Need a book? Engineering books recommendations...

# RE: lateral torsional buckling really happens!

• To: <seaint(--nospam--at)seaint.org>
• Subject: RE: lateral torsional buckling really happens!
• Date: Thu, 4 Jan 2007 15:52:32 +1030

```Thanks Josh.

Looks like I may have some reading to do. The Australian cold-formed
structures code, requires Cb=1, for combined actions (bending and axial),
though some cold-formed shed designers here seemed to have missed this.

Taking the segment between inflexion points under such conditions, is
probably even less desirable. If I understand correctly the theory is based
on a single span simply supported beam subject to a uniform moment. The
point of Cb is that most beams have a distributed moment, and therefore that
part of the beam which is not fully stressed provides some assistance to
that which is. But if axial compression is present as well that is no longer
true, hence Cb=1.

Looks like I need to check how well the codes separate:

1) Flexural-Lateral Buckling and
2) Flexural-Torsional Buckling

My understanding is that the flexural-lateral buckling is a consequence of
compressive stresses in the plate elements. Whilst flexural-torsional
buckling is a consequence of the section shear-centre being offset from the
centroid.

Doubly symmetric sections like I-beams, have shear-centres coincident with
the centroid and are therefore less susceptible to twisting. Whilst
cold-formed c-sections, and hot-rolled channels have shear-centres offset
from the web and external to the section. Load applied to the flange is
therefore eccentric and produces a twisting moment. For the I-beam it is
mainly manufacturing defects responsible for eccentricities that produce
section twisting.

Which leaves me wondering whether the inflexion point defines a valid
segment for lateral-buckling, but is invalid for torsional-buckling.

Then to complicate matters further, if we get it all restrained we have the
problem of distortional-buckling.

Regards
B.Tech (mfg & mech), MIIE, gradTIEAust
mailto:sch.tectonic(--nospam--at)bigpond.com
Roy Harrison & Associates
Consulting Engineers (Structural)
PO Box 104
Para Hills
SA 5096
South Australia
tel: 8395 2177
fax: 8395 8477

-----Original Message-----
From: Josh Plummer [mailto:josh.plummer(--nospam--at)cox.net]
Sent: Thursday, 4 January 2007 14:12
To: seaint(--nospam--at)seaint.org
Subject: RE: lateral torsional buckling really happens!

Yes, it is an explicit rejection of the inflection point. Not by me, but
rather by AISC. From Appendix 6 of the 2005 AISC specification:

"Beam bracing must prevent twist of the section, not lateral
displacement....
Lateral bracing systems that are attached near the beam centroid are
ineffective....
The inflection point can not be considered a brace point because twist
occurs at that point (Galambos, 1998).....
The beam brace requirements are based on the recommendations in Yura
(1993)."

Now, I have heard some say that if you use the inflection point, but assume
a Cb of 1.0 then you should be okay.  I haven't run the numbers, but if it
works out to a similar capacity then the design these folks have been doing
should be fine and they don't have much to worry about.  But, to me, that's
still sloppy thinking.

I guess I'm coming down on AISC's side on this one.  To me this is all a
stability issue.  A cantilevered column has a longer unbraced length than
the column itself.  But for some reason a lot of folks get bent out of shape
regarding a similar concept with LTB buckling.  Go figure!

Josh Plummer, SE

-----Original Message-----
Sent: Wednesday, January 03, 2007 4:38 PM
To: seaint(--nospam--at)seaint.org
Subject: RE: seaint Digest for 2 Jan 2007

Josh

Is that an explicit rejection of the inflexion point by AISC. I would be
interested to know titles of such papers.

Here in Australia our steel structures code (AS4100) only uses physical
restraints to mark the ends of segments being checked. But our cold-formed
steel structures code (AS4600) which is supposedly based on the AISI
specification uses the inflexion point to define a segment being checked.

Other than the codes account for the distribution of moment slightly
differently, I've never understood why the inflexion point is acceptable for
sections more prone to flexural-lateral-buckling and
flexural-torsional-buckling, and seemingly not acceptable for sections less
prone.

Clearly at the inflexion point, a compression flange has changed to a
tension flange, so why would the plate buckling length (for flange and
portion of web) be any greater than the length of such segment? Depending on
the assumed end conditions for the column analogy the effective length may
be much greater than the segment. The inflexion point is not however
considered a point of torsion restraint.

So typically have lengths:
Lx[XX axis - strong] = span
Ly[YY axis - weak] = length of segment between inflexion points and or
physical restraints.
Lz[ZZ axis - torsion] = span

Though some references suggest no reason that Lz shouldn't equal the smaller
Ly value. I don't adopt that approach if I have to set Lz=Ly then, I provide
a fly-brace to provide torsional restraint.

Some references also suggest that girts/purlins provide full restraint and
therefore imply only need to check the un-restrained inner flange. That also
seems poor practice, since it may not hold true, and depending on how the
girt/purlins are connected, they may not provide any torsional restraint.

