Eng-Tips is the largest engineering community on the Internet
Intelligent Work Forums for Engineering Professionals
Newly Constructed Gym Has Roof Collapse in New Mexico 12
- Thread starter jerseyshore
- Start date
- Status
I also don't think the pilaster was the problem. However, the depth of the block/pilaster affects the available bearing surface and degree of eccentricity. Given the construction method, the bearing surface is limited to less than half the pilaster depth.
- Thread starter
- #163
jerseyshore
Structural
Did anyone submit a public records request yet for the plans with the City of Albuquerque? This thread is going to re-engineer this whole structure 10x over before we get any new info I bet.
Permit plans would be nice, however no guarantee they match as built. We also need approved shop drawings for the truss.
They are keeping quiet on this whole thing, which smells like all parties involved are concerned about their role in this.
Everybody hunkering down in preparation for litigation. Without whistle blower we may not get any more information for a long time.
They are keeping quiet on this whole thing, which smells like all parties involved are concerned about their role in this.
Everybody hunkering down in preparation for litigation. Without whistle blower we may not get any more information for a long time.
TheGreenLama
Structural
What is this? Someone above mentioned the roof panels may have fallen on a construction lift. Could it be something else? From another angle this (steel) piece is inclined at a large angle from the vertical, so it wouldn't have been a vertical brace. Could it have been a steel beam, added later when unusual deflections were noticed, connecting the bottom chords of the trusses together to provide lateral support? The location looks to be exactly at midspan.
From a construction/build standpoint alone, I don't see how you can set the hollow core concrete roof planks without some temporary lateral bracing for the two long span trusses. You would have to set them in the correct install pattern sequence, to avoid 'tipping' the truss. Once truss is tipped out of plane, oh no!
So if lateral bracing likely required for installation, and hollow core deck ends, which appear to be a friction float on the top chord of truss, would 'tip' with any load differential between opposing ends of hollow core planks.
It appears GreenLama is onto something, and a lateral mid-span brace may have been added for construction or resolution attempt after failed framing inspection?
The two steel cables showing thru windows on parapet wall is odd as well?
Classic Example of where a Good System Engineer Construction Manager working solely for the Customer paying the bills is employed to catch the disconnects between design throwing it over wall to GC, and GC then has to figure out how to actually assemble the design safely, and in proper order.
A Great System Engineer/Construction Manager, is money well spent. Now choosing poorly can also WRECK a project. I have personal knowledge of Practicing PE Structural Engineer Construction Manager, that saved his customer lots of $$$ over catching costly design flaws and build-ability issues of design. And this was huge Manufacturing Campus from ground up.
Edit: The kicker is he makes more money as CM than he does doing design work for himself. Now he is unusual person that has design expertise and vast Project Management Field Experience. So not every DEI hire has this level of talent, skills and experience.
So if lateral bracing likely required for installation, and hollow core deck ends, which appear to be a friction float on the top chord of truss, would 'tip' with any load differential between opposing ends of hollow core planks.
It appears GreenLama is onto something, and a lateral mid-span brace may have been added for construction or resolution attempt after failed framing inspection?
The two steel cables showing thru windows on parapet wall is odd as well?
Classic Example of where a Good System Engineer Construction Manager working solely for the Customer paying the bills is employed to catch the disconnects between design throwing it over wall to GC, and GC then has to figure out how to actually assemble the design safely, and in proper order.
A Great System Engineer/Construction Manager, is money well spent. Now choosing poorly can also WRECK a project. I have personal knowledge of Practicing PE Structural Engineer Construction Manager, that saved his customer lots of $$$ over catching costly design flaws and build-ability issues of design. And this was huge Manufacturing Campus from ground up.
Edit: The kicker is he makes more money as CM than he does doing design work for himself. Now he is unusual person that has design expertise and vast Project Management Field Experience. So not every DEI hire has this level of talent, skills and experience.
TheGreenLama
Structural
SymPIe, I'm not saying this would have been a good idea. The slenderness ratio would have been huge. And, since both trusses are identical, have identical loads, and identical boundary conditions, if the primary buckling mode pushes center span point out, you would just get both trusses to buckle simultaneously.
The report you noted above says they noted a dipping of the truss. No mention of a horizontal component. But that doesn't mean there wasn't any.
If this is a retrofit, it reminds me of the FIU bridge collapse. They have a nearly completed structure that suddenly shows signs of unusual movements. So, to try and salvage it, and keep to construction schedule, they go in and start trying things, without fully understanding the issues. All this does is put construction workers lives at risk.
There's a lot of speculating here. The way the roof panels fell it just appears that they break across the midspan, implying something substantial running beneath. You can envision a case where the two trusses start moving towards each other at the midpoint, and someone says, Hey, all we have to do is block them somehow.
The report you noted above says they noted a dipping of the truss. No mention of a horizontal component. But that doesn't mean there wasn't any.
If this is a retrofit, it reminds me of the FIU bridge collapse. They have a nearly completed structure that suddenly shows signs of unusual movements. So, to try and salvage it, and keep to construction schedule, they go in and start trying things, without fully understanding the issues. All this does is put construction workers lives at risk.
There's a lot of speculating here. The way the roof panels fell it just appears that they break across the midspan, implying something substantial running beneath. You can envision a case where the two trusses start moving towards each other at the midpoint, and someone says, Hey, all we have to do is block them somehow.
