Need a book? Engineering books recommendations...

Return to index: [Subject] [Thread] [Date] [Author]

GLB Failure

[Subject Prev][Subject Next][Thread Prev][Thread Next]

I had a similar situation several years ago in which GLB's in a panelized 
roof industrial building failed.  The night before there was a  rain storm in 
the vicinity and the contractor/developer/owner wanted to blame the collapse 
on a lightning strike (I wonder why!).  The engineer on the project was a 
civil does-it-all engineer.  A quick preliminary check indicated that the 
glue lams were considerably undersized, however, since the failure was 
localized in a rather large building, it seemed that this in and of itself 
probably was not what actually precipitated the collapse.

After determining which GLB probably failed first, the failure location was 
examined.  The bottom lamination had a finger joint that, for the most part, 
pulled apart; the next lamination had a knot; and there was another finger 
joint in the third lamination which also failed by partially pulling apart.  
Here was an undersized beam that had three strength reducing characteristics 
in the three most critical laminations at the most highly stressed location 
in the span.

My conclusions were that the beam failed because it was undersized *and* had 
these strength reducing characteristics concentrated in the most critical 
location in the beam.  I could not say, one way or another, whether the beam 
would or would not have failed if either of these conditions did not exist.

In your description of your failure, once the bottom lamination finger joint 
failed, that lamination started to behave like the leaf of a spring in your 
car and had to be held to the other laminations by tension.  Since neither 
glue nor wood perpendicular to the grain has any significant strength in 
tension, this is where the horizontal and sloping cracks would occur and 
since the grain would slope upward on one side of failure location, the 
effective beam section would get smaller and smaller as well as having lower 
quality lumber.

Frequently kitchen areas, at least here in Arizona, are cooled with large 
evaporative coolers, I was wondering if there was one installed on or near 
the failed beam?  Together with saturated pads and water in the reservoir, 
evaporative coolers can weigh up to 1100 pounds.  Stove hoods over commercial 
stoves are no lightweights either.

Hope this gives you some ideas to look at.

A. Roger Turk, P.E.(Structural)
Tucson, Arizona

Bruce Resnick wrote:

. > To all:
. > 
. > A question regarding possibile causes of the failure of a glulam beam:
. > 
. > At the request of a client, the other day I looked at a glulam beam which 
. > had failed in the roof of a one-story building. (Fortunately, we did not 
. > design this.) The building is about 20 years old, and the beam appears to 
. > be about 5-1/8" x 24". The beam is about 42' long supporting a panelized 
. > wood roof system.  There is nothing of note weight-wise on the roof.
. > 
. > The beam has "broken" at about midspan and is now temporarily shored.  The
. > finger joint in the lowest lamination has separated about 3", and there 
. > are 4 or 5 other major cracks in the beam at the failure point.  The 
. > cracks are primarliy horizontal but are diagonal enough to be extending 
. > through the laminations not along the glue lines. The beam has sagged 
. > about 6" due to the failure and was probably hanging from the roof until 
. > it was shored. The beam is over a kitchen area.  Reportedly, someone 
. > heard a "pop" and noticed the ceiling was sagging. The failure was 
. > noticed when the ceiling was opened to inspect the beam.
. > 
. > We are recommending that the beam be removed and replaced from above, 
. > which the owner is willing to do.
. > 
. > However, I am curious as to possible failure causes.  There is no unusual
. > equipment on the roof and the building engineer says they have no ponding
. > problems.  Could the failure be related to heat or gases from the kitchen?
. > Could it be a latent defect?
. > 
. > Any thoughts and/or similar experiences would be appreciated.
. >