Blacksburg gym collapse 1

[hokie66]

I haven't seen this posted on the site. Can anybody shed any light? It is difficult to tell how the roof was framed. A bulb tee is seen in one photo, so probably a Tectum deck or similar.

http://roanoke.com/multimedia/bhsgymcollapse

 

[Ron]

hokie...can't shed much light on it except I saw the bulb tees as well. Also saw what looked like pieces of formboard, so could be poured gypsum deck. Both Tectum and poured gypsum were common in that construction era.

Looks like a couple of feet of snow on the roof. Considering the way it sheared rather than sloughing off, it was likely pretty well packed.

Agree with Easterling on one count...you don't often see such full collapse. His other comments are a bit presumptuous.

I'll be surprised if they take the time necessary to determine a true progression of the failure. My sense is they want to get it cleared and reconstructed ASAP.

 

[hokie66]

I didn't actually agree with Easterling about the total collapse, but that would depend on the structural system used. The Tucker High gym in Richmond fell in a heap, but that was a thin shell hyperbolic paraboloid. Hartford coliseum, due to poor connection detailing all around the perimeter. This looks like a space frame or three chord trusses, but I really can't tell. It could also be a lateral stability issue. I hope they find out and let us all know.

 

[hokie66]

Found this report, for what it's worth. The whole emphasis seems to be on inadequate welding, poor connection design, and low strength steel in the truss. A bit hard to tell from the photos, but if we can assume there is a small SHS column built into the block pilaster like in the other picture taken in the auditorium of the same school, I imagine rotation of the truss/column joint has bent that column enough to destabilise the whole thing. The report of a falling block and the extensive cracking in the pilaster show that there was a lot of distress there prior to the truss failing.

http://www.mcps.org/newsreleases_bhs/FinalForconReport.pdf

 

[Ron]

hokie66...thanks for digging that one up. It's always good to get follow up on these things.

Now the "fun" starts. The defense will have a field day with the report. One statement in particular

The butt weld between the support and the horizontal member was not a full penetration weld, a violation of ANSI/AWS D1.1 for structural welds.

will create a world of grief for the writer when the depositions come about. I understand what he means, but the defense will use the statement on its surface to discredit his other statements. (He implies that only full penetration welds meet AWS requirements). Further, he leaves out one important example of a prequalified joint profile in his list of approved ones...the single bevel, perhaps the most common after a square butt.

The next argument will be whether it was the chicken or the egg that came first. Then they'll get into the argument about whether the 9% yield strength deficiency means anything, since the design stress was a lot lower than that. And on, and on, and on....!

The cell phone photos were of good quality!

 

[hokie66]

The photos certainly are of good quality. The coach must have a better cell phone than mine, and knows how to use it. I think those photos are the best evidence existing.

I wonder what legal recourse the school board has, as the gym was built 35 years ago. The contractor, steel fabricator, and structural engineer would be in the firing line. I think the contractor still exists, but don't know about the others.

What hit me first about the report was that the investigating engineer for the school board basically just brokered a testing company to do all the work, and depended entirely on the conclusion of the testing outfit, without any further explanation or discussion of the structural behaviour of the roof system.

Based on the coach's photos, I suggest the bottom chord, supposed to be a zero force member at that point, instead placed a couple on the tube column, yielding it as the roof deflected, thus making the end unstable, resulting in the masonry distress. I would like to examine the drawing for five minutes to see if my alternative theory makes sense. I'm not a metallurgist, but I believe a case could be made that the gross distortions in the seat were due to impact when the roof came down.

 

[dik]

Really interesting (in a Chinese sense) report... It'll take a couple of hours to scramble through it... I had difficulty getting meaningful information from it, initially.

There are no levels of stress indicated... or I missed it on first glance. I cannot determine from the report what the consequences of the inadequate 'butt' weld are or the lack of full penetration. I also could not determine the consequences of a low Fy... and if the issue is buckling, then, other than rotation, what the consequences of the above are.

I'll try to take a more thorough gander at the report over the weekend. On first glance, there may be a bit of a problem in defending the report, other than the obvious, that the roof came down.

Dik

 

[Ron]

hokie66 and dik...agree. Report is weak. Depending solely on the results of others and compiling a cover letter report will get this guy fried in litigation. Apparently there was a prior report that outlined the structural conclusions.

I also agree, hokie, that the distortions were post-failure rather than concurrent or causative. That makes all the effort (and cost) to explain the metallurgical and profile issues for naught.

If the defects are latent (which they appear to be), then they might have recourse against those responsible for the defect, unless Virginia has a statute of repose as well as a statute of limitations. The way that works in my state is that the clock starts running when the defect is discovered. You then have 4 years to bring that defect to a lawsuit(statute of limitations); however, if the defect is over 10 years old, time has run out (statute of repose), even if the statute of limitations has not.

 

[dik]

It could be more interesting if the cause of the failure and the defect were not related...

