Miami Pedestrian Bridge, Part II 55

[bridgebuster]

Attached is a research paper about testing of a 160' prestressed roof truss. During the testing it was noted that the first crack was a horizontal crack at the first diagonal. To me, it supports hokie66's theory of a node failure. In looking at Sheet B-17 of the DB proposal, in Detail A, note that the centroids of the roof member (top chord) and two diagonals do not coincide but that may not mean anything.

In the attached paper, the truss that was investigated had HS rods as verticals; they also pointed out the need for confinement reinforcement.


You can take that to the bank.

 

[gte447f]

In the pic (screen shot from NTSB investigation video) posted above by Meerkat007, the PT rod anchor plates from Member #11 are visible. The lower one, which is the one that was being stressed and still has the hydraulic ram attached appears to have moved a good ways toward the upper end of the #11 as the the PT rod erupted out of the top chord with the ram still attached. Ingenuity proposed in part 1 of this thread that the rod had not ruptured, but had erupted out the top after a failure in the dead end anchorage zone in the node at the far (lower) end of #11. That appears to be the case. There is also the spalling off of the entire bottom face of #11. Don't know if that happened prior to, simultaneously to, or after the dead anchorage zone/lower node failure.

 

[TehMightyEngineer]

EPCI: The typical shear design includes plain concrete shear capacity combined with steel shear capacity. Prestressing adds a third element; compression. Compressing the member provides additional shear capacity. You are correct that severe cracking reduces or eliminates the plain concrete shear capacity, for this reason high-seismic concrete design cannot take credit for any plain concrete shear capacity due to the extreme spalling and cracking expected in a high-seismic loading event.

If shear was indeed the failure then a sudden loss of compression in the member and/or any severe cracking would reduce a significant amount of the shear capacity of the member.

Professional Engineer (ME, NH, MA) Structural Engineer (IL)
American Concrete Industries

https://www.facebook.com/AmericanConcrete/

 

[structuralengr89]

Bridgebuster,
the title of that article is "Truss-Girder"...that is not a truss but a beam with holes near the center where vert. shear is low. It is solid webbed as it approaches the reactions where vert. shear capacity is needed...We don't have that with this truss

 

[racookpe1978]

But the diagonal truss members whose ends are visible show a "clean" finish: No rebar coming out of the exposed ends, no cracking, no distortion. Just smooth concrete surfaces as if the ends of the angled trusses were wedged into notches in the upper and lower surfaces at both ends. The PT cables continue through into what is now open air, but no rebar is visible.

That makes little sense.

 

[SwinnyGG]

There is also the spalling off of the entire bottom face of #11. Don't know if that happened prior to, simultaneously to, or after the dead anchorage zone/lower node failure.

Question for the more concrete experienced:

Is it possible that this extreme spalling on the under side of this member is indicative of a massive internal moment generated by near-instantaneous removal of one of the PT tendons and its associated compression load?

In other words:

-#11 starts out with 2 tendons which are apparently symmetrically placed about the beam's neutral axis
-Tendon A fails, leaving Tendon B intact and leaving an eccentric load in the member
-#11 is subjected to a giant internal moment and a bunch of material spalls off of the tension side

It also looks like there's a fair amount of steel inside those members beyond PT elements. Has no one found a reinforcement schedule yet?

 

[dik]

EPCI: Further to TME's comment. Some jurisdictions do not allow the concrete to take any shear if the stress is beyond a certain limit.

Dik

 

[dik]

Just speculating, a couple of scenarios:

The original procedure was that #11 tensioned rod was to be de-tensioned, once in place and was forgotten. A crack developed and reminded the contractor that #11 was to be de-tensioned because it was a major compression member and didn't need the prestress. In process of detentioning, coincidentally or due to a change in the loading regime; the collapse of the walkway was precipitated.

alternatively, if the stressing of the rod in #11 was increased and the rod failed, the sudden release of this energy may have caused vibrations that precipitated the collapse.

I still cannot wrap my ears around the 'fake' stays... If not intended, I would have designed a connection that allowed movement. It doesn't make sense that they had no purpose.

I did renovations to Polo Park shopping centre about 30 years ago... and when we were demolishing one of the small structures that had a huge sign connected to it, I discovered the 'connection' of the sign to the building was simply a pipe sleeved in a larger pipe. I talked to the engineer that did the original building and he explained that it was a city requirement that a sign of that size had to be attached to a building and he knew better than to attach a huge sign to a little building, so, he provided a slip connection that looked like an attachment.

Dik.

 

[dik]

Has no one found a reinforcement schedule yet?

Not yet, need shop drawings to see what should have been placed.

Dik

 

[3DDave]

jgKRI - the massive spalling is from the rod being torn out the side during collapse. The retaining rings of rebar are severed as the rod zippered down the beam. The B-Roll of the NTSB, as mentioned above shows the rod still attached to the lower end. Here's a link to the investigation of that area:

https://youtu.be/aeJKqojmHgY?t=130

I think this means the lower member #11 PT bar did not fail because it would not be sufficiently anchored to act like a pull string and rip out the bottom of the beam.

