Whenever someone comes up with a new type of bridge, there is no way of knowing if the bridge will work or not. The stakes are pretty high too – no one wants to build a bridge that ends up like the The Tacoma Narrows Bridge:

Leon Moiseff, the man who designed this bridge, was no hack. Previously, he had also designed the Golden Gate Bridge, among others. Here is how he described the Tacoma Narrows Bridge:

“The most beautiful in the world.” That’s how engineer-designer Leon Moisseiff described his 1940 Tacoma Narrows Bridge. The statement represents much more than one engineer’s opinion about his own work. Moisseiff’s words reflect important architectural design trends and artistic tastes of the 1930s, as well as three decades of suspension bridge design.

Moisseiff cared deeply about bridge aesthetics. Bridge designs, he said, needed to be “safe, convenient, economical in cost and maintenance and at the same time satisfy the sense of beauty of the average man of our time.” Moisseiff believed that engineers should try “to develop the beauty of their structures” by emphasizing “the essential, to interrupt rhythmically the monotonous and to indicate the minor importance of the auxiliary . . . and attain the pleasure of good form.” Bridge designers, he said, should “search for the graceful and elegant.”

Why do we need new bridge designs? The main reason is not aesthetic – the primary driver is that new methods of building bridges allow us to cross spans that previous designs couldn’t cross. That is why we see new forms of bridges being designed and built – each advance allows us to get across longer stretches, or to get across gaps that have unique physical problems.

New bridge designs let us do new things – but the risk is that we don’t know that they’ll work until we build them. This is true of many innovations – it is often very hard to know in advance if they will be successful once they are at full scale. How can we deal with this? We can learn a bit from bridge designers.

The first thing that they do is they figure out small-scale experiments that allow them to test possible new designs. Here is how Henry Petroski describes preparation for building the Britannia Tubular Bridge (shown above) in the 1840s, in his chapter in Seeing Further: The Story of Science and the Royal Society edited by Bill Bryson:

The experimentalist-engineer William Fairburn… was responsible for conducting scale-model strength tests to establish the preferred shape and detailed design of the wrought-iron tubes. He began with small-scale models to compare the relative strengths of different shapes and arrived at the conclusion that the rectangular cross-section was the best. The model tubes were tested by handing from their centre weights that represented the load of a heavy locomotive. Weights were added until the tube failed, which revealed the weakness of the structure and thereby provided guidance for how to modify it in the next model. By progressively increasing the scale of his models, Fairburn was able to establish trends of behaviour, and from the experimental data the theorist Eaton Hodgkinson established an empirical formula by means of which he could extrapolate to the requirements for the full-sized tube.

To build a full-scale model and test it to destruction would have been essentially to build the bridge itself. So, as is typical in the engineering of large structures to this day, there comes a point when judgment dictats that the model testing must end and the real thing begin.

Small-scale testing is one method for experimenting to support bridge design innovation. However, there are often things that we can’t anticipate, and consequently can’t test. The problems with the Tacoma Narrows Bridge were caused by the huge amount of wind that whips through the Narrows. This is what generated the huge torsion which tore the bridge apart. The design had been thoroughly tested for strength, but not for wind.

Of course, now we can use things like computer simulation to test even for factors such as wind – simulation is part of a class of tools that can help us with this kind of innovation. This allows for even more extensive experimentation than can be undertaken through physical methods. Still, we can only model things that we know about. The Millennium Bridge in London was simulated extensively before it was built, and yet it still had problems with excessive swaying.

It turns out that this swaying was caused by lateral movement induced by synchronised walking. Once this was discovered, possible changes were simulated, until a method for passively damping the movement was designed and installed. Subsequently, the bridge has been stable.

There are three key innovation lessons in all of this:

1. Test out your big ideas by devising a small-scale experiment first. It is often difficult to figure out how our innovations will work at full scale. But a good first step is to figure out a way to experiment with trying them on a smaller scale first. This is what Hindustan Unilever did when introducing their Shakti Program in India – this is a radical new sales-distribution model. They tested it first with just seventeen women to see if it would work, then they scaled up over time. We can learn a great deal from experiments.
2. We don’t always have to do live experiments – there are now many innovation technologies available that will enable us to conduct experiments through methods such as rapid prototyping and simulation. Arup Engineering used simulation extensively in the initial design of the Millennium Bridge, and just as extensively in the retrofitting of the motion damping system. With these technologies, we are able to experiment even more extensively than previously.
3. Finally, we can’t test everything before we launch – at some point we have to use our judgment about whether or not things will work. Sometimes, there will be problems from sources that we never could have anticipated. This may cause failure – as in the case of the Tacoma Narrows Bridge. The key lesson here though it that when we discover a new way to fail, we must ensure that we learn as much as possible from it. After the Tacoma Narrows Bridge collapsed, subsequent bridges using that type of design included stabilisers that prevented the twisting motion. Consequently, no other bridges have collapsed from that particular problem.

Experimenting is a key tool in innovation. We can use it to discover problems and test solutions before we have a massive failure in public. Experimentation can reduce the risk of innovation. So think like an engineer, and figure out ways to test out your ideas before you fully implement them.