Federal officials concentrate on a possible design flaw and warn bridge engineers nationwide to watch weight of construction work.
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.
When the I-35W bridge fell during evening rush hour last Wednesday, Progressive Contractors Inc. (PCI) of St. Michael, Minn., was at work on a $9 million job to make surface repairs. Crews from the well-known contractor had been on the bridge for six weeks.
Piles of sand, gravel were on bridge
An aerial photo taken about three hours before the accident showed several large company trucks, along with sand and gravel piles, in the two closed southbound lanes just south of the center of the bridge. The company has insisted the repairs did not cause the bridge's failure.
The photo also shows traffic in each direction confined to two lanes, with the remaining two lanes in each direction closed for construction. At the time the photo was taken, a string of widely spaced traffic can be seen moving in each direction.
At Wednesday's MnDOT news conference, construction engineer Liz Benjamin confirmed that workers had dumped large sand and gravel piles on top of the bridge the day of the collapse.
As had been routinely done on previous days, those piles were put on the bridge before 2 p.m., according to a source familiar with the construction project. Those piles were big enough to make concrete for a section equal to a quarter of the length of the bridge and a quarter of its width, the source said. Even if the piles weighed 100 tons, the source said, it would be lighter than the weight of three loaded semitrailer trucks.
"We continue to believe that our concrete repair work was routine, that it was done well, and that there was nothing about it that should have caused a bridge to collapse," Tom Sloan, a vice president for PCI, said in his first comments since the collapse.
He said the company had interviewed every employee on the bridge when it collapsed and had not found anyone who could substantiate reports from a first responder who said workers told him the bridge was wobbling before it fell.
The steel truss underside of the 40-year-old bridge was built with hundreds of gusset plates -- metal plates of various sizes where vertical, horizontal or diagonal beams are tied together. The NTSB said Wednesday that its investigators "observed a design issue" with the plates but wouldn't specify the location or explain the potential flaw.
The source familiar with the construction project said investigators are questioning whether the plates should have been thicker in the original design.
State transportation officials shed no light on the topic. From the outset, all parties to the investigation have noted that the failure of one connection or member in the bridge's steel superstructure could have caused it to fall because the bridge was built without redundancy.
Plates prone to fatigue
The New York Times reported that a consultant hired by the state after the collapse to investigate what had gone wrong first discovered the potential flaw. However, a MnDOT spokesman said he had no knowledge of that being accurate and that nobody from the consulting firm had told MnDOT of such a finding.
"No one from the firm has informed MnDOT of that finding," MnDOT spokesman Kevin Gutknecht said. He said he did not know the reason why the firm, Wiss, Janney, Elstner Associates Inc., was hired by the state. He said the firm was retained late last week.
A bridge expert intimately familiar with the I-35W bridge said the structure's gusset plates are prone to fatigue caused by twisting or bending of a truss member.
"What gusset plates tend to do on beam bridges is they create out-of-plane bending -- things work back and forth," the expert told the Star Tribune. "The beam is not designed to see that kind of stress and that tends to increase fatigue-prone details."
The expert said that the NTSB's interest in studying the material makeup of the gusset plates could mean that investigators are checking the quality grade and strength of the steel.
"A gusset plate is fabricated out of what would be considered a lower quality, a lower tensile strength steel than what it's being attached to," the expert said. "... Some places of the bridge matter less than others -- a secondary piece of bracing wouldn't have to be of the same strength as, say, a main member."
What's crucial, the expert said, is how the gusset plates were attached. When the I-35W bridge was built in the mid-1960s, welds were commonly used to fasten truss members. Today, gusset plates are usually bolted to truss members, the expert said. The expert said that bridge inspectors considered the welding of the plates to be a design flaw in the bridge.
"A majority of the repairs we see made are due to poor design detail," the expert said.
In the latest Minnesota Department of Transportation inspection report of the I-35W bridge, deficiencies were found in at least seven gusset plates. The flaws included heavy flaking, pack rust, pitting, section loss, a weld overlap and one case of loose bolts, according to the June 2006 report. Some gusset plates on the bridge are as wide as 5 feet.
At an afternoon news conference Wednesday, Bob McFarlin, MnDOT's assistant to the transportation commissioner, would not respond to whether the investigation into the bridge collapse is now focused on the plates or whether the latest NTSB report leads in that direction.
"I don't want to rush to judgment," McFarlin said.
Other areas of focus
The NTSB emphasized that the gusset plates are only one area of inquiry. It is working with the Federal Highway Administration to conduct a computer-aided structural analysis of the bridge and that part of the investigation is expected to take "several months," the agency said Wednesday.
"We are continuing to make progress on this investigation, and each area of inquiry gets us closer to ultimately determining the cause of this tragedy," NTSB Chairman Mark Rosenker said.
The NTSB's 19-member team interviewed vehicle occupants, construction employees and crew members on a dinner cruise ship that was in the Mississippi River lock near the bridge at the time of the collapse, the agency said.
On Monday, the NTSB team used a helicopter to shoot high-resolution pictures of the fallen bridge superstructure at its north end. Several fractures were observed, but nothing that looked like a failure that could have caused the collapse, the NTSB said.
A MnDOT fact sheet released Wednesday said that while the bridge roadway slab was 9 inches thick, PCI in most instances removed only the top 2 inches and replaced it with new concrete.
"Other concrete removal work was done using 45-pound jackhammers. Nothing larger was used to remove the concrete," the agency stated. MnDOT officials said the company had stopped using the jackhammers earlier in the day or the day before.
Terry Williams, a NTSB spokesman, said he could recall no instance in which the NTSB had concluded that an existing bridge had collapsed because of repair work.
The speculation whether the repair work was a contributing factor meanwhile continued.
Kent Harries, an assistant professor of structural engineering and mechanics at the University of Pittsburgh, said from the pictures he has seen, it does not appear that there was an excessive amount of equipment on the bridge at the time of the collapse. He also noted that commuter traffic was down to two lanes in both directions, so the bridge was handling about half of its normal weight load.
Staff writer Rochelle Olson contributed to this report.