 

[VolsCE84]

This is my first post, so bear with me. Based off the proposed the transporter was moved and caused #11 to be prestressed. I assume 2 PT bars ( and the pictures show 2) let us assume 280K each. That gives 560k of compression. Now let us set the structure. We all can agree #11 is now a compression member, based on my rough figures it is about 36 dergees and has a length of about 28.5 feet from node to node. Based on 950k of vertical load we now get an additional 1120k of compression loading on #11. Now we have 1680k of unfactored load based on 580k +1120k. In pure compression or around 3.3ksi in the member. That may not sound like much for 8500psi concrete in compression. However, when looking at the numbers even with the high concrete strength, I cannot get the member to work when figuring it as a column. This leads me to think that we may have a KL/r issue. However, the member on the opposite side looks to be longer and had similar loading conditions. Plus is survived impact loading from the crash. To de-tension a bar we must add tension to loosen the nut. It does appear the PT bar broke during what i think was an attempt to de-tension the bars. ( in photos I don't see any anchors for these bars in the bottom. They have been dead end in the bottom as if we did not have enough going on in this area + 2 holes for anchor bolts.) If for some reason during the de-tensioning process the bar broke. Could this be like an impact loading on half of the member? When we look at the NTSB photos we see what looks like pure shear. The aggregates are sheared along a line of between the 2 PT bars. My apologizes for rambling on.

 

[bridgebuster]

@structuralengr89 - what we have at FIU isn't a truss either but a beam with web cutouts.

 

[SocklessJ]

I’m not a bridge guy, but while we’re speculating about the design (which might not have been the problem, we’ll have to see)...

I don’t know how you go about designing those nodes. Strut and Tie? Probably not a lot of textbook examples for STM models resembling this situation. The fact that there are multiple bars in each truss diagonal makes it even more complex, as others have noted earlier.

You’ve got a highly stressed, determinate structure whose local behavior is complex and analyzed using a lower-bound method. If the mild steel around those nodes isn't sized and oriented just right...

I still can’t believe how fast this thing failed, though. From the video of the collapse, it almost looks like the bottom flange fails in tension. Maybe the nodal failure was at the support. Maybe the temporary PT in member 11 was helping confine the joint by adding a vertical clamping force across a horizontal plane just above the longitudinal PT anchorage. Think shear-friction design. The base of Member 11 simply shears across a horizontal plane and punche out the end of the structure. The problem with this theory is that the shear plane I'm thinking of is way too large.

 

[waross]

In answer to dik, No, I'm not structural. I have spent most of my working career in the field, on the ground.
I have seen a lot of stuff go wrong. Today, everything is planned so that nothing breaks.
When I started, a lot of equipment was stressed to the breaking point.
You always asked yourself;
"If something breaks, where will it end up?"
"Don't be there!"
I always had a keen interest in the root causes of any jobsite failure as my future safety may depend on it.
Example, not a failure but an illustration;
We had a lot of anchor bolts to set in concrete for the structures of a substation.
The bolts were inserted into the holding templates and the nuts were spun on by hand to position the bolts vertically.
The templates were positioned on the forms and the concrete was poured.
A worker was tasked with removing the nuts in preparation to erecting the columns.
Why is it taking so long?
It only took half this long to place the nuts and the worker is only half finished.
When the concrete was poured, very small splashes of concrete were deposited on the threads of each bolt.
Now instead of spinning the nuts off by hand, there is just enough drag that each nut has to turned all the way off with a wrench. The worker was working hard but taking off now takes about 4 or 5 times as long as putting on.
What's the point?
The Florida crew has probably been installing the nuts for the most part by hand. They have probably done most of the post tensioning on this project.
It is possible that a spec of concrete or a grain of sand or other material became lodged in the threads and this nut could not be removed by hand.
I have seen too many crews that would keep increasing the tension trying to free the bolt to disregard this possibility.
If the excess tension pulled the anchor on the other end of the PT rod through the concrete, that may explain the failure of the strut and the spalling on the bottom of the strut.
This may not be the reason for the failure but it is too possible to be discounted offhand.

How about the concrete strength?
I have seen too many times concrete that did not meet the design strength. (Too much water makes it easier to work with.)
I have seen too many concrete samples that did not reflect the strength of the installed concrete. (Don't add the water until the inspector leaves with his sample.)
Stuff happens in the field.
Sometimes more stuff happens with design build projects.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[gte447f]

SocklessJ, BAretired has hit on this also, a failure of the tension chord at or near the lower joint of diagonal #11. I think this is very plausible, and I think the direction and manner of the movement of the bottom PT rod in #11 as the collapse happened could support this theory.

 

[LittleInch]

Am I the only one who is concerned about the mismatch of diagonal supports? The members 2,4,6,8 and 10 all point more or less the same angle one way resisted only by nos 11 and 9 pointing the other.

Does this not induce a force which is essentially trying to push the top section to the right (North end) heaping even more misery on the no 11 support? (see Ingenuities diagram on 18 mar 20:55). Also this is where the initial collapse was focused, but may have been due to other issues.

I've read all the posts and watched the videos lots of times and the speed of collapse is just frightening and seems almost impossible to say what actually broke first.

Certainly fiddling about with the PT bars in that overworked no 11 strut was clearly not a good idea, but is this simply a flawed design from the get go?



Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[racookpe1978]

Well, ALL of the angled truss members are "pointed" back vertically to align with the "not needed" (missing) cable stays/pipes going back up to the "not needed" (missing) vertical pier.

 

[Rabbit12]

VolsCE84 I think you're on the right track but just note, some of the prestress force would be lost because of compressive strain in the concrete diagonal.

 

[SocklessJ]

The more I watch the slo-mo video, the less I like my previous theory. I'm gonna go back to thinking that differential bar tensions in member 11 caused an eccentricity that led to local, and then complete compression failure in member 11.

Once the bridge started deflecting, the angle of the compression member decreased , causing the load to further increase and cause collapse. Could be as simple as that.

 

[dik]

waross... was just joking... we structural guys often have the same list you posted...

Dik