›

Part 1 20 TheStructuralEngineer Professional guidance November 2014 Quebec Bridge collapse The Quebec Bridge collapse: a preventable failure (part 1)

the Quebec Bridge Company in 1887. The He also insisted on maintaining full control of company’s role was to plan and deliver the the project, and when it was suggested that bridge, and the technical challenge was his work should be subject to peer review, he signifi cant. At its narrowest point the St refused, claiming, “this puts me in the position In the fi rst of a two-part Lawrence River was 3.2km wide and up to of a subordinate, which I cannot accept”1. article, Sean Brady takes 58m deep. The river was fast moving (as Furthermore, the poor fi nancial position of readers through the fast as 14km/h), had a tidal range of 5m the project would result in Cooper agreeing and, during winter, ice became stuck in the to charge only half his usual fee, a decision chain of events that led channel and piled up to a height of 15m. that essentially cost him his ability to hire to the collapse during The team that would eventually deliver the appropriate staff . construction, with 75 structure fi rst met at an American Society of He did, however, hire Norman McLure as Civil Engineers (ASCE) meeting in Quebec in his on-site inspector, a decision that would deaths, of the Quebec 1897, where the Quebec Bridge Company’s have far-reaching consequences. Due to ill Bridge. principal engineer, Edward Hoare, met the health, Cooper didn't actually visit the site chief engineer from the Phoenix Bridge after superstructure construction began, Company. Controversially, Phoenix off ered to instead relying on the young McLure as his A stark example of how technical and human complete the initial design of the bridge for eyes and ears. factors can combine and cause failure is that no fee, provided it was awarded the contract Inexperience constructing a bridge of this of the Quebec Bridge, which collapsed twice for the bridge’s fi nal design and construction. size was not, however, limited to McLure. during construction: the fi rst time in 1907, This is indeed what occurred, with the Hoare had never worked on a bridge resulting in the deaths of 75 workers; the Phoenix Bridge Company’s proposal selected longer than 90m — hence his desire to second failure occurring in 1916 with a further from the six proposals submitted, being cited retain Cooper as a consultant — and the 13 fatalities. as the “best and cheapest”1. Phoenix Bridge Company was also relatively The Quebec Bridge was one of the world’s At the same ASCE meeting, these men inexperienced for a bridge of this scale. most ambitious bridge engineering projects. would also meet the eminent bridge engineer, Worryingly, it had experienced a number of It was overseen by one of the world’s Theodore Cooper, who off ered his services collapses during construction on previous most celebrated bridge engineers and its as a consultant. In time, Cooper would also projects. collapse would reverberate through the be selected as the chief consultant for the construction industry for decades to come. project, overseeing all of its technical aspects. Construction Its aftermath would see the formation of the Cooper’s reputation as a bridge engineer Construction of the superstructure began in American Institute of Steel Construction was immense. He had been an inspector July 1905, with a signifi cant technical issue and the American Association of State during the construction of the Eads Bridge manifesting itself early the following year. Highway Offi cials, and it would also give rise at St Louis, which pioneered the use of steel The steel members arriving on-site were to the tradition of all graduating engineers in major bridges, and was assistant general heavier than expected. McLure brought the in Canada receiving the Iron Ring, a ring manager of the Keystone Bridge Company issue to Cooper’s attention and a self-weight worn on the pinkie fi nger of their working before setting up his own consulting calculation error was identifi ed. hand to remind them that, regardless of practice in New York in 1879. In addition to The initial design span of the bridge was the sophistication of modern structural his practical expertise and experience, he 488m. However, in 1900, after completion of engineering, their fundamental duty is to made a signifi cant academic contribution to the tender process, Cooper increased this prevent failure. bridge design. He developed a methodology span length by 61m to 549m. There were In Part 1 of this two-part article, we will for dealing with railway loads on bridges, sound technical reasons for the change, examine the events that led to the fi rst and he wrote papers promoting the use including that an increased span resulted collapse of the iconic bridge. In the next of steel as a bridge material and prepared in piers being closer to the shore, with a issue, Part 2 will explore the complex general specifi cations for iron and steel subsequent reduction in construction depth. interaction of human and technical factors bridges. At approx. 60 years of age, Cooper There is speculation, however, that an that contributed to the catastrophe. felt the Quebec Bridge project would be the altogether more human desire contributed to crowning achievement of his career2. Cooper’s decision3. The initial span of 488m Crossing There would be a number of interesting was shorter than the 521m span of the Forth While the idea of bridging the St Lawrence aspects to Cooper’s engagement. Despite the Bridge in Scotland, which was completed River had been around since the 1850s, it fact he would be located in New York, over in 1890. Increasing the span length of the was only in the 1880s that interest gained 800km from the site, he accepted the role of Quebec Bridge to 549m would make it the momentum, culminating in the formation of inspecting the steel fabrication and erection. longest cantilever bridge in the world — as it

