Learning from Bridge Failure

The Institution of Civil Engineers are hosting a seminar on Learning from Bridge Failure in London on Wednesday 24 September 2008. The agenda includes talks and discussions on past failures, sharing knowledge between public bodies, the Minneapolis bridge collapse, press reporting of failures and how things could be improved in the future.

With the notable recent exception of the collapse of the De la Concorde overpass in Quebec (pictured), which was investigated and reported on comprehensively, problems with bridges (and other structures) are often either unreported or very poorly reported. This may be because details of a failure are not fully shared with the press, or are withheld due to ongoing legal action.

The concern that appears to have prompted this seminar is that if details of failures aren't shared with other bridge owners and engineers, then the opportunity to learn lessons and prevent similar failures from happening may be missed.

In Britain, failures in related industries are in some cases very well investigated and publicised. Notable, the Rail Accident Investigation Branch publishes thorough reports on rail related safety failures. There is a legal duty to notify the RAIB of an accident and to assist their investigations (Railways (Accident Investigation and Reporting) Regulations 2005). There are equivalent bodies for the air and marine industries. However there is no equivalent body which would investigate highway or footway bridge failures - the closest is the Health and Safety Executive. A key difference is that if the HSE investigates a safety failure, it is generally with the aim of prosecuting those responsible. An investigation by the RAIB is intended to share knowledge of the causes of failure, not to apportion liability (although, by making recommendations on what the people involved should do differently, it can't avoid attributing blame).

The problem with all the above bodies is they will only investigate after a failure. If failure led to death, a coroner's inquiry might be sufficient to determine the reason, but would not provide exhaustive technical detail. If the failure is significant enough, a body can be set up to investigate even where none already existed, as happened following several failures of steel box girder bridges with the appointment of the Merrison Committee in December 1970.

There's little doubt that bridge failures have been one of the main sources of technical improvement in bridge design (in addition to Merrison, the cases of the Millennium Bridge, Tacoma Narrows and others are relevant - see Henry Petroski's book "To Engineer is Human" [Amazon.co.uk]). But it's hard not to suspect that there is a great deal of very useful information gathered by forensic engineers and expert witnesses in legal cases and rarely if ever more widely disseminated. While there is a confidential reporting route (CROSS) intended to encourage information to enter the public domain, it seems mainly to address generic problems rather than specific, detailed cases.

I doubt much can be done to remove the obvious impediments to sharing information more widely: client confidentiality; commercial and competitive advantage of limiting the spread of knowledge; fear of costly legal action etc. If anything, the last of these impediments has been made significantly worse by the introduction of corporate manslaughter legislation. One option might be to introduce specific legal limits on liability for negligence (but it seems unlikely that the benefit of encouraging knowledge-sharing will outweight the adverse effects on victims). Another would be to consider mandatory investigation and reporting regimes similar to those that operate in other industries (but this is expensive).

It will be interesting to see whether the ICE seminar can come up with anything which has a more realistic chance of success!