 

[Ron]

True...and that could be the case. It could be just a simple overload.

 

[JedClampett]

I was wondering the same thing as you. How do you differentiate between the initial failure and the subsequent damage?
As far the discrepancy in the yield strength, I agree it can over rated, but even in the 70's it was standard to get between 45 and 50 ksi for A36 steel. I know that because in the nuclear business we used to get mill certs all the time. If there was one thing you could count on, it was a minimum of 36 ksi for A36 steel. The chemical analysis didn't tell me anything. What is wrong with the steel? And couldn't they get a historical version of A36? Call ASTM for god's sake. It just seems sloppy.
It seemed unusual to have a mechanical and metallurgical analysis done (or at least by a ME), but no structural. For all I know the member with the low yield strength had no force in it. What's the calculated stress vs. the allowable?
I'd like to see a better report, not for the lawyers, but for a learning experience.

 

[Ron]

As stated in the final report, they had issued a structural assessment report on June 4, 2010. I found a copy of the report (attached).

The only conclusion the engineer makes is that the failure was caused by a non-overload condition and reaches no conclusion as to the probable cause of the collapse!

Was there deterioration or other conditions that led to the collapse?

Oh well...read for yourselves..

 

[JedClampett]

OK there was a report. It seemed as if they didn't want to dig too deep into the design. Eight inches of snow weighs 23 psf. What if there was drifting? The firemen said that there was snow and ice on the roof. Ice weighs about 57 pcf. Isn't it a coincidence that the failure happened under snow load condition?
So the code in the 70's required a 35 psf load. What stress level did that cause? Maybe the designer missed that load case and only designed for a reduced 20 psf LL (12 psf?)based on the area.

 

[hokie66]

Ron,
Kudos for finding the earlier "report". It would be interesting to see that photo 7, which he says shows the column bent by the bottom chord.

Jed,
You wouldn't reduce snow load in Blacksburg, Va. It really snows there. We had it drifted to 2nd story windows in the dorms in 1965.

 

[JedClampett]

My point was that if the snow load was applied incorrectly or missed altogether, the next controlling load case was the standard LL.
If it turned out that the maximum stresses were 7000 psi with 35 psf applied to the roof, you could move on to other causes. To blow off any design investigation because there have been greater snow accumulations in the past shows an amazing lack of curiosity. I'd like to give them the benefit of the doubt that they ran a calculation and it was obviously not the design. Plus they reroofed the gym in 1999. How did that roof weight compare to the old one?

 

[Ron]

Attached is photo 7. Doesn't particularly look like the description...appears the buckling was lower. Also appear to be burn marks on the column. Lots of electrical around....hmmm...could there be a connection of is it just another chicken or the egg conundrum?

Here's a link to the report and all the photos...

http://www.mcps.org/forcon/

 

[hokie66]

Thanks again, Ron. That photo doesn't correspond to the description in the report. But then that isn't surprising.

The columns were wide flange sections, not tubes as I thought. And there was nothing bracing the tops of the columns, if you don't count the unreinforced blockwork, which I don't.

I still haven't seen a photo showing the connection or lack of such between the bottom chord and the column. If that connection didn't allow rotation, I think it would have initiated the collapse.

This building obviously has had problems before, as evidenced by crack monitoring of walls and concrete beams, also permanent steel props.

 

[Ron]

Agree. I find their conclusions to be absurd and poorly supported. Obviously they had all the information that we don't have, but the physical indications that are presented in the photos don't seem to support the conclusions.

The final report relies on the metallurgy and physical testing which are irrelevant if the loading was not more than has been seen in the past, since there was no indication of a fatigue implication. They clearly do not believe that overload is the cause; yet how can they deny that as an implication...after all, if the snow had not been present, it isn't likely that the failure would have occurred. I suspect they have underestimated the actual loading at the time of failure.

Interesting!

 

[dik]

FWIW... In the several collapses I've looked at, and measured the snow loading at the time, there has only been one instance where the snow loading exceeded code design loads.

I made up a PVC pipe 'plug cutter' that I use to sample the depth of snow in various locations.

Without knowing what the overall framing was, the collapse photo appears to be an overload; if it were a local condition, then I would suspect there would be a redistribution of forces that would normally preclude a collapse... but, I'm not sure based on info presented.

Dik

 

[hokie66]

dik,
In this case, there were two trusses, spanning 150', which carried most of the roof. So when one of those became unstable at one end, down she came. There was still a lot of snow and ice on the fallen roof after the collapse, so I think they would have known if excessive snow buildup was the problem.

The investigators attributed the collapse to web crippling of the truss seat, which they just called buckling. That could be, as there was little bracing the top of the trusses at the end from twisting. Maybe a combination of web crippling with prying due to the bottom chord being in contact or connected to the column. The webs were certainly crippled, but it's anybody's guess as to whether that happened first or not.