The Interstate 35 Bridge
Doug Ross has aerials of the Interstate 35 Mississippi River Bridge before it's collapse to show the layout. Slide posted a series of photos taken from the ground after the collapse. James Lileks, who runs Buzz.mn has extensive coverage. Noah at Blanked Out lives in sight of the bridge and actually saw it collapse. Here's his account: (Hat tip: Captain Ed, who lives in the area and has wall to wall coverage on the collapse.
Description of collapse: Surprisingly quiet, but my entire warehouse apt shook quite a bit. Bridge fell so very slowly - poof. Yelled and yelled but no one answered. Only a handful of ppl in water, all got out ok and then began to help as well. Ran into woman who thought school bus was trapped on east side in fold of highway. Couldn’t confirm it or get to the other bank to help. ... I live the closest to the 35W Bridge than anyone in Minneapolis and I watched it come down from my roof.
I spent 20mins getting people out and off the bridge. I’m being evacuated.
PS This was not a bomb.
posted by wretchardthecat at 8/01/2007 07:01:00 PM
28 Comments:
Interesting visual comparison between older 4 lane bridge and the newer 8 lane overpass. The 8 lane bridge appears to be made out of toothpicks compared to its massive half-capacity cousin.
looking up river
looking down river
Wager this will be one for the textbooks. Perhaps someone relied too much on computer simulation and cost-cutting optimization v. old-fashioned (over) engineering?
The Star Tribune has another series of photos taken after the disaster. Lileks adds that:
Governor holding a press conference right now; says the bridge was inspected last year; not scheduled for overhaul until 2020. No one is ascribing the collapse to the construction, but am I the only person to think that it might be a factor, somehow?
I'm sure the subject of its construction is going to be revited somehow but it may be a while before any kind of definitive engineering post-mortem emerges.
"Wager this will be one for the textbooks. Perhaps someone relied too much on computer simulation and cost-cutting optimization v. old-fashioned (over) engineering?"
Doubtful, the bridge was early 1960s. I doubt that it was done with computers, it was still the slide rule era -- see Apollo XIII.
Info about bridge here.
Bridge opened in 1967.
Seymour Cray (at Control Data in Minneapolis at the time) brought out the CDC 6400 in 1964, and they had a finite element package that they offered as a service on what was the supercomputer of that era. And simpler portions of the design equations had already been coded in FORTRAN and run on IBM machines prior to this.
Hopefully just a coincidence.
I live about 200 yds from the bridge. If you don't have an overlooking view you can't really get to anyplace to see it. Access to neighboring pedestrian bridges (10th Ave Ped Bridge & Stone Arch) are closed off. Still cops everywhere. Everything indicates a phenominal coordination effort between all metro and surrounding agencies. I had a small armada of rescue boats from surrounding communities drive past my office (1 mi away) not 45 min afterwards.
Several friends in and around the scene before and shortly after didn't notice anything giving any warning.
10th Ave Bridge (for cars) is totally closed off for on-site coordination. That's the big cement arched bridge next door. 35W was a single span over the river, versus having a piling in the middle.
I'd driven the bridge during construction. I only saw surface work. Refurb/replace driving surface. Nothing more than an inconvenience to drivers.
...Glad I walk to work...
-Sierra Victor
I live in the Twin Cities, and just graduated from the U of M. Interestingly, the bridge just next to I 35W is on the states registry of historic bridges. It is that old, I believe WPA era.
Something odd about the collapsed bridge is that at the foundations there are these hinge like braces to the foundation blocks. I've driven and biked underneath it many times along the now blocked West River Road, and it always looked flimsy . . . and with those wierd hinges. Almost emphemeral compared to the older 10th St Bridge next to it.
Here is a write up on the bridge and a brief report on the failure today.
Here's an engineer's speculative take on what happened from KTSP and he focuses on the 2006 inspection finding of "out of plane cracking" in the beams and diaphragm connectors under the roadway. See the link for more info.
Prediction: No bureaucrat will lose one dime of their pension or health care benefits over this.
Thanks for the photos, Ari Tai.
Bridges are flexible structures. Sections have hinged supports on one side because they must flex to deal with varying loads and thermal expansion and contraction. The other side isn't a bolted connection but sort of rests on its support.
The boom, boom, boom reported by witnesses is consistent with a progressive failure within a torque box connecting steel beams. The culprit in such cases is sub-standard or missing bolts.
