August 10, 2007

Foreshadowing Of Disaster In Cleveland?

Yesterday, I wrote about the NTSB's new interest in the gusset plates that held the St. Anthony Bridge's girders together as a possible cause of its collapse. The Cleveland Plain Dealer reports today on a near-collapse of an Ohio bridge that started with construction work and ended with failed gusset plates eleven years ago (h/t: CQ commenter Mike):

Two failed bridges. Two scarily similar scenarios.

Last week, the Interstate 35W span over the Mississippi River in Minneapolis collapsed under the weight of rush-hour traffic and construction crews. Federal investigators now wonder whether the design of steel plates joining beams is to blame.

Eleven years earlier, the eastbound I-90 bridge over the Grand River in Lake County failed. The reason: the same steel plates, called gussets. They had corroded, then buckled after crews blasted them during painting preparations. ...

The spans are Warren truss bridges, made of diagonal compression members joined by gussets. Both bridges are nonredundant, meaning that if one part fractures, the whole structure can fall down. At the time of the failures, both bridges had work crews and equipment weighing them down.

Was this a missed warning sign? It's hard to know for sure. The NTSB apparently didn't think so at the time, and they had some good reason for that conclusion. The Ohio bridge incident differed in some particulars. For one, it wasn't rated for the kind of heavy construction equipment that the crew had parked on it during the work. Secondly, they used steel shot to blast away corrosion from the superstructure during the work, which seriously degraded the integrity of the gusset plates. Nothing like that had been done to the St. Anthony Bridge before its collapse.

Still, the similarities seem rather compelling. Once the Ohio bridge had one gusset plate fail, a number of them bowed outward. No one seems to know why the entire bridge didn't fail, but it did sink three inches and had to be closed for almost six months. It calls into question whether the NTSB should have reconsidered the Warren truss design and warned states of potential issues with the thickness of gusset plates on similar bridges, such as the St. Anthony Bridge in Minneapolis. Ohio focused more of their inspection on these plates after the 1996 incident, but the word did not appear to get to other states.

There are no national specifications for gusset plates, either. Designing support structures is part science and part art, as designers have to assume many variables for use, traffic weight, age, and corrosive elements in the environment. However, the St. Anthony Bridge had a uniquely long span over the river with no center supports. As the Plain Dealer notes, there are no charts for correct specifications on gusset plates.

In other news, divers retrieved three bodies from the river yesterday, including a mother and he baby and a man who died trying to save others:

Peter Hausmann, a father of four from Rosemount, survived the collapse and escaped from his van into the murky, turbulent waters, according to a source involved with the investigation. In the resulting chaos, he apparently swam toward victims in another vehicle in an attempt to render assistance, the source said. ...

Hausmann spent about three years doing missionary work in Kenya and maintained ties to Africa, working on AIDS projects and building a church. If he was trying to rescue someone, it would be typical of Hausmann's selflessness, his friends and co-workers said. "Pete is the type of guy who would do anything to help someone," said Jeff Olejnik, Hausmann's boss at Assurity River Group in St. Paul.

Another friend echoed that sentiment. "That would be Pete," said Gerry Fisher, a friend and former co-worker of Hausmann. "If there was a last act of Pete on this Earth, that certainly would be consistent [with who he was]."

The other two victims identified were Somali immigrant Sadiya Sahal and her 22-month old daughter Hana. The Star-Tribune has a heartbreaking photo of Sadiya and Hana, and published a portion of a letter Sadiya wrote about how excited she was to come to America for a better life. Both stories put the tragedy in stark personal terms, and reminds us how many people will need healing and support from this community.

UPDATE: Numerous commenters and e-mailers point out the obvious, which is that the salt water mentioned in the article would have come from road salt during the winter, which is also true here in Minneapolis. I took out that "difference" from the paragraph above. Also, the other gusset plates bowed outward, not "blowed".


TrackBack URL for this entry:

Comments (14)

Posted by Mahon | August 10, 2007 8:48 AM

Salt water? Must have flowed up Niagara Falls.

Posted by Mike | August 10, 2007 9:01 AM

> The Ohio bridge had to deal with exposure to salt water, for instance, while the Mississippi is obviously a fresh-water river.

The bridge (about 30 miles east of Cleveland) is over the Grand River, which flows into Lake Erie. Here is a satellite picture:,+lake+county,+oh&ie=UTF8&ll=41.735626,-81.10414&spn=0.003166,0.00677&t=k&z=18&om=1

These are also fresh-water bodies. The 'salt water ' would ahve been runoff from road salt, which I expect would also be in use in Minneapolis.

Posted by John Cunningham | August 10, 2007 9:08 AM

Good reporting, Cap'n, but as the previous guys noted, no salt from the Lake. However, Ohio road crews do use a stupendous amount of salt during the winter. When I moved back to Ohio from Alaska, I was struck at the salt crusts which accumulated on my truck when there was even a dusting of snow. And bridges get salted more heavily than roads.

Posted by docjim505 | August 10, 2007 9:16 AM

Was it determined that the Ohio bridge had trouble due to a lack of taxes?