Based on the assembly video, it's not clear if the bearing engagement on each protruding pilaster was equal, and this may have influenced the failure mode. After engaging the left support, the free end of the truss sweeps just inboard of the other, protruding pilaster. The truss must then be moved longitudinally to engage that bearing surface.
I believe this questionable truss support is the one that lost engagement at the time of the collapse, while the other end was still supported. Check out time interval from 0.20 to 0.25s
I don't think this illustration of a truss end support can be relied on, as the actual one is a proper, protruding pilaster.
Kevin Kelleher, P.E. (retired)
Internal Mechanical Eng'g Consultant
DuPont ESD Specialists
I believe this questionable truss support is the one that lost engagement at the time of the collapse, while the other end was still supported. Check out time interval from 0.20 to 0.25s
I don't think this illustration of a truss end support can be relied on, as the actual one is a proper, protruding pilaster.
Kevin Kelleher, P.E. (retired)
Internal Mechanical Eng'g Consultant
DuPont ESD Specialists
SWComposites
Aerospace
Yeah, but the FEM showed positive margins.
In addition to video not showing adjusting bearing depth on last end of truss, how much of the bearing margin depth was lost by swinging trusses into pockets with parapet wall already built, versus dropping trusses in from above before parapet wall built on top of bond beam?
Pocket had unused bearing depth closer to centroid than swinging technique allows utilization of.
Pockets are common on tilt-up reinforced concrete wall panels, however those walls are stronger and more durable than spliced rebar stuffed into grout filled cinder block cores. Especially outer face of cinder block.
Pocket had unused bearing depth closer to centroid than swinging technique allows utilization of.
Pockets are common on tilt-up reinforced concrete wall panels, however those walls are stronger and more durable than spliced rebar stuffed into grout filled cinder block cores. Especially outer face of cinder block.
LittleInch said:
Is there a real time video available? May provide more details of what appears to be steel lined openings for the pilasters, and how the workers were using fasteners? to secure the end of the truss.
Kevin Kelleher, P.E. (retired)
Internal Mechanical Eng'g Consultant
DuPont ESD Specialists
I have seen the full 31 second video on Facebook, and my eyes don't see a steel lining on pocket, and I offer up the fact that they enlarged the pockets during erection. Those images were posted above and I will post my adding arrow to one of those pictures below, showing they chiseled out the side of pocket to enlarge it so truss would swing into place. Later images show the pocket growing even more than what is pictured here..... They did not chisel thru steel, IMHO.
I wonder if the original design featured the truss top chord extending into the wall, which might have been shortened later for constructability reasons. It's hard to imagine a designer intentionally allowing the truss to bear only on the outer edge. It's plausible that the designer might have assumed the trusses would be set into pockets, without realising that the builder was going to construct the walls in one go, preventing full length trusses from being dropped in.
We've definitely encountered scenarios where the shop drawings include subtle modifications for easier construction - things shorted a little, etc - and these changes go unnoticed by those reviewing and approving the shop drawings. Then you arrive on site and realise there's a potential problem. Sometimes, these changes are significant and aren't caught by anyone (eg Hyatt walkway collapse)
Of course, this is just speculation, but it's a scenario I've seen play out based on my experience.
We've definitely encountered scenarios where the shop drawings include subtle modifications for easier construction - things shorted a little, etc - and these changes go unnoticed by those reviewing and approving the shop drawings. Then you arrive on site and realise there's a potential problem. Sometimes, these changes are significant and aren't caught by anyone (eg Hyatt walkway collapse)
Of course, this is just speculation, but it's a scenario I've seen play out based on my experience.
My thinking is steel trusses were very long lead item, and the GC did not want to have to pay mason to re-mobilize for parapet wall as separate step, and did exactly what you have seen, and school was pressing to compress schedule so they bought in on what may have been a read bad decision?
Bearing plate should be sufficient, IMHO.
Bearing plate should be sufficient, IMHO.
Tomfh said:
It does seem like the top cord was initially too long to fit into the original pockets, based on the left pocket apparently being manually deepened. This allowed the right end to swing inboard of the protruding, original pocket/pilaster. It may appear that opening was also widened, perhaps for installer satisfaction vs true function. The installation video shows this was not necessary.
The left pocket on the other previously installed truss appears to have also been deepended (but not widened).
Q: Why didn't they lift the truss, with both ends on the other side of the pilasters? Then the far chord could have been slid horizontally into the open end of the pocket, vs outside of it.
Q: By outer edge, does this imply line contact across the width of the cord, at the outer edge of the bearing surface? Hard to control where line contact will occur, based on camber at cord end and plane of the bearing support.
Kevin Kelleher, P.E. (retired)
Internal Mechanical Eng'g Consultant
DuPont ESD Specialists
On the pilaster opposite of which has been posted a few times,you can see what appears to be two studs protruding up through the bearing plate, as well as what appears to only be 3" to 4" of bearing area as they sat it down. Without any room to get in from behind, I cannot see how any bolted attachment of the truss to pads could exist. If it was a welded joint, would Aprox 4" of weld per side actually suffice here? Could 2 studs holding the plate down be enough to prevent tilting?
The whole arrangement to me looks like once they got it in there, they assumed it would stay once they blocked it back in.
The whole arrangement to me looks like once they got it in there, they assumed it would stay once they blocked it back in.
- Status
Similar threads