BTW, I listened to the linked report. It speaks of fatigue cracks in the approach beams, not those at the center of the bridge. From the pics, the approaches to the bridge are still intact.
Finally: I can't believe this failure had anything to do with being overly-dependent on finite-element design. Computer-aided engineering was so new in the 1960s that no engineer worth his salt would have relied on it without checking the calculations manually.
solomon2,
You mention the effects of thermal expansion.
With the temperature extremes of Minneapolis (+90s to well sub-zero?), could the wide range of temperatures experienced by the bridge over the years have played a role?
The abutment on the right bank is leaning into the water. Looking at the debris - looks like the whole system slid to the left. The failure could have been on the bank itself.
---------------
Interstate bridge collapses occur every few years.
http://archives.cnn.com/2002/US/05/26/barge.bridge/index.html
I can recall the bridge on the Red River going down in the 70s.
Finite element analysis in the early 1960's was in its infancy. In that period, when this bridge was being designed, the consumers of FEA were largely aerospace and nuclear - nobody else had the money, the computer/software resources or the time.
Both the bridges my business associate worked on in the late 60s have now been replaced. He used hand methods and early electronic calculators, maybe the odd comptometer. I wouldn't be surprised if MNDoT was in the same boat at the time.
For me, what makes this bridge both elegant looking and scary is the lack of redundancy. One chord buckles or has a tension crack run uncontrolled and its down. Built like an airframe, maintained like a bridge. Modern practice puts redundancy back in and details for fatigue/fracture. Contemporary practice is based on experience and experience is based on the lessons of failure and failure avoidance.
The use (if used) of salt may be a factor. I live in the eastern shore of MD; not too much snow, so not too much salt use. But my brakes failed, the front and rear break lines just beyond the master brake cylinder, a couple of months ago on my 1998 truck due to salt according to my mechanic. Nowadays the highway departments supposedly mix silicon with the salt to make it stick better/longer to the road surface. Therefore, when splashed up inside of your vehicle it will not wash off easily through normal use. Supposedly relatively new cars are experiencing this problem. My brake lines are carbon steel and the bridge bolts/structure is also carbon steel, I would imagine. I also wonder if the original paint, lead based most likely, had been removed. The new paints have not inhibited corrosion as well as did the lead based ones.
I notified the US DOT and truck manufacturer of my problem for what it was worth.
"Contemporary practice is based on experience and experience is based on the lessons of failure and failure avoidance."
Oops, contemporary really should be contemporaneous. No slam intended on MNDoT - I'm sure they used the best methodology and tools of the time.
For those interested, a good non-numbers intro to engineering is J. E. Gordon's Structures: Or Why Things Don't Fall Down. Another is Henry Petroski's Success through Failure: The Paradox of Design
It is most unlikely thermal expansion and contraction had anything to do with this bridge failure.
Road salt - here, the use of the potassium acetate deicing system installed in 2001 - probably had no effect, either.
I followed the link of one story to an analysis from several years back of the bridge that concluded early replacement wasn't necessary; however, the report did say in Summary item 7 that computer simulation didn't match measurements unless some "composite" action was taken into account. The composition and boundary conditions they had to stick in to this 2D model may give a very rough idea (qualitative, not quantitative) of how fatigued the bridge was.
Fatigue only happens when you have load reversal. I think that in this sort of bridge that rarely happens. However, because of the night-time road work, the span of the bridge devoid of vehicles may have experienced it via vibration through the rest of the bridge that was still in operation. Pretty far-out, I know, but I imagine it could joggle a bolt loose.
Let me make it clear that all my comments about bolt failure are a speculation. It's waaayyy too early to tell. One newspaper story quoted a motorist who drove under the bridge. He reported hearing weird creaking noises and then the bridge fell behind him. That sort of thing is more indicative of stress (not necessarily fatigue) failure; popping bolts would then follow.
The NTSB will collect what they can and key items will go under a microscope to see what happened. I'd wager that they'll find every piece of steel of import, no matter how small. The cause will eventually be found.
LGF has a link to a video of the collapse.
Looks like the bridge shifts a bit horizontally then the whole central span falls at once.
