The investigation into the collapse of the St. Anthony Bridge has taken an intriguing and somewhat unexpected turn. The NTSB has issued an alert based on a potential design flaw that could have caused the catastrophic collapse — and that no inspection would likely have caught:
Federal officials investigating the Interstate 35W bridge disaster said Wednesday that they are looking at a possible design flaw in some of the steel plates under the bridge and issued an alert that added weight from construction work may have been a factor in its collapse.
Opening a new window into last week’s fatal bridge collapse, the National Transportation Safety Board (NTSB) said that one of its areas of inquiry involves the design of steel connecting plates known as gusset plates; the material makeup of those plates; and the loads and stresses they bore.
Hours later, Secretary of Transportation Mary Peters said the NTSB indicated that the stress on the bridge’s gusset plates may have been a factor in the bridge collapse and that one possible stress may have been the weight of construction equipment and materials on the bridge.
Peters issued the first national alert to stem from the disaster, telling bridge engineers nationwide to “carefully consider the additional weight placed on bridges during construction or repair projects.” An NTSB official stressed again that the probe is in its early stages and that the design of the gusset plates is just one of many areas of inquiry.
From the beginning, people wondered whether the construction activity on the bridge contributed to its failure. The work had nothing to do with the underlying structure of the bridge, however, and that appeared to be merely a coincidence. The NTSB apparently doesn’t believe in coincidences.
So far, though, the construction team doesn’t appear to have done anything unusual. In resurfacing projects, it’s not unusual to have the raw materials dumped nearby for easier access by the crews. As the Star Tribune notes later in the story, that same material would have been mixed as concrete later and applied to 1/16th of the bridge surface, which would have only distributed the weight somewhat more broadly. Even if the material weighed 100 tons, it would not have been as heavy as three fully-loaded semis, which travel over that bridge constantly.
The gusset plates also came under suspicion yesterday. The NTSB observed a “design issue” on one in the wreckage, but refused to specify the issue or location. The paper reports that investigators believe the plates were too thin for the loads they had to handle. The bridge managed to stay up for 40 years, but increasing traffic and heavier loads may have caused one to fail — and since the bridge had no redundant support systems, one failure in the right place could have caused the catastrophic collapse.
Would inspections have caught this problem? If it was a design flaw, one would suspect that all of the gusset plates would have had the same problem and may not have shown signs of failure until one actually broke. The New York Times reported that a consultant hired by MnDOT after the collapse identified the problem, but MnDOT says that they have no such report. Earlier reports and consultant recommendations did not focus as much on the gusset plates as they did on other parts of the metal, especially the girders themselves.
If the collapse came from a gusset plate failure, especially one where no inspection revealed any problem, then increased inspections probably would not have helped. It seems more likely that the design of this bridge had more to do with its early failure, and that other bridges of the same or similar design should have immediate attention.
The gusset plate piece doesn’t fit earlier news on 10 years of inspection reports and warnings. Also very few big structures fails “suddenly”, there are usually some warning signs of the coming failure. The problem is to distinguish between normal wear and serious structural problem.
Similar thing happen in Montreal a year ago, when pieces of concrete started to fall off the overpass hours before all overpass collapsed.
I’d be interested to see commenters with engineering background expand on this.
I presume there was some requirement that the plate be able to carry a specific quantity of stress?
Does it sound like the quantity was mis-specified, and that the plates were fit for that specification but not fit for actual bridge use? Or was the requirement correct, but the wrong plate used, one that was not suitable to the requirement?
I am not sure where the gusset plates are exactly but I did wonder if tearing out the old cement might have affected some of the containment of the remaining structure. If already weakened, this might have, indeed been what caused the last straw.
UPDATE:a potential design flaw that could have caused the catastrophic collapse I-35W Didn’t Collapse Because Of Lack Of Funds….
UPDATE: The Captain has 2 more very important pices for you to read. The first being Gusset Plates And Heavy Weights, which discusses the fact that the NTSB has issued an alert based on a potential design flaw:
Federal officials investigating the Int…
gusset plates are flat pieces of plate that tie the various beams, t sections, z sections and what not together.
they are odd looking plates with lots of holes in them.
their strength is great in two dimensions, if they deflect they become three dimensional objects and loose most of their strength.
if one or more gets twisted (like from a pier settling or shifting) its all over.
