An extensive subway system with lines fanning out in many directions speeds transportation in small, crowded Singapore. On April 20, 2004, during construction for the new Circle line of the subway, a deep excavation suddenly collapsed. The catastrophe killed four people, twisted steel beams, swallowed two construction cranes and knocked out a substantial chunk of the main highway running over the tunnel.
"This was one of the largest failures of a civil engineering project under construction in about 50 years," said Professor Andrew Whittle of the Department of Civil and Environmental Engineering in a Jan. 9 IAP class, "What Caused the Collapse of the Nicoll Highway in Singapore?"
Singapore authorities promptly began an intensive inquiry into the disaster, and Whittle was chosen by the Singaporean Land Transit Authority to be one of four international experts to investigate the accident and determine the causes. During what Whittle called "some of the most intense meetings I've ever attended," the experts studied the evidence and wrote a report on the collapse. Twenty ongoing excavation projects were temporarily shut down for design-practice reassessment, resulting in very large costs associated with delays and changes.
"Failure is always a combination of factors," said Whittle, who listed a series of cascading errors.
The original design misinterpreted the local geology and overestimated the soil shear strength in its analysis. The structure was therefore underdesigned to resist lateral earth pressures. Excavations for the subway extended more than 100 feet below the ground surface where the strongest marine clay was "weaker than the weakest clay in Boston," Whittle said.
There were also errors in detailing the structural connections for the structural bracing system; the collapse occurred when one level of bracing was overloaded and there was inadequate capacity to redistribute the loads among the remaining supports. Although large wall deflections occurred during the excavation, the measured strut loads were smaller than expected. As a result, the project engineers were apparently unaware of the potential for a catastrophic failure.
The complete findings from the official committee of inquiry were published in a 460-page report with detailed graphs and photos.
A version of this article appeared in MIT Tech Talk on February 6, 2008 (download PDF).
