Viewpoint: Failure follows success - failure in design engineering

Materials World magazine
2 Apr 2012
stylised graphic of Brooklyn suspension bridge

Professor Henry Petroski of Duke University, USA, discusses the significance of failure in design engineering.

We know success when we see it, and nothing is quite like it. Successful products, people, and business models are the stuff of best-sellers and motivational speeches, but how often can their stories be replicated?

Success is a dangerous guide to follow too closely. Imagine what would have happened if Titanic had not struck an iceberg and sunk on her maiden voyage. Her reputation as an ‘unsinkable’ ship would certainly have been reinforced. Imagine further that she had returned to England and continued to the North Atlantic without incident: her success would have been evident to everyone, and competing steamship companies would have wanted to model their new ships on her.

Indeed, they would have wanted to build even larger ships — and they would have wanted to build them more cheaply and sleekly. There would have been a natural trend towards lighter and lighter hulls, and fewer and fewer lifeboats. Of course, the latent weakness of the Titanic’s design would have remained in her and her imitators. It would have been only a matter of time before the position of one of those derivative ships coincided with that of an iceberg and the theretofore unimaginable occurred.

Incontrovertible failures reveal weaknesses. A success is just that — a success. It is something that works well for a variety of reasons, not least of which may be luck. But a true success often works precisely because its designers thought first about failure. Indeed, one simple definition of success might be the obviation of failure.

Engineers are often called upon to design and build something that has never been tried before. Because of its novelty, the thing cannot simply be modelled after a successful example, for there is none. This was certainly the case in the mid-nineteenth Century, when the railroads were still relatively new and there were no existing suspension bridges capable of carrying them over wide waterways and deep gorges.

The suspension bridge seemed to be the logical choice for the railroads, but existing suspension-bridge roadways designed to carry carriages and pedestrians were light and flexible and many had been blown down in the wind. British engineers took this absence of a successful model as the reason to come up with radically new bridge designs that like the famed Britannia Bridge, were often prohibitively expensive to build and technologically obsolete almost before completion.

The German-born American engineer John Roebling looked at the history of suspension-bridge failures in a different way. He studied them and distilled from them principles for a successful design. Taking as his starting point the observation that wind was the greatest enemy of such bridges, he devised ways to keep the bridge decks from being moved to failure by the wind.

Ironically, his masterpiece, the Brooklyn Bridge, served not as a model of how to learn from failure but as one to be emulated as a success. Subsequent suspension bridges designed by others over the next half century successively did away with the stay cables, trusswork, and deck weight that Roebling had so deliberately used to fend off failure. The culmination of this paring down of a successful design was reached in the late 1930s, when bridges were built longer, lighter and more slender.

At first there were just warnings that something was amiss. The Golden Gate Bridge, shortly after it opened in 1937, showed a surprising flexibility in strong winds. At the same time, bridge designers were striving for an aesthetic of shallow decks with no stiffening trusswork. Suspension bridges built in remote locations with little traffic were designed with not only shallow but also narrow decks, and these were naturally very flexible.

The deck of the Bronx-Whitestone Bridge, completed in 1939 just in time to carry traffic to New York World’s Fair, undulated in the wind, as did that of the contemporaneous Deer Isle Bridge in Maine. Engineers disagreed on the cause and remedy of the unexpected motions, and also on exactly how to retrofit the bridges with cables to check them. No one appears to have feared that the bridges were in imminent danger of collapse.

The Tacoma Narrows Bridge, completed in 1940, behaved in much the same way, with its deck rising and falling in great undulations. The fun of it all actually increased traffic beyond expectations so that it came to be nicknamed the Galloping Gertie. The fun lasted for only four months, however, when the bridge deck began to move in a new way. It started to twist with great amplitude, and after only hours of such motion its deck collapsed into the arm of Puget Sound that it had been designed to cross.

Stories like those of the Titanic and of the Tacoma Narrows Bridge provide classic case histories of the danger of having undue confidence in models of success — and following them to failure.

Further information

Henry Petroski, the Aleksandar S Vesic Professor of Civil Engineering and a Professor of history at Duke University, North Carolina, is the author of a number of books on engineering and design, including An Engineer’s Alphabet: Gleanings from the Softer Side of a Profession and the forthcoming To Forgive Design: Understanding Failure.