SHEAR FAILURE at temporary supports is being blamed for the catastrophic collapse of a partially constructed concrete box girder bridge in South Africa last July.
Early investigations into the disaster had pointed to insufficient temporary prestressing in the incrementally launched deck as the most likely cause (NCE 16 July).
But according to sources close to the project, consultants working for Concor Construction and the bridge's designer VKE Engineers have now agreed that the collapse was triggered by the lower concrete flange failing above temporary bearings on the intermediate pier.
The collapse happened in Mpumalanga province in the north east of the country as main contractor Concor Construction was close to completing the second span of the seven-span, 300m long Injaka bridge. Fourteen people died - among them the bridge's 27 year old designer - and 13 others were left seriously injured.
The recent conclusions by experts for the contractor put the collapse down to the inadequacy of the junctions between the lower flange and webs of the deck segment to resist shear forces from the 1m2 temporary bearings on the bridge's second pier.
Their investigations have shown that failure occurred first at the western bearing over pier 2, causing the deck to list sideways. According to a Concor Construction worker who was on a temporary platform next to the bearing before jumping to safety, the deck fell 300mm to the top of pier 2.
Moments later, the latest theory suggests, concrete sheared at the eastern bearing, effectively removing the support at pier 2 (see diagram). This meant that the 90m long deck was propped only at the abutment and by the steel launching nose resting over pier 3.
The dramatically increased sagging moment in the deck would then have caused the 27m long steel nose to buckle and bend 45degrees a third of the way from its tip, and the deck to fail over pier 2. The rest of the deck would then have collapsed progressively, dropping workers and a party of site visitors 30m to the ground.
Latest investigations do not explain why the segment failed where it did. Possibilities include insufficient reinforcement at the fillet to resist forces from the bearings, or a localised problem with concrete quality.
A third potential cause is that the temporary bearings produced higher than expected local shear stresses because their size or location was different to that assumed in the original design.
Responsibility for the accident is also yet to be determined. The formal inquiry was due to begin on 30 November (NCE 26 November), but was postponed after VKE's legal team claimed the chairman might have been influenced by consultations with the official investigating engineer.
A new chairman, South African Department of Labour senior official Larry Kloppenborg, has been appointed to the inquiry, which will start on 25 January at Graskop magistrate's court, 25km south of the accident site.
Proceedings will be moved on 1 February to Pretoria Industrial Court, 400km south west of the site, to accommodate VKE and Concor Construction's legal teams and expert witnesses who will be called to testify.