Skip to comments.Minneapolis I-35W Bridge Collapse Failure Analysis
Posted on 08/21/2007 3:20:52 PM PDT by jeffers
click here to read article
Your piers are labelled correctly, but not your connection points.
The mainspan (span 7) center vertical strut is member U1-L1. The next strut south of there is U2-L2. The next strut north of the center strut is U2’-L2’.
Your cantilevers and trusses are generally correct, but the two center cantilevers are drawn a little generous. In the proper labelling scheme, the pier 6 cantilever only extends out (towards midspan 7) to U6-L5.
I’ve looked very hard at a triggering failure in the southern approach. Over and over again, I have to rule it out, at least as the most likely sequence.
Primary reason, as you note, the pier 6 superstructure would have rotated, top towards the south, if span 6 sagged midspan. That would have lifted the south end of span 7 well up above the horizontal plane of the deck, and we simply do not see this in the video.
If it happened that way, the top chord fractured north of pier 6 almost immediately, before any (assumed) visible deflection took place in span 6, and before said deflection could induce (actual) visible elevation of the span 7 south end.
I won’t rule out a triggering failure at the pier 5 crossbeam/endbeam rocker assembly, given the history there.
But the video does not visibly support such a sequence, and therefore, it’s not the first sequence in my assessment of the likeliest possibilities.
It’s not the best image I’ve seen but you’re looking at the south endbeam in this photo.
The junction between the south end of the main trusses and normal beam and post approach spans did not fall directly over pier 5. It was located south of pier 5, and was designed as a somewhat complex cantilever assembly.
The far end of the beams you see (resting on the ground) were between pier 4 (seen in the image) and pier 5, (not seen).
For illustration as to why the visible beams buckled to the near side of the visible pier 4, imagine pier 3, pier 4, no main trusses, and no pier 5.
In such a scenario, the visible beams sit with one end on pier 3, and are cantilevered out into free space past pier 4. Load the free end of the cantilever in excess of design limits, and yes, the beams will buckle just to the pier 3 side of pier 4, as you see in the image.
That is where the bending stresses reach maxima. The reason for this is due to the lifting reaction placed on the pier 3 end of the beams. If the only stress in play was the gravity load on the free end of the cantilever, then the beams would buckle over pier 4. But the dead load and connection assemblies at pier 3 resist the gravity load imposed on the free cantilever, and in effect, bend the beams over pier 4 like a fulcrum.
There’s more to this, involving the nature of the pier 4 beam connections and how they are permitted to move and restricted in movement, but this should outline the basic idea.
Would telling you that I saw the equipment sitting on the bridge a couple of times a day when I crossed it during June and July substantiate the claims?
Can you remember what you saw, and where it was?
Especially interested in heavy loads between mid span 7 and just south of pier 6.
Removal could have had small cushman type debris removal staged to another point on the bridge and then hi-loadered into dump trucks. Cushman’s wouldn't’t be a big load nor would small piles of debris staged for reloading. Only one truck would have been filled at a time and then it would depart. Others waiting would have been empty, not full as some have speculated in discussing “trucks full of gravel”
In the repaving operation, if concrete, we could have seen ready-mix trucks lined up too far out on the structure and a paving screed. I haven’t heard there were ready-mix trucks in the collapse debris vehicles.
Conversely, if asphaltic paving or wear course was underway, a “paving train” would have consisted of a crawler driven small paving machine and trucks loaded with heavy asphalt lined up to dump in it. These heavy trucks could have been staged off the structure or they could have lined up on the closed lanes.
It appeared that small sections were being done in stages so I don’t really think that any of the foregoing were the case, but there has been so much talk of construction material loads that at least we should discuss what we might be looking for in the debris or observations.
I can imagine that MDOT and the contractor aren't talking about specifics at all until the forensics and formal investigations are done.
I can imagine that the contractor and his liability carrier as sweating blood.
Something I noticed studying the images, was that the roughened deck surface visible after the collapse appeared to be confined to the SW corner of the bridge, maybe from pier 5 and south, in the southbound lanes.
I imagine the contractor had an overall schedule, and other parts of the contract work could have been underway elsewhere, but that corner seemed top be the primary focus of work in progress from what the post collapse imagery showed.
With that type of repair it can have a new surface in either concrete topping or asphaltic topping depending upon the substrate of the paving. Both operations can either be done in small segments or in large scale paving operations as previously described.
In either topping we are not looking for "trucks loaded with gravel" as some news accounts describe, but instead, we are looking for trucks loaded with hot asphalt or ready-mix trucks. The former would have been servicing a paving machine:
The machine and the trucks servicing it should have been smaller that what is shown above but this shows the operation and machinery type.
A concrete paving operation could have been a small vibrating screed with almost no weight or something of this type
Again, this is a larger unit that what would be used for bridge topping but it shows the sort of machine utilized and its appearance for those not familiar with this process.
