Condemned to repeat?

The hotel had only been open for a year when the tea-dance, a popular Friday event to kick off the weekend, turned into a living nightmare. It was just past 7pm and the band was in full swing. As many as 2000 people had gathered to dance and to watch. Suddenly there was a loud crack and two suspended walkways, one at the fourth floor level and one at the second, and both with dancers on them, fell onto the crowd in the atrium. It was the US's most devastating structural failure: 114 died and more than 200 were injured.

Investigators had no trouble identifying why the "skywalks" fell. Vertical rods attached to the ceiling held up the fourth-floor skywalk. Those rods screwed into box beams on the skywalk. Another set of vertical rods hung from the fourth floor box beams and supported the second floor skywalk. This doubled the pressure on the fourth-floor box beam connections.

This design was a late change. The original design had long vertical rods hanging from the ceiling with each skywalk clinging to them. But this presented assembly difficulties so the subcontractor changed the design and the engineers, disastrously, approved them. To understand the effect, one should imagine a man hanging onto a rope with another hanging off his ankles. The first man might cope with his own weight, but not the weight of two.

In 1984, two engineers were found guilty of gross negligence, misconduct and stripped of their licence to practice in the state of Missouri.

Could this happen today? Scarily, the actions leading to the disaster were almost identical to those leading to the Tay Bridge collapse 100 years earlier (see page 18). In both cases the design of connections had been largely left to the contractors. Frankly, a similar disaster is possible today although the introduction of the CDM regulations, which require a planning supervisor to co-ordinate all elements of a project, should, in an ideal world, reduce this risk. Repeat: in an ideal world.

1970: Erection mistakes of novice contractor kill 35 in West Gate Bridge collapse, Australia

The contractors had never built a bridge like this before and when it started to bend out of shape they employed such Heath Robinson remedies that it brought the 110m span of the bridge crashing down in Victoria, Australia.

An enquiry found that the erection method proposed by the original contractor required extra special care. But when another, less experienced contractor finally came to do the work, the designers didn't check the proposals properly, nor did they step in to put a stop to the dangerous procedures.

Among the 35 dead were the designer's resident engineer and site engineer.

The contractor decided to construct the bridge in two longitudinal halves. This created an asymmetric section and, for whatever reason, the contractor overlooked the implications. These two halves were then to be jacked into place and bolted together along the centreline of the bridge.

Asymmetric bending caused the two sections of the bridge to bow apart when they were jacked into place, making it impossible to bolt up the longitudinal splice.

The contractor tried various methods of forcing the two halves of the bridge together including using kentledge (typically one cubic metre blocks of concrete). This extra weight caused the bridge to buckle. The contractor then made a major error of judgement and tried to straighten the buckle by adding more kentledge and unbolting a transverse connection across the bridge.

The bridge started making alarming noises and, shortly afterwards, the weight of concrete and the unbolted transverse splice caused the weakened bridge to collapse completely.

Could this happen today? Contractors still have overall responsibility for ensuring a safe method of construction, and the CDM Regulations require that the methods of building are planned before work starts, but it really depends on whether such an inexperienced contractor could blag their way into such a responsible job, and be unscrupulous enough to ignore what the CDM Regulations stipulate.

1907: Designer arrogance blamed for Quebec Bridge collapse that killed 82

The whistle had just blown signalling the end of the day for workers when a sharp report like a cannon shot stopped everyone in their tracks.

Dumbfounded, they watched 19,000 tons of steel forming the south anchor, cantilever arms and partially completed centre span of the Quebec Bridge slide into the St Lawrence River, killing 82 workers.

The bridge, designed to be the longest in the world at that time, was to have been the crowning glory on the career of Theodore Cooper, the most renowned bridge engineer in North America. It turned out to be a shameful monument to his ego.

Cooper's first action was to change the design by moving the piers closer to shore, thus increasing the length of the bridge from 1600 to 1800 feet.

There were two reasons for this. First, piers constructed in deeper water would be vulnerable to heavy ice floes. Second, building the piers closer to shore would be quicker and cheaper.