It as always seemed to me that the adequacy of restraint, is a lot more
subjective than it ought to be.

Regards
B.Tech (mfg & mech), MIIE, gradTIEAust
mailto:sch.tectonic(--nospam--at)bigpond.com
Roy Harrison & Associates
Consulting Engineers (Structural)
PO Box 104
Para Hills
SA 5096
South Australia
tel: 8395 2177
fax: 8395 8477

-----Original Message-----
From: Josh Plummer [mailto:joshp(--nospam--at)risatech.com]
Sent: Thursday, 4 January 2007 03:28
To: seaint(--nospam--at)seaint.org
Subject: RE: seaint Digest for 2 Jan 2007

Andrew -

I'm curious about one other thing related to this failure investigation....
AISC has claimed for years that the inflection point should not be used as a
"brace point" for the LTB of a beam.  If those pictures demonstrate WHY you
don't want to use the inflection point for bracing, then count me as another
person who would love to see them.  To me, that's a much more effective
argument than anything they say in the AISC spec, commentary or seminar
series.

For what it's worth, most of the engineers that I talk to (outside of heavy
hurricane or seismic regions) seem to scoff at me when I tell them that it
may not be a good idea to consider the inflection point as bracing.  They
come back with the same old lines that contractors use, "I been doing these
structures for 30 years and I always use the inflection point as bracing and

Since AISC is getting more and more of a reputation as being ruled by ivory
tower academics, it'd be nice to be able to show these guys a real world
example of why you should or shouldn't make these sorts of assumptions.

Sincerely,

Josh Plummer, SE

RISA Technologies
joshp(--nospam--at)risatech.com

--------------------------------------------------------------------------
From: "Andrew Kester, PE" <akester(--nospam--at)cfl.rr.com>
To: <seaint(--nospam--at)seaint.org>
Subject: lateral torsional buckling really happens!

I just got back from looking at a tornado damaged PEMB in Daytona Beach, =
FL. It happened on Christmas Day, go figure, it is Florida... We also =
practically had to run the AC, Scott in Michigan would like to know.

It tore off the wood framed and metal deck mansard canopy on the east =
end, but did not damage most of the rest of the walls and canopies. I =
went on the roof carefully while I waited for the owner to arrive and =
unlock the door. Parts of the roof were buckled upward and other =
downwards, in excess of 2-3 feet. It was really amazing, I have never =
seen anything like it. Not a piece of deck was missing though some had =
been torn in half due to the partial collapse. There were also some =
projectile holes in the deck.

Once inside, several of the 3 foot deep beams as part of the moment =
frame had failed in several places in what to me seems bottom flange =
compression failure. They were not on the ground but they were =
deflecting and were severely distorted. Many of the purlins were now =
horizontal as they had failed in lateral buckling I would assume.  I did =
see some bottom flange beam bracing along some of the beams. One of the =
worse areas of deformation was near where a beam failed at the column =
connection, and these were the deep tapered beams that are deeper at the =
columns then got less deep towards the center. In another part of the =
building the drop ceiling was still intact. 30 yards away a single story =
shingle roof was completely intact. Two blocks away an apartment complex =
had several buildings unscathed, but in the center of the complex it =
looked like bomb was dropped on a couple of the buildings. Tornados are =
definitely random, much more so then hurricanes, so it seems in my =
forensic experience.

I design canopies and roof beams in high uplift areas and always have to =
remember to check them in uplift due to the lack of compression flange =
bracing, and will usually just upsize them a little to handle that. Of =
course you can put bottom flange bracing but that is usually more work =
and expensive then a slightly bigger beam.But in my head I always think =
there is no way that would ever be the failure mode, the deck would get =
ripped off or the light gage framing before the load would ever be =
extreme enough to cause that. I have been wrong.

Andrew

Andrew Kester, PE
Lake Mary, FL

******* ****** ******* ******** ******* ******* ******* ***
*   Read list FAQ at: http://www.seaint.org/list_FAQ.asp
*
*   This email was sent to you via Structural Engineers
*   Association of Southern California (SEAOSC) server. To
*   subscribe (no fee) or UnSubscribe, please go to:
*
*   http://www.seaint.org/sealist1.asp
*
*   Questions to seaint-ad(--nospam--at)seaint.org. Remember, any email you
*   send to the list is public domain and may be re-posted
*   site at: http://www.seaint.org
******* ****** ****** ****** ******* ****** ****** ********

******* ****** ******* ******** ******* ******* ******* ***
*   Read list FAQ at: http://www.seaint.org/list_FAQ.asp
*
*   This email was sent to you via Structural Engineers
*   Association of Southern California (SEAOSC) server. To
*   subscribe (no fee) or UnSubscribe, please go to:
*
*   http://www.seaint.org/sealist1.asp
*
*   Questions to seaint-ad(--nospam--at)seaint.org. Remember, any email you
*   send to the list is public domain and may be re-posted