Posted by RBMN | August 10, 2007 9:20 AM

Re: docjim505 at August 10, 2007 9:16 AM

No, I think it was determined to be caused by a shortage of public transit.

Posted by pk | August 10, 2007 9:46 AM

if you look at a truss bridge you will see that all of the members form triangles.

a triangle is the only geometric form that will not distort without one of the legs bending (compression) or breaking (tension). this allows tremendous strength in a relatively light weight design.

it seems to me as though the railroads used a number of warren trusses but are replacing or have replaced them.

perhaps a smarter railroad/bridge man than i would like to comment.


Posted by IB Bill | August 10, 2007 12:14 PM

Equipment failure follows a bathtub-shaped curve. That is, things fail at the beginning of their use and at the end of their expected life in far greater percentages than during normal use. Maintenance tends to start that bathtub-shaped curve over again.

Based on this, I'd look at two things first: The role of that construction crew and connecting devices used to connect the trusses themselves. I suspect you'll find the problem in one of those two places, or both.

BTW, good post, Captain Ed.

Posted by Conrad | August 10, 2007 12:48 PM

Federal investigators now wonder wheather the design of the steel plates joining beams is to blame?

I would not be surprized if the Fereral investigators conclude after the investigation that the collapse was the result of several deficiencies caused by age, wear, overload, corrosion, lack of maintenance, etc.

What bothers me about this is the aparant apathy involved with the people we trust to manage the public assets.

How can we blame a certain part that lasted for 40 years under increasing load conditions as a design flaw ?

Posted by pk | August 10, 2007 1:50 PM

there is this thing called "fretting corrosion". that is when a close fitting metal part is forced into another metal part there is a tiny bit of corrosion formed. when this moves away because of rain, snowmelt, vibration (stand on a modern bridge and you will see what i mean) it exposes a new surface to corrode. as time marches on this has a tendency in extreme cases to "rot away" considerable material. in this world when material goes away it makes the part weaker, especially the fasteners.

the bolts (rivets), commonly referred to as fasteners, holding gusset plates are particularly suceptable to this.

rivets are the strength of the native metal that they are made of (you cannot have heat treated hot rivets because the riveting process would anneal the rivets at installation). pop rivets (cold riveting ) was not in industry in sizes big enough for bridge work at the time that this bridge was erected. they are now. however it takes king kong and six apprentices to get the gun up there.

terrifically strong bolts were available then but they were expennnsive.

keep in mind that there will be more than a hundred thousand bolts in these structures and they will will be big ones (1&1/2" or larger).

rivets and bolts allow the gussets and members the chance to shift and accomodate considerably before they are up against a hard place.

welding makes for a rigggggid structure that has very little accomodation.


Posted by burt | August 10, 2007 5:33 PM

One didn't need to have the Cleveland bridge to recognize the obvious that salting the road would cause corrosion. Another cause of corrosion is the bird dung found on this bridge. I understand the gusset plates were a dissimilar metal to the rest of the structural steel used in the bridge. This would mean the material of the fasteners holding the structural steel parts together were dissimilar to the material of at least one of the parts. These places where dissimilar metals touch are a primary locus of corrosion and since they are hidden cannot be visually inspected.

Civil engineers are required to design according to codes which frequently are reduced to charts. Competent engineers either know how to supplement the charts when necessary or they find someone to help them. Lack of charts may be an adequate excuse in a court of law it is hardly an ethical excuse for a dangerous incompetent design.

Posted by Wise Ol Bird | August 10, 2007 8:47 PM

Just curious.
Anyone here ever design a bridge?
Inspect a bridge?
Blow a bridge up?

Posted by pk | August 11, 2007 1:20 PM

Even small bridges are huge projects.

one guy works on foundations.

another guy works on structure.

in this case another guy would work on deck treatment.

they sign the printzzzz.

the print package for the proposal will probably be about an inch and a half thick 30"x42".

a checker will check the prints and initial off.

a boss will examine the work and approve by initialing.

someone else (the owner of the firm?) will sign off on the whole project.

one of this bunch (probably the owner) is liscened by the state and is called a professional engineer. he stamps the prints as ok.

something turns up wrong a man from the state liscencing entity comes down and tears his liscence to confettii and confiscates his official stamp.

the liscience is worth $$$$$ to engineers and so they are verrrrrry careful not to lose it. the loss of it also means that they are incompetent.

and so no individual is going to claim that he designed a bridge unless it was a very very small one or unless he is a named architek (those guys have towering egos). they have a distinct tendancey to be consensus projects.

my lifes work was in a parrelel universe with many cross overs to bridge work.


Posted by pk | August 11, 2007 1:31 PM

Hey Burt.

i have direct experience with electrochemical corrosion between dissimilar metals but never thought that the various steel alloys would react to each other in that fashion.

do you know or are you speculating.


Posted by Charles D. Quarles | August 13, 2007 8:55 PM

Hi guys.

Let us not forget that iron is, If I remember correctly, the one metal whose oxide does not strongly adhere to the metal surface. This will also apply to most steels (iron alloys).

Post a comment