, then yielding is something to consider, although local
> stability is also a candidate. If the item was in service a long time,
> you should be asking why it didn't fail sooner. So you have to ask
> yourself about fatigue. Along those lines you need to look closely at the
> actual failure. Did it yield or is there evidence of fracture?
Yes. We've done plenty of thinking about a failure mechanism. Fatigue has
certainly been ruled out - the stub simply was undersized for the actual
loads (dead load) let alone the design loads. We are certain that this
caused the failure - the issue is whether failure would have occurred if a
stiffener was in place to improve the section capacity.
> You need to separate code compliance from yielding. Code compliance is
> based on some sort of assumption of ultimate loads. You need the address
> the loads that existed at the time of failure. If you're suggesting a
> design detail that wouldn't meet Code as a remedial measure, you're
> wasting your time. Opposing counsel will eat you alive.
I only mentioned code non-compliance to prevent a response such as "the
detail doesn't appear to meet the design standard and therefore is no good!"
> And ultimate
> strength has no bearing on bending failure which is governed by the
> establishment of a plastic hinge. Yield stress governs plastic failure.
You've made a good point here. At yield, the top of the stub would develop
a hinge - not wait until the stress reached ultimate strength.
Yenem Engineering Services
54 John Street
Gooseberry Hill WA 6076