Noah Kunin, eyewitness to the collapse
It will take time for officials to untangle the cause of the I–35 collapse.
A surveillance video, failure analysis software program and construction and inspection records could save months off the National Transportation Safety Board investigation of the fatal Interstate 35W bridge collapse in Minnesota. However, Mark Rosenker, NTSB board chairman, added it will take a year to analyze the Aug. 1 collapse that killed at least four people. "We will not rule out anything," he said Thursday evening. "What we begin to do is rule things in."
The Minnesota police department gave NTSB a surveillance video of the steel truss bridge collapse. A Federal Highways Administration official gave NTSB a copy of the failure analysis software he used for his dissertation at the University of Minnesota that resulted in a 2001 report analyzing fatigue cracking on the bridge. The report concluded that "fatigue cracking of the deck truss is not likely, which means hat the bridge should not have any problems with fatigue cracking in the foreseeable future."
"We will go line by line over the report," said Rosenker. "We will look at all the bridge inspection program provisions and see if they are robust enough."
The results could have implications for the current federal bridge inspection program, which includes guidelines for states with inventories of bridges eligible for federal aid. "MinnDOT may well have done everything prescribed, but that may not be enough," said Rosenker.
MinnDOT has hired Wiss Janney Elstner, Northbrook, Ill., to do an independent forensic investigation. The Army Corps of Engineers will lead the clean-up effort. MinnDOT officials are working on a design-build contract for emergency replacement. In a later press conference, Gov. Tim Pawlenty (R) said that divers had to cease operations for a while Thursday because the Army Corps slowed the flow of the river via a dam upstream, creating some suction. Several vehicles have been identified underneath the fallen bridge. Three other steel arch truss bridges, 106 "deficient"–rated bridges and the rest of the state's bridges as needed will be inspected.
Bearings corrosion and fatigue cracks in approach spans, since fixed, led to a non–dangerous "structurally deficient" federal rating for the I–35W bridge over the Mississippi River. Structurally deficient, under federal guidelines, means there are elements of the bridge that need to be monitored and/or repaired. At least 30 were still missing Thursday afternoon, a day after the 581–m–long eight–lane crossing collapsed during evening rush hour, sending vehicles tumbling 60 ft into the river.
There had been no evidence of additional or growing cracks in welded connections in inspections conducted between 2004 and this year, say Minnesota Dept. of Transportation officials. Based on a study by URS Corp., MinnDOT began a thorough inspection of welded connections inside the two steel arch deck trusses this year, Dan Dorgan, MinnDOT director of bridges and structures, said at a press conference. The inspections began in May and were to be completed after Progressive Contractors, Inc., finished its $9–million contract in September. St. Michael, Minn.–based PCI, which was performing routine resurfacing work, was still missing 1 of 18 workers as of late Thursday.
"Our bridge repair crew of 18 workers was performing routine resurfacing work on the Interstate 35W bridge across the Mississippi River," said PCI bridge division vice president Tom Sloan in a statement. "They were preparing to pour the final two inches of concrete surface. Before they could start pouring, the bridge collapsed. One of our workers has not been located. Three were hospitalized, and others were treated at the scene. All of the survivors are severely shaken up."
He added: "This was a steel structure that supported a concrete deck. We did not do any work on the steel structure. As we have done successfully many times all over the Midwest, we simply repaired the concrete deck by removing deteriorated concrete, patching, and resurfacing."
This bridge is unique because it was constructed with a single 458 ft steel arch to avoid putting piers in the water which would impede river navigation, according to the U.S. Dept. of Transportation.
The bridge was built between 1964 and 1967 by Industrial Construction, says Dorgan. The bridge, designed in house, consists of steel multi–beam approach spans, concrete slab approach spans and steel deck truss main spans. According to a 2001 report by the University of Minnesota that examined fatigue cracking, the twin steel deck trusses comprise a latticework of riveted plates and I–beams for diagonal and vertical members. "The truss members have numerous poor welding details. Recent inspection reports have noted corrosion at the floor beam and sway brace connection, and rust forming between connection plates," said the report.
The investigators found that peak stress ranges were less than the fatigue thresholds at all details in both controlled and open traffic tests. If fatigue problems were to develop due to increased loads later, cracks would first show up in a floor truss, where they "should be readily detectable, floor trusses are easy to inspect from the catwalk." If cracks went undetected, it added, the bridge could most likely tolerate the loss of a floor truss without collapse, "whereas the failure of one of the two main trusses would be more critical."
The report recommended that the main truss members it identified as having the highest stress ranges be inspected thoroughly every two years. It also recommended inspections of reinforcing welds on lower chords and diagonals of all the floor trusses. "Since they can be inspected easily from the catwalk, they could be inspected every six months," the report concluded.
Speaking on a general level, Stuart Sokoloff, principal with Construction Technology Services, Garden City, N.Y., which provides forensic engineering and failure analysis, noted that "most often when there's a collapse it has to do with the connections and not the actual steel girder." He also notes that "welded connection failures tend to be catastrophic," but are not as easy to find in a visual inspection. "If there is direct shear force on a bolted connection, you'd start to see a prying deformation. A visual hands on would identify it... For a welded connection, unless it underwent some sort of failure, you can't tell what strength is left."
In a hypothetical case, repair work could induced lateral forces that push a truss out of line, he notes. "Trusses are meant to be 2–dimensional entities. The deeper you go from the chords, the stronger the truss you get." However, something like a jackhammer or simple resurfacing doesn't seem likely to cause such a catastrophic failure, he adds.
Richard Stehly, an expert in bridge engineering and co–founder of St. Paul, Minn.–based American Engineering Testing notes that PCI crews were doing maintenance work on the bridge – – patching, partial depth replacement, milling and overlay. " It's hard to see how a partial–depth replacement would be a cause," he says, though nothing can be ruled out.
The critical problem, as noted in the 2001 report, is the lack of redundancy in the truss design. "Two planes of truss support the eight lanes of traffic. If you lose one you lose the whole thing," he says. "The design was fairly common" 40 years ago, he says. "Today you wouldn't build a truss like that."
The possibility has been raised of a train causing vibrations on a track adjacent to piers. While not ruling that out either, Stehly notes that it is a track siding, where trains aren't likely to be going more than 10 mph. But typically, "there's no one factor" in a situation like this.