Hoan Bridge fails in Milwaukee
On December 13, 2000, motorists crossing the northbound approach span of the Daniel Webster Hoan Bridge in Milwaukee, Wis., reported that the bridge seemed to be "buckling." Transportation officials immediately closed the bridge to ascertain the reason for the bridge's 4-foot deflections. What they found was that two out of the three girders supporting the two northbound lanes had suffered an abrupt and total fracture, while a large crack developed in the third. All three cracks occurred simultaneously and at the same location on each girder. Not only did this situation call for immediate action in order to close the bridge to traffic and prevent injury, but also for a thorough investigation.
Flange fractures
History
The Hoan Bridge is a multispan, three-girder bridge constructed between 1970 and 1972. HNTB Corporation designed the structure. The girders are hybrid, noncomposite plate girders consisting of A36 webs with 400 series steel used in the flanges. Partly because the bridge was noncomposite and partly due to the somewhat large girder spacing of 24 feet, the girders were very large for this span. The bottom flange in the positive moment span of the center girder was a hefty 3 inches by 28 inches. These represent the largest flange dimensions, but at all other locations, flange thicknesses were no less than 2 inches. All three of the girders were about 10 feet deep. The lateral system of the bridge consists of K-bracing, with members that are considered very large for a bracing system, even with respect to standards at the time of construction. The span that failed was 217 feet long.
Flange fractures
The three-mile bridge was not actually in use until 1977 due to the construction of the surrounding roads. The bridge is part of Interstate 794 and carried about 35,000 vehicles per day, but it was designed to be part of a freeway loop carrying up to 85,000 vehicles per day. The estimated design life for this type of bridge is 50 to 60 years.
Investigation
The kind of abrupt fracture that occurred in the Hoan Bridge is the worst type of failure for any structure. If all three girders had fractured, the span would have simply collapsed. Current bridge and building codes are written to ensure that the members will yield and show large deflections before actually failing. This preferred failure mode clearly did not occur in the Hoan Bridge, and the reason it didn't was the primary question for investigators. There is no evidence or known record of this type of immediate fracture in multiple girders ever occurring. From an engineering standpoint, it was imperative that a reason for the near-collapse be determined immediately, so that bridges with similar characteristics could be sought out and repaired if necessary.
The failure of the Hoan Bridge was investigated by a number of different bodies. The Wisconsin Department of Transportation played a significant role, and outside specialists were consulted. The Federal Highway Administration's Turner-Fairbank Research Center in McLean, Va., Lichtenstein Consulting Engineers from New York, and Lehigh University in Bethlehem, Penn., were all involved.
Problem gussets
Failure
At first investigators believed either the steel used for the girders or the workmanship was of poor quality, and thought that this may have contributed to the bridge failure. Another potential factor considered was the temperature, since excessively cold temperatures can cause a brittle failure. Most experts also believed fatigue cracking had a large influence on the failure. This would be the most obvious explanation because the structure in question was a bridge, and fatigue cracks near the point of failure had been reported in a recent inspection.
The failure was reported at about 7:00 a.m., after a night when the temperature was four degrees below zero, according to the National Weather Service. However, Milwaukee experiences below-zero temperatures 15 days a year on average, and the bridge was in use during the two coldest days on record for the city, at 26 degrees below zero. By this rationale, the cold, as the sole cause of failure, could be eliminated.
A web crack
The forensics team conducted a thorough set of tests and analysis on the materials and actual construction. They then compared the results to the specifications for steel bridge design at the time, and determined that neither the steel nor the workmanship was responsible for the failure.
Results
From extensive modeling and investigation of the actual bridge materials at the microscopic level, investigators returned the astonishing result that the girders did not show any evidence of fatigue cracking. Fatigue, of course, may have been a minor factor, but the actual cause of the failure was deemed to be "constraint-induced fracture, (with the crack) initially in the web gap" of the interior girder, explained Bill Wright, high performance steel specialist with the Federal Highway Administration. The crack started to propagate in the gap between the web stiffener and the flange, where the lateral bracing framed into the girder. The weld detail at this location called for three welds, all perpendicular to one another, to run together.
At the time in question, it is likely that two heavy salt trucks passed over the bridge side by side. The combination of three welds at the web gap constrained movement in all three directions and caused a stress concentration.
The finding of the agencies involved in the investigation was that the cold weather caused the steel to be more brittle than usual. A large enough load was applied, not only to start a crack at the web gap, but to cause the crack to propagate all the way through two girders, including the extremely large flanges, simultaneously. A crack developed in the third girder but arrested in the bottom flange.
Dr. Robert Dexter, a structural engineering professor at the University of Minnesota, explained that "cracks have occasionally occurred, in bridge girders and other structures, due to such highly constrained welds. These cracks usually occur during fabrication and so they are usually caught long before a structure is put into use."
To relieve these constraints along the rest of the bridge and prevent a similar disaster, WisDOT modified the bracing scheme to allow for more flexibility. Currently there are ongoing projects, not only in Wisconsin but elsewhere, to check for this type of detail in other bridges. At this point, the Hoan Bridge appears to be an anomaly, and it seems this kind of near-disaster only occurs under very specific circumstances. The experience may have left some members of the public apprehensive, but the fact that the bridge did not fall is both a relief and a testament to the power of design techniques that engineers continue to improve upon.
—Nate Erpestad
(The release of a full technical report by the Wisconsin Department of Transportation is pending. If you would like a copy of the full report, please contact Nate Erpestad with CTS at 612-626-2862, or via e-mail at erpe0012@umn.edu.)