(Forensic investigations are carried out by hypothesizing the cause of a failure or accident based on the evidence available at the time - as limited as this might be, and revising the hypothesis as more evidence comes in. This successive hypothesizing and revising might be done several times during an investigation. The following is an example of this process.
Counsel benefits from a process like this early in a case – ideally before deciding to take the case, when an expert studies the evidence, then, based on the available evidence, identifies and evaluates the technical issues and the cost to investigate these)
The cross bracing was inadequate. I concluded that last March, a few days after the failure. (Ref. 1) I used the bridge failure to illustrate how a hypothesis – an idea, could be formed about the cause of an engineering failure based on very little evidence. In this case, all I had were some on-line photographs .
But, inadequate for what? To resist the service plus construction loads – weights and pressures on the bridge, as required by the Canadian Highway Bridge Design Code? If there’s no bracing at all - none, these loads would cause the girders to buckle sideways in the order of 375 mm, not the approximately 1,000 mm or more seen in the photographs.
The 1,000 mm was determined from the photographs by scaling like we do on maps. The known 3.0 metre depth of the girders at the middle section – see Sources below, is like a scale or ruler in the photograph. This depth is about the same as the spacing between the girders, maybe a little less. I decided the spacing was 3.5 metres. I saw that the girders had buckled about 1/3 of the way into the 3.5 metre spacing – 1,000 mm or more.
The 1,000 mm buckling indicates a greater load was acting on the girders than perhaps was required to be resisted by the Code. Where did the greater load come from to cause the 1,000 mm?
The only thing attached to the girder – at the top, that can be seen in the photographs is a sling at the end of a crane’s cable. The cable is attached to a crane’s telescopic boom. The boom would sway and flex a little in the wind that was blowing that night which would cause the sling below to tug on the girder – a point load in engineering. Construction cameras show the girders intact at 2:00 in the morning and buckled at 2:15.
I hypothesized last March that this repetitive tugging caused the girders to buckle like they did.
The 1,000 mm magnitude of the buckling and the fact that crane booms sway in the wind supports this idea that tugging on the girders was the source of the greater load.
It would be of interest to know if the bracing that was in place – and seen bent in the photographs, was adequate to prevent buckling, except for the 1,000 mm due to the tugging. There are simple calculations that bridge design engineers do to determine the bracing needed to prevent buckling in the order of 375 mm.
I did think about a sudden foundation soil failure causing the crawler crane to subside and the cable to tug on the girder as a result. I dismissed this idea because the crane had been there a while lifting 40 tonne girders into place. Foundation failure would have occurred some time before because of these heavy lifts if the foundation soils were inadequate.
The only way I would revise my hypothesis is to note that the crane operator did not contribute to the failure because he was not working.
The initial hypothesis
The bridge failed because the middle crane’s boom moved in the wind – possibly also due to the crane operator’s actions, causing the cable to periodically tug at the middle section of beam #6 and eventually cause it to bend. This caused the middle sections of beams #5 and #4 to bend as well because they were connected to #6 by some cross-bracing. The cross-bracing was inadequate to resist the force from the tugging indefinitely and eventually failed too. The middle sections of beams #3, #2, and #1 did not bend and fail because they were adequately cross-braced.
The bridge failed because the middle crane’s boom moved in the wind causing the cable to periodically tug at the middle section of beam #6 and eventually cause it to bend. This caused the middle sections of beams #5 and #4 to bend as well because they were connected to #6 by some cross-bracing. The cross-bracing was inadequate to resist the force from the tugging indefinitely and eventually failed too. The middle sections of beams #3, #2, and #1 did not bend and fail because they were adequately cross-braced.
I studied various photographs on-line including construction photographs taken at the time of the failure.
I spoke with Barry Bellcourt, the Road Design and Construction Manager for the City of Edmonton, a few weeks ago, also Bryon Nicholson, Manager of Special Projects. Barry mentioned the litigation and the city’s position.
I also learned from him that the bridge consists of seven, 40-tonne girders. Each girder consists of two 7.5 metre long end sections and a 43 metre middle section. The end sections are 4.5 metres deep arching up to 3.0 metres at the middle section. The sizes are approximate.
I saw and photographed the underside of the repaired bridge girders from Groat Road in early August when I was in Edmonton.
I understand it was windy the night the girders buckled and that was the reason workers were not on the job.
I spoke with four companies in Nova Scotia that operate cranes. I learned that crawler crane booms move in the wind; flex and sway. There is greater movement sideways because there is less strength that way. Telescopic booms move more than lattice booms because of the greater surface area. Booms are lowered to the ground in strong winds. One company doesn’t operate its cranes in winds of 50 km/hr or more.
I also talked with Amjad Memon, a structural engineer with the Nova Scotia Department of Transportation, about the Canadian Highway Bridge Design Code.
- Wind, construction crane and inadequate cross-bracing caused Edmonton bridge failure: An initial hypothesis. Posted March 27, 2015
- Why, in a recent blog, didn’t I seem to consider foundation failure as a possible cause of the Edmonton bridge failure? Posted April 3, 2015
- Bridge beams that fail are sometimes like balloons filled with water – squeeze them and they pop out somewhere else. Posted May 20, 2015
- Google: Edmonton bridge failure, Groat Road, Buckling, etc. to see photographs of the buckled girders.