The 35W bridge's undersized gusset plates were staring bridge inspectors and engineers in the face for the 40 years of the bridge's life. They inspected the bridge, watched it age and documented it with photos.
But inspectors weren't trained to know that under-sized gussets posed a threat to the bridge. As NTSB Acting Chairman Mark Rosenker said, at the time, even computer modeling didn't consider gusset plates.
"These finite, element analysis that had been done over the years would never have found this deficiency," Rosenker said.
NTSB investigators found that the bridge designers apparently did not calculate how much load the half-inch gusset plates could carry. The original designers of the bridge, Svedrup and Parcel, certified the design and handed it to MnDOT without ever determining how much weight the gusset plates could withhold. Then, MnDOT examined the designs, but not in detail. The problem slipped into the background until August 1st, 2007.
Member Debbie Hersman said this error of design was not just the fault of the bridge designer, it was also the fault of state and federal transportation officials.
"I think that you have to rely on the expertise of others," Hersman said. "But as the stewards of the public dollar, and who are responsible for making sure that the project was built to specification and for the tax payers, there is an obligation for the government to trust and verify."
The other four members of the board agreed. They voted 5-0, that MnDOT bore a share of the responsibility for the bridge's failure with the bridge's designer. Even though at the time, states did not know much about gusset plates and weren't in the practice of reviewing designs in detail.
Acting Chairman Mark Rosenker said the board does not assign blame, however.
"We think it is important to recognize that there was oversight that should have been done and at that time it didn't move down, didn't drill down into the full detail that may have been able to find this," Rosenker said.
State DOTs to this day do not include gusset plates in an evaluation of a bridge's structural health, or in its load rating. The NTSB recommended state agencies do that. They also recommended that a system be put into place so that state DOTs have a quality control system in place to detect design errors.
In Washington, MnDOT Comissioner Tom Sorel said MnDOT still needs to digest the report, particularly the findings that MnDOT may have been able to avoid the collapse with better oversight.
"We've got to look at that a little further," Sorel said. "Because I think, as the board says, and as the members said many times, that at the time those were the procedures and the guidelines in place. So that's the reality of the situation."
But he said this is the definitive report and it has implications beyond the 35W bridge. The NTSB found that other states have also built bridges that had unnoticed design flaws. Board Member Kitty Higgins believes this is a legitimate problem.
"I'm not saying it's widespread but I don't think we can say it's not common," Higgins said. "I think we don't know. What we know is there are design errors that escape the review process and we're doing something about it."
The NTSB said corrosion was not a factor in the bridge's collapse, but investigators pointed out that it was a troubling safety issue on the bridge. They pointed out Friday that MnDOT's inspections failed to detect the extent of corrosion on the gussets.
The board also recommended engineers evaluate the gusset plates on all of their truss bridges and train their inspectors to recognize problems in gusset plates. Then, after roughly 15 months of investigation and 17 hours of deliberation, the NTSB approved their findings on the 35W bridge collapse.