Minneapolis Bridge had been recently passed inspection

August 3, 2007
Minneapolis Bridge Had Passed Inspections
By MATTHEW L. WALD and KENNETH CHANG

The eight-lane bridge in Minneapolis that collapsed on Wednesday had been diligently inspected for years and had always passed, state officials said yesterday.

It did not, however, get stellar grades for its condition.

Additionally, officials said the bridge’s design had been considered outmoded for decades because a single failure of a structural part could bring down the whole bridge. About 11 percent of the nation’s steel bridges, mostly from the 1950s and 1960s, lack the redundant protection to reduce these failures, federal officials said.

Over all, the bridge was rated 4 on a scale of zero to 9, with 9 being perfect and zero requiring a shutdown. An inspection report last year said the supporting structure was in “poor condition,” far from the lowest category. Hundreds of other working bridges are in similar shape, but the report did indicate that the bridge had possible issues that needed to be regularly inspected.

The bridge has been inspected annually since 1993, but independent engineers acknowledged yesterday that there are well-known limits to how useful an inspection can be. Bridges, they said, are prone to a variety of problems, and some are hard to spot. At the Minnesota Department of Transportation, shaken engineers made it clear that they knew something crucial had somehow been overlooked.

“We thought we had done all we could,” said Daniel L. Dorgan, bridge engineer at the department’s bridges division. “Obviously something went terribly wrong.”

On Thursday, the United States Department of Transportation said it had told all states to inspect bridges similar in design and construction to the one that collapsed, or to review inspection reports. The department said there were 756 such bridges.

In 1982, a bridge inspector looked at the Mianus River Bridge in Greenwich, Conn., and did not see the metal-fatigued pin that would break nine months later, collapsing three lanes of Interstate 95 and killing three.

In 1987, a New York Thruway bridge near Amsterdam, N.Y., also had a clean bill of health, but inspectors had never gone underwater into the Schoharie Creek to look at the bridge’s footings, where flood waters had scoured the concrete base. When the footings slipped, the bridge fell. Ten people died.

Today, inspectors use ultrasound to check the pins in bridges similar to the Mianus one, and bridge footings receive much more attention.

"I think bridge inspections are the best they’ve ever been," said David Schulz, director of the Infrastructure Technology Institute at Northwestern University in Chicago.

The cause of the Interstate 35W bridge failure on Wednesday will probably become clear through metallurgical examinations of the wreckage, experts said, but recovering the metal parts will be delayed by the search for human remains and the need to keep investigators safe in the swirling waters of the Mississippi.

The bridge was undergoing repair work this summer, and Mr. Schulz said he would be stunned if the work did not play a major role in the collapse. “It’s too much of a coincidence,” he said.

But Mr. Dorgan said he saw no connection between the repair work, which was taking place mostly on the roadway, and the collapse of the steel support structure far below.

Mr. Dorgan said the bridge was believed to be in good enough shape until 2020, when it was due for either a major overhaul or replacement.

Parts of the bridge were considered structurally deficient because of corroded bearings and tiny metal cracks that had been spotted years ago but were considered stable. The rating of “deficient” is a common one that indicates the need for regular inspection and does not mean the bridge is dangerous, said Thomas D. Everett, a top bridge official with the Federal Highway Administration.

The most visible threat to a bridge is usually corrosion. But metal fatigue — the weakening of steel by the repeated weight of heavy trucks bouncing across the bridge surface — is harder to see. Bridges in northern climates are particularly vulnerable to metal fatigue because steel becomes more brittle and prone to cracking when it is cold.

“A crack is very difficult to observe visually,” said Steven J. Fenves, a guest researcher at the National Institute of Standards and Technology, part of the Commerce Department, and a professor emeritus of civil engineering at Carnegie Mellon University. “There may be paint over it, or maybe many layers of corrosion over it. It may be in an invisible place, in the second plate, not the outermost plate.”

The possibility that metal fatigue could cause a bridge to fail was not even considered by bridge engineers in the 1950s and 1960s, when the Minneapolis bridge was designed and built, Mr. Dorgan said. Research at Lehigh University in the 1970s showed that stresses could be much larger than had been thought. The I-35W bridge, which had been designed according to less rigid standards devised by the American Association of State Highway Officials in 1961, had components that would not be included in a bridge built today.

Fatigue cracks appeared in the approaches to the bridge, but no significant problems were detected in the center span. A study in 2001 by University of Minnesota researchers concluded: “The bridge should not have any problems with fatigue cracking in the foreseeable future.”

In a study completed in 2001 by the Federal Highway Administration, 49 working inspectors from around the country visually examined test bridges in Pennsylvania and West Virginia. The inspectors correctly identified fatigue cracks only 4 percent of the time.

Additional techniques like X-rays, ultrasound, magnetic particles or dye can help identify cracks.

Mark V. Rosenker, chairman of the National Transportation Safety Board, said yesterday that his agency would determine whether the criteria for inspections were adequate. “They may well not be enough,” he said. Or the procedures may be adequate but may not have been followed, he said.

Safety board investigators will use computer modeling to study the failure, Mr. Rosenker said, and will also reconstruct parts of the bridge from the wreckage to determine the cause. The board’s final report could be many months away, he said.