The extra length of the bridge was accommodated by changing the specification to allow higher stresses in the structure. He did not recalculate the weight of the bridge. A senior government engineer suggested his department check the calculations but Cooper dismissed this suggestion on the grounds that this would make him a subordinate.

In February 1906 it was revealed that the weight of steel in the bridge had increased by nearly 5000 tons, almost 20% of the original estimate, increasing the already high stresses in the bridge by about 10%. But Cooper allowed construction to continue as the only alternative was to start building the bridge again. Eventually the stress in the bridge became too great and the almost completed structure collapsed.

The Royal Commission of Inquiry blamed Cooper for failing to carry out proper calculations. As a result, Canadian engineers now wear iron rings to remind them of their responsibilities. Now, no individual engineer could behave like such a prima donna, but 60 years later the inquiry into West Gate Bridge found a similar catalogue of errors.

1879: Contractor design weakened the Tay Bridge before a train plunged into the firth, killing all 75 passengers on board

He received a knighthood for his endeavours but 19 months after his bridge was built, Sir Thomas Bouch was blamed for causing the worst structural collapse in British engineering.

At 7.15 pm on 28 December 1879 the central spans of the Tay Bridge collapsed during a storm as a train was crossing, taking all 75 people on board to their deaths. The bridge was nearly two miles long with 85 spans giving an 88ft clearance above high water in the navigable channel and making it the longest bridge in the world at the time.

The exact cause of the collapse is still debated. The Court of Inquiry did blame Bouch for failing to make adequate allowances for wind loading but it wasn't that simple. He had originally designed a masonry bridge but discovered that the ground was weaker than expected, requiring more expensive foundation piers.

Bouch decided to use lighter, cast iron piers with a longer span in between to overcome the problem. Each pier was made of six cylindrical cast iron columns, arranged in a hexagon.

Critically, Bouch left it up to the contractor to figure out how these columns would be connected. He also neglected to specify a ring beam to connect the columns at their tops.

The contractor decided to cast the columns with integral lugs, into which the network of bracings could be bolted. But the foundry (owned by the contractor) did a poor job. There were air bubbles in the metal and the bolt holes didn't line up. The poor workmanship could have been caught by supervision but, alas, supervision was poor.

The investigation found that almost all of the lugs broke off the columns before the collapse, leaving each pier vulnerable to splaying open like a petal. What caused the lugs to fail is still not clear. Gale force winds? The train travelling too fast, or hitting a kink in a girder, thus jarring the whole bridge?

The debate continues, but what is clear is that asking a contractor to design such an important part of the structure left the bridge terribly vulnerable.

Although a structural collapse of this magnitude has not occurred since, the conditions do exist for a repeat performance: compromises to save money are still made, poor workmanship and inadequate supervision still occur and important elements of design are still delegated to contractors.

1994: Bizarre coincidence of wrong assumptions kills six in Ramsgate ferry walkway collapse

Only four months after it had been built, fatigue cracking caused a passenger walkway to collapse just after midnight, sending six people trying to board a ferry plunging 10m to their deaths. Seven were injured.

It was a bizarre coincidence of wrong assumptions that no one detected. A month earlier an insurer's inspector found the walkway, which connected the terminal in the Port of Ramsgate to the ferry, to be in good working order.

The HSE's investigation found that both the designers and the design checkers had made almost identical conceptual errors. The walkway had been designed to rise and fall with the tide. It was built from box section trusses and connected at the seaward end to a frame on a pontoon using horizontal axles that were free to rotate. Similar connections were used at the landward end but there the supports were designed to slide as well. This allowed the walkway to move up and down, but the walkway was torsionally stiff and no allowance had been made for rotational movement that, in practice, caused the walkway to lift off its bearings.

They had both assumed that the connection between the axles and the walkway would only be subject to shear forces and not to bending and they had also overlooked the effect of torsion and assumed that all four axles would carry equal loads. It was also found that lubrication of the bearings was deficient partly as a result of poor design and partly because some automatic lubrication devices had been left off.

The designers, fabricators, checkers and The Port of Ramsgate were charged under the Health and Safety at Work Act and all found guilty. But the possibility of competent and experienced designers and checkers making the same simple mistake was considered so remote that no changes in working practice were recommended. It's a chilling thought, but a similar failure could happen again.