Hassan Astaneh's analysis of the 35W bridge collapse concludes that MnDOT, the construction company PCI and URS Corporation, the consulting firm that evaluated the bridge, could have prevented the collapse.
Astaneh is a professor in University of California Berkeley's Civil and Environmental Engineering Department whose work includes studies on the collapse of the World Trade Center for the National Science Foundation. He's been hired as a consultant by one of the lawyers representing some of the victims of the bridge collapse. His paper on the 35W bridge disaster is the keynote address at International Conference on Steel Bridges. Astaneh provided MPR News with an advance copy.
Bridges are designed to bear one and a half to two times their intended weight, Astaneh says. That margin is a safety net that accounts for design errors.
MnDOT had already eaten into the bridge's safety net when it added two inches of concrete to the deck in 1998, Astaneh claims. The 35W bridge deck was nine inches thick. Normally a bridge's deck is six to seven inches thick, Astaneh says.
"The weight of the bridge comes mainly from the weight of the deck, the roadway. And of course the truss itself," he says. "So you cannot simply add another two inches, that's like 30-percent extra weight. That can translates to 10 to 15-percent weight of the total bridge."
In January the National Transportation Safety Board reported the bridge's gusset plate at node U10 was undersized. Last August, Astaneh's research found the same thing. But Astaneh says despite this design flaw, the plate would have held up if it weren't for the extra weight.
That is absolutely unbelievable, absolutely unheard of. Any bridge engineer would just be completely shocked. We never do that, especially steel bridges.
Construction company PCI also had loaded 270 tons of construction materials on the bridge on August first. PCI was re-decking parts of the bridge that afternoon. Even if U10 had been designed properly, that construction material should not have been on the bridge, Astaneh says.
"That is absolutely unbelievable, absolutely unheard of. Any bridge engineer would just be completely shocked. We never do that, especially steel bridges. It's like a car, a car has a capacity. It can take only two tons, 10 tons, 20 tons. There is a meaning to this capacity," he says.
In his analysis, the weight of those materials fractured gusset plate U10 and led to the collapse.
MnDOT and the NTSB have both said it is unlikely a bridge inspection would have discovered that the U10 gusset plate was under-designed and bowed. Astaneh disagrees. He says an inspection ladder is located next to the gusset plate.
"There is no way for any inspector to go and inspect this bridge, for any reason, and climb that ladder and not see, just two feet away, the gusset plate buckling," he says.
In a report submitted to the Legislature by the law firm Gray Plant Mooty, one engineer said he did notice the gusset plate had bowed, but he didn't report it.
URS Corporation also could have caught the bowed plate, Astaneh says. URS studied the 35W bridge from 2003 until 2007. Their engineers conducted a fatigue evaluation and redundancy analysis. MnDOT also hired the firm to supervise bridge inspections conducted by MnDOT. But URS did not consider gusset plates in its evaluations.
That practice goes against the Federal Highway Administration's standards for evaluating bridges, Astaneh says.
"The URS Corporation did not look into gusset plates at all. But the standard clearly says in a couple of places any time they talk about evaluating, or investigating or inspecting existing bridges they emphasize it has to be members and connections. Connections for truss bridges means gusset plates; that's the only connections they have," he says.
In an e-mail exchange with MPR News in late January, MnDOT's Lucy Kender stated that typically truss analysis does not include connections. She explained that was why the gusset plates weren't included in URS's evaluation of the collapsed bridge.