You can have fatigue without loads being completely removed. The dead load of the bridge and deck is permanent and substantial. A bridge life span is projected based in part on number and weight of vehicles crossing. With the intermittent stress of each crossing, a little more damage is accumulated. Many bridges and roadways are loaded beyond design projections from the get go. The engineers' conundrum is that if they overdesign relative the spec, they can be accused of goldplating. Everybody likes safety, but who likes the bill?
I guess it is because I was in DC during the 1992 presidential campaign, but this brings to mind the “Infrastructure!” war cry used by the Democrats back then. I even recall a banker’s wife in Richmond, VA, being interviewed on NPR and saying she was voting for Bill Clinton because she was terrified that a bridge might collapse under her some day.
The question is: What was really done? Did the Democrats wave the “Infrastructure!” warning around and then forget about actually doing anything about it, as was the case in so many areas?
I do recall thinking ‘They have to be kidding!” when the Infrastructure Scare was being run up the flagpole. The amount of highway construction at the time seemed to be absolutely incredible to me.
In any case, you can bet they will trot it out again.
Former NTSB guy was interviewed on TV and said that the bridge was not designed in a fail safe manner. In other words, failure of one part could lead to failure of other parts. I suspect that this was the deficiency mnetioned in the inspection.
By the way, On Fox News they had a surveillance camera view of the main span falling. Was time lapsed but still seemed to be very fast.
In any case, investigating a bridge failure is not like an aircraft or rocket mishap investigation, which I have done - you always have all the parts with a bridge. They will find out what happened pretty fast.
I did structural engineering in the eighties, specifying bolts and welds for massive steel columns and beams in power plants. I always worried that I may have made a mistake, somewhere, at some time. This bridge collapse is horrifying. One has to feel bad for the civil engineers and contractors who designed, built and maintained, and for those users who trusted.
The trusting and failure are so difficult to accept.
The dead load of the bridge and deck is permanent and substantial.
That's the trick here. Although the current road work wasn't structural, they did remove entire sections of pavement - creating holes in the bridge open to the sky - which would have reduced the dead load of that area substantially.
Look at the video.
There were not a lot of cars on the bridge when it fell.
If they did remove whole sections of the bridge deck over the past few months, (per Soloman2) that radically changed the vibration characteristics of the bridge. That made it capable of vibrating at a much different – higher – frequency than it would when it was complete, with the deck intact.
Vibration produces fatigue by subjecting structural members and their joints to loads that are not normally seen. And this bridge was steel; steel vibrates much more readily than concrete, wood, or aluminum. Force = Mass times Acceleration. More movement equals more force. High cycle fatigue comes from the vibration repeatedly bending an area when the structure resonates at its natural frequency. I am a mechanical engineer rather than a civil one but I doubt that fixed, cantilever, steel bridges take such vibration into account; the vibration normally would not be there with the bridge deck intact.
I recall having to sit on the old Woodrow Wilson Bridge over the Potomac one day while they raised the drawbridge to let a ship go by. I was astonished when the bridge suddenly started shaking. They had put the drawbridge back down and the rhythm of the oncoming traffic on the other side was causing the bridge to shake. That bridge was designed for that kind of vibration in mind; I doubt that the one in Minneapolis was.
The answers to the questions are currently under water and tons of concrete. I'd guess the two arch sections of the bridge nearest the concrete supports on each end will probably tell the story. The investigators may look at general corrosion, undetected fatigue and/or stress corrosion cracks, fasteners and joints, welds, asymmetric loads during repairs and a whole laundry list of other factors. Likely they'll use materials testing, NDE crack testing, FEA simulation, and integrate/check with witness accounts and photo/video plus reviews of the original plans, upgrades/repairs, inspections, etc. The WTC investigation documentation ran to 7000+ pages. I hope they can pinpoint the probable cause more easily than that and forestall similar disasters elsewhere.
"If they did remove whole sections of the bridge deck over the past few months, (per Soloman2) that radically changed the vibration characteristics of the bridge. That made it capable of vibrating at a much different – higher – frequency than it would when it was complete, with the deck intact."
HUH?
Its the low frequency oscillations that kill you because they can dump energy into higher frequencies and into other modes.
The danger in altering the bridge is that it changes its harmonic frequencies, dampening, and harmonic coupling in unknown ways.
Based upon Sisyphus' observations of the East end rolling up before the bridge fell it appears the anchor on the East bank failed.
Post a Comment
<< Home