From what I remember is they had what I think they were generators built on trailers, trucks that came and went and road cutting saws. There were several of each of these distributed the full length of the bridge on both sides. They took up a lot of the 2 outside lanes in both direction, where they were working.
This is kind of vague, but hopefully it is helpful.
I would think that the truck mounted items you saw were truck mounted air-compressors as opposed to generators.
Compressed air generators and the trucks are not heavy virbration or dead load.
Compressed air is used to blow off the light debris and dust after the larger chips are removed behind any scafication operation. That gives a clean surface that the topping / resurfacing can bond to as it is placed.
Regarding the official bridge labelling scheme, you are correct. The drawing I based my assumptions on is a poor one but the text, on deeper investigation, does indicate the correct labelling scheme, and I apologize for any confusion in this area.
I also note that an earlier poster’s scheme actually is correct.
I pity the efforts you had to go through to make sense of my analysis in this respect. It’s hard enough to understand when the numbers line up.
As you note, a different scheme doesn’t change the basic conclusions, though I would clearly state, it moves the suspected problem area to the U10 connection point of the east truss.
This opens the door for some new data, and I am still working on other areas of analysis, including construction traffic and further identification of structural members in post collapse imagery.
I doubt these avenues of analysis will lead to concrete conclusions, and have to decide whether to publish sheer speculation or let important, but unresolved, issues lie unmentioned.
I note with interest your assessment that other scenarios are possible. I have looked at some of these, and for the most part, found one or more conflicting datapoints that appear to rule them out, though U have not and will not do so categorically.
I’d be interested in your read of other failure mechanisms that remain possible, given the available pool of empirical evidence.
I believe that posting them, even as speculation, yields more benefit then harm, and opens the door for further discussion.
Thanks for responding. I was afraid everyone had finally left this discussion! I think it’s important to realize there is a line between what we know or can logically infer, and rank speculation. As long as we maintain an awareness of that distinction I think it’s reasonable to postulate and conjecture.
Your fine detective work on identifying the steel members of the structure post-collapse is 100% valid IMO. It may be too much to ask that we’ll see hi-res pictures of the scene as the rubble is removed. I would think they would remove the concrete and rebar first, exposing the steel underneath. There’s a lot under there we’ll never see.
Regarding alternative collapse sequences, my thoughts on that come from reading the URS bridge analysis report. They identified 52 critical beams that, if one of them failed, would cause others to fail like dominoes. They listed them in their report but I need to go back and read it again. IIRC they identified secondary failures in some members that were not in close proximity to the initial failed member. That stood out when I first read it and I must read it again. Their study, of course, is predicated on certain assumptions that may turn out to be invalid. Garbage in, garbage out. If there initial data was bad then the results are suspect.
My model is not a finite-element model. It’s visual only so right now I don’t envision plugging it into any stress analysis software to analyze. Speaking of the picture I posted above, it occurs to me that perhaps the deck had already begun to fall in the first frame of that video. That could potentially explain the “curvature” of the roadway that I cannot replicate with the intact model. And speaking of video, I read yesterday that someone thought the camera wasn’t a video camera at all, but rather one that snaps still pictures in sequence, either at timed intervals or triggered by motion detectors. I don’t want to speculate further on that without knowing more about the actual camera system they use.
Glad you’re still here.
There are a total of 26 frames of video, and in the original (posted) sequence it lasts a total of about 12.4 seconds, meaning that the time between frames is about 0.477 seconds.
It is interesting to note that there is no bridge movement between the first and second frames, meaning that the video has captured the entire collapse from the start (which begins between frames 2 and 3, and is already underway at frame 3. It is also interesting to note that there is a very slight upward shift of the camera between frames 2 and 3 (coincident with the beginning of the collapse).
It is also noteworthy that the bridge segments closest to the camera (at least from initial observation) appear to be falling straight down. It is difficult from this perspective to tell if the east (far) side of the bridge was falling ahead of the west (near) side.
Even though the bridge appears to have broken somewhere on the south end, out of camera view, these pictures can perhaps still provide some additional information and insight into the collapse.
I downloaded those as soon as I got your e-mail.
Those and the other images you linked me to are very good ones, and have already allowed me to further the investigation.
Some things are becoming very clear, expect some minor changes in my opinion of where the earliest known failure occurred, but nothing major.
I’m going to take a few of those video images and manually extend the superstructure skeleton to the right to include pier 6 and you’ll see what I mean.
The focus is narrowing and the possibilities diminishing, and there are critical superstructure elements visible in both the older images, and to a better degree in the newer ones.
It will take time to put it all together, and I am obligated over the weekend elsewhere. Maybe mid next week?
In any event, thanks for your help with those images, your contribution to our informal analysis cannot be overstated.
I’ve been away for a couple weeks, are there any current discussions on the collapse? How about current images from the debris cleanup?
Interstate 35W Bridge Photos
Thank you for your info above.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.