TTSE35_20-21SE35_20-21 Health&Health& SafetySafety QuebecQuebec v3.inddv3.indd 2020 223/10/20143/10/2014 11:0911:09 www.thestructuralengineer.org 21

Figure 1 Quebec Bridge prior to collapse

remains today. Whatever the reason, it was discovered that the Phoenix Bridge Company had carried out the self-weight calculations for the fi nal design based on the initial proposed length of 488m and had not updated its self- weight calculations for the revised length. The actual bridge would weigh 325MN, as opposed to the initial calculated weight of 276MN, an increase of 18%1. Cooper would

assess that the calculation error increased SKITT MATTHEWS JAMES the stresses in the bridge by 7%, yet he concluded that the design was adequate and southern cantilever arm suspended 223m had occurred prior to installation. construction could proceed1. over the St Lawrence River (Figure 1), the Cooper’s message to cease construction However, evidence that the design may not scene was set for catastrophe. would arrive at the offi ce in Phoenixville at have been adequate began to present itself 1:15pm on 29 August. It was ignored; the in June 1907, 23 months into construction Collapse principal engineer was out. He returned at of the superstructure. McLure identifi ed On 27 August 1907, a foreman noticed 3pm and, despite being aware that work was that a number of the bridge’s compression the signifi cant bowing and immediately still proceeding on-site, simply waited for members had mid-point defl ections — suspended construction. McLure sent McLure to arrive that evening at 5:15pm. They they were bowing. This bowing was most a message to Cooper advising him that discussed the issue, then decided to make a pronounced in the compression members construction was suspended until Cooper decision the following morning. on either side of the bridge piers (known reviewed the matter. The following day, The decision came too late. At 5:30pm that as A8L and A9L). These members carried McLure caught the 1pm train to New York to evening, while the meeting in Phoenixville some of the largest compression loads in discuss the matter personally with Cooper. was wrapping up, the critical A9L member the bridge. In early June, it was found that When Cooper arrived at his offi ce at 11am buckled, initiating the progressive collapse member A9L had a bow of 1–2mm, but later the following morning, 29 August, McLure of the 17 000t southern arm of the bridge that month these defl ections had grown to was waiting for him. They had a discussion into the St Lawrence River. The bridge took 19mm. McLure notifi ed both Cooper and the and Cooper sent a message to the Phoenix 15 seconds to collapse and was heard over Phoenix Design Company’s chief engineer. Bridge Company in nearby Phoenixville 10km away in Quebec City. Only 11 of the Phoenix was of the view that this bowing stating that no further load was to be added 86 workers on the bridge at the time would was pre-existing and present when the to the bridge until he had resolved the survive, with a number of bodies never being members left the fabrication shop, thus being situation. recovered. unrelated to in-service stresses. McLure McLure was then to travel to Phoenixville disagreed: he was convinced the bowing to meet the Phoenix Design Company that Sean Brady is the managing director of Brady had only appeared after the member was evening. He assured Cooper he would send Heywood (www.bradyheywood.com.au), based in installed, a view supported by the fabrication the same ‘cease construction’ message to Brisbane, Australia. The fi rm provides forensic shop. Cooper appears to have been the bridge site, but he didn’t, and the site and investigative structural engineering confused by the bowing — the possibility remained unaware of Cooper’s wishes. services and specialises in determining that the bridge was overstressed was simply While both Cooper and McLure proceeded the cause of engineering failure and non- not considered1. He believed that the bowing on the basis that work had been suspended performance. may have been present prior to installation, on the site since 27 August, this was not the

but also speculated that these members may case. Unknown to them, the same foreman REFERENCES: have been impacted by other members being changed his mind and recommenced work 1) Delatte N. J. (2009) Beyond failure: lifted into position during construction. He the next morning — the same day McLure Forensic case studies for civil engineers, asked McLure to check for evidence of such caught the train to New York. He changed Reston, USA: American Society of Civil impacts. McLure found none and the cause his mind because prolonged delays or Engineers remained unresolved. stoppages would result in a ‘decrease in But by late August 1907, there was no morale’ on-site, with the risk of workers 2) Middleton W. D. (2001) The Bridge at Quebec, Bloomington, USA: Indiana University Press doubt that an issue existed. In the space of leaving the project. He also proceeded two weeks the bow in the A9L member had on assurances from Phoenix, which still 3) Åkesson B. (2008) Understanding bridge increased from 19mm to 57mm1. With the appeared convinced that the member bowing collapses, London, UK: Taylor & Francis

TTSE35_20-21SE35_20-21 Health&Health& SafetySafety QuebecQuebec v3.inddv3.indd 2121 223/10/20143/10/2014 11:0911:09