Gussett plates under scrutiny.
Only a week into the federal investigation of the deadly Interstate–35 bridge collapse, the probe seems to have taken more twists and turns than the mighty Mississippi.
Some of those twists have observers even more confused as to what caused a structural failure when the 1,907–ft–long bridge fell into the Mississippi River during evening rush hour on Aug. 1.
The latest development, potentially flawed gusset plates along the bridge's 988–ft–long three–span deck truss, calls into question the original design of the bridge, which stood for 40 years before failing. Investigators are also looking at whether or not those gussets were weakened by summer deck repairs.
Investigators with the National Transportation Safety Board say they have observed "a design issue" with the gusset plates "in particular locations," without giving any more specifics.
"It sounds like they are thinking that [the gussets] might have been the wrong size," says Gene Corley, senior vice president of Skokie, Ill.–based CTLGroup. Thickness and fastener quality are likely under scrutiny, he adds.
Jacobs Engineering Group Inc., Pasadena, Calif., which in 1999 bought Sverdrup Corp., did not return phone calls. The unit designed the bridge, before the structure opened in 1967, as Sverdrup & Parcel.
Prior bridge inspections over the last decade show that the structure had a combination of rivets and bolts in gusset plates, scores of which were located along the main parallel and floor trusses, and had a history of corrosion problems. But there is no mention of gusset fatigue.
Rivets, which are not widely used anymore, are "pretty undependable," Corley notes, because torque can't be checked as easily as with a bolt. "So there is more chance of getting a fatigue crack started," he says.
Even so, traditional "hot" rivets of the time "would be darn tight," says Richard M. Gutkowski, a professor of civil engineering at Colorado State University.
The problem of using rivets "is not an issue of tightness, it's an issue of not being able to establish the tightness," he explains.
Confusing the issue is modern evidence from computer analysis showing that gusset plates and other parts of bridges have unintended benefits, such as helping to strengthen a truss while tying it together.
Today's bridges are designed with a reduced safety factor, "because you have more predictability," Gutkowski explains. So gusset plates seem to be an unusual focal point of the probe.
"It's a head–scratcher, that's for sure," he says.
"I'd be a little bit surprised," echoes Nicholas Altebrando, bridge practice leader for New York–based STV Goup Inc. "Poor fatigue details inside the boxes were what people were worrying about, since they're not visible."
"Gusset plates are part of primary inspection," Altebrando adds. "You would inspect them all."
Even so, catching a problem with a riveted gusset plate, designed to hold truss joints together, isn't easy, engineers say, because stress may not show on the surface.
"Welds are checked, for sure, but the state of the strain and stresses in a gusset plate wouldn't be anything that's visibly evident," says Gutkowski.
Gusset plates constructed in the era of the I–35W bridge typically ranged from 1/2–in. to 3/4–in. in thickness and were hot–riveted, engineers say.
Over the years, inspectors found no cracks on the bridge's fracture–critical trusses. Several cracks were drilled, though, in the multi–girder steel approach spans.
Corrosion and bearing problems also needed attention, but the main truss superstructure showed no signs of fatigue that would cause a collapse, according to multiple reports.
Engineers say that most of the problems found in inspections are typical for an aging bridge. Without knowing what caused the collapse, "Inspection is most likely an issue on this bridge no matter what," Corley concludes.
Working around the submerged superstructure, forensic engineers are also digging deeper into the bridge deck for clues.
NTSB has taken core samples of the pavement "to get a better picture of the deck thickness," and is studying construction records for items that might show what kind of strain $9–million overlay project may have put on the bridge just prior to its collapse.
Though both gussets and deck repairs are mentioned together, NTSB has not yet said if they are linked. It also has not yet made any statements of what caused the catastrophe, which killed at least six people and injured dozens.
That report could take one year to 18 months, investigators say.
Yet soon after NTSB released its statement about the gussets on Aug. 8, the U.S. Dept of Transportation sounded an alarm for states "to carefully consider the additional weight placed on bridges during construction or repair projects."
"Given the questions being raised by the NTSB, it is vital that states remain mindful of the extra weight construction projects place on bridges," says Transportation Secretary Mary E. Peters.
Several structural engineers have said they think that paving activity is an unlikely factor in the collapse, however. The eight–lane bridge was rated for a full load of cars and trucks, hardly the same as what deck repairs in four lanes would require, they say.
And the workers "weren't even supposed to be near the main members," says Corley.
Mike McGray, president of St. Michael, Minn.–based Progressive Contractors Inc., ( http://www.pciroads.com/) which was performing the overlay project when the bridge collapsed, claims that his firm did nothing wrong.
"All I can say is that the procedures that we were using were the same procedures we used prior to the collapse," McGray says.
PCI had already finished repairing four lanes of deck and were in the process of fixing the other four then the bridge fell.
Though NTSB Chairman Mark V. Rosenker has called Progressive "an experienced bridge and deck repair contractor," the DOT's warning is the first official indication that the staging of materials, equipment and people on the bridge may have contributed to its failure.
Workers were getting ready to place material in the southbound lanes when the bridge fell. The majority of the work was staged on the southern side of the bridge, near the river's west bank.
"We've done thousands of these overlays," says McGray. "We just proceeded as we had in the past, always under the inspection of the owner."
Meanwhile, Minnesota Dept. of Transportation has announced five bidders for a design–build replacement project.
Also on the list is C.S. McCrossan Inc., Maple Grove, Minn.
MNDOT has not yet released the names of design firms bidding on the project.
Flatiron and Johnson Bros. are working with Tallahassee, Fla.-based Figg Engineering Group, according to Charlie Humphries, vice president of business development for Johnson Bros.
Costs for the new bridge are still unknown, though MNDOT says it expects the project to exceed $150 million.
Humphries says cost estimates are still too early to tell. "We need to get more parameters from the department," he says.