C
and furthermore:
i have seen gussets twist in an earthquake. what 30 seconds of shaking?
i have also seen gussets straight and flat one day and twisted the next.
steel has this thing about yield points. a piece of steel will have all of the characteristics of a rubber band until the force on it exceeds the first yield point. in other words if the force is below the yield point it will return to its original shape to the thousandths of an inch when the stress is released.
if the stress exceeds the first yield then the part is permanently deformed. and it will never be as strong as in the original condition. if the stress exceeds the ultimate yield point than the piece breaks in two.
the high strength lo alloy steels (read structural stuff) are made with high ductility ratings so that if they are over stressed they will bend and twist and not shatter.
if the problem is twisted gussets than someone who knows about this should have seen the problem and fixed it immediately or closed the bridge.
C
As people who use oversize wrenches find out, some metal parts (like bolts) can seem real strong, until they break in half. If the gusset material was just plain inferior, in some way, there may not have been clear warning signs to find with a simple inspection.
In relation to the loading on the bridge, I’ve wondered if the particular distribution caused by the construction caused some unusual loading that the structure (perhaps already in a weakened condition) couldn’t deal with. Having several lanes closed to traffic, I’m sure the bridge was nowhere near its rated dead weight. But, for instance, traffic confined to an outside lane, and heavy construction vehicles and materials nearby, could have edge-loaded it enough to produce an unusual side load or torque.
While we’re just spitballing it here, one other thing that came out of the past inspection reports was the recommendation to re-deck the bridge if the bridge itself wasn’t scheduled for early replacement. Watching the security camera footage and the way the bridge fell straight down all at once still suggests to me a delamination of the deck from the steel beneath. If a gusset plate failed it seems to me there would be more of a progressive failure from that point out.
Regarding the chorus asking for higher gas taxes and castigating us for “neglecting infrastrucure”, some higher gas tax for better roads is a tradeoff I would take. But for blaming the bridge failure on neglect.
I am thinking this bridge will come out as one of those “unknown, unknowns” (some unforseen design problem probably in some weird combination with a bunch of other things) rather than “known, unknowns” (inadequate bridge inspection or giving a passing grade to a failing bridge).
I am not a civil engineer, but the civil engineeers I know are smart people, and you figure that at least all of the obvious things are covered in terms of bridge design, redundant structure, inspection procedures. There are all of those weird confluences of things that you hadn’t thought about that get you (I do in fact develop computer software, and sofware is famous for ‘bugs’ that are often in this category).
Drudge on Sunday night said the firm repairing the roadway, “dumped” tons, about 10-pounds, of loose gravel. So they’d have it “when needed.” And, he said that ends the guesswork. The bridge came down because the people fixing the lane beds HAD NO IDEA ABOUT ANY OF THE ENGINEERING.
I think that’s about right.
Add to this, that it’s nearly impossible to get engineers to “sign off” on “clean bills of health.”
They’re like doctors. Who know you can drop dead for a catastrophic event, soon after you get your “annual.”
One thing for sure, this catastrophic event will make companies that bid on “jobs” a bit more careful. Since they aren’t in business to be too concerned with stuff beyond what they know how to do.
By the time engineering finishes sifting through all the debris, I think you’ll see mock up models that describe the bridge’s first rumbles. To its smashing down.
there is one thing.
locomotives don’t jump ditches worth a damn.
falling bridges don’t make very good gliders.
both fall straight down when they go.
C
An interesting article in today’s Cleveland Plain Dealer, describing a non-catastrophic failure of gusset plates in a similar Ohio highway bridge in 1996:
http://www.cleveland.com/news/plaindealer/index.ssf?/base/news/1186736754295500.xml&coll=2
A section of the bridge suddenly dropped about 3 inches when some of the gusset plates buckled.
First, at this point in time, any opinions on the reasons for the collapse are pure conjecture. With this in mind, there are a few things that I would like to point out.
If the contractor was storing construction materials on the bridge this could have been a catalyst in initiating the collapse. 100 tons of material spread equally over the entire bridge surface is a distributed load, and a load that the bridge potentially could have handled. 100 tons of material in one location on the bridge is a point load. This point load would act as an eccentric load affecting the bridge structure in a way that was not anticipated by the bridge designers.
The gusset plates and rivets seem like a good place to start in looking for the source of the collapse. Structural failures usually happen at the weakest link in the chain and usually the weakest points in a structure are in the connections (bolts, welds, clips, rivets, gusset plates…).
Captain Ed’s conclusion that similar bridges should be inspected / studied is spot on. There are probably clues to the Minneapolis collapse in other bridges.
I would be surprised if the investigating engineers don’t come up with a reason for the collapse. They will figure it out.