How do we know that bridges will stay up?

20 February 2018


How confident can we be that a bridge will stay up?


Chris Smith put this question to engineer Allan McRobie...

Allan - You can be fairly confident about it. It’s been designed by a structural engineer and then checked independently by quite a few others. When you would not be quite so certain is if it’s a fairly innovative bridge that is a type we haven’t seen before. There’s a thing in reliability theory called the bathtub curve I think they call it, which says if something's going to fail it sort of fails either early on on day one or two, or it’s going to be a long time, many years.

The early failures are there’s some dodgy detail or there’s a bad bit of design and you don’t really know till you’ve built it and then load tested it, so don’t go on bridges on their opening day is a general rule. If it makes the opening day, it’s going to be fine on day two, day three, etc. But then after several decades you start to get corrosion, you get fatigue, cracks will open up, etc., you’ll get material degradation. So then it becomes more likely to fail but we do have good bridge inspection regimes.

For really long bridges the hardest thing is keeping it there in a hurricane or a storm. It’s all about the wind design. It’s quite easy to hold a traffic jam in the air for several kilometres, you just need really big cables, but to hold it there during a hurricane is a really difficult piece of engineering, the way the structure and the wind talk to each other. And so don’t go on big bridges in hurricanes. You would be blown off .

Chris - How does an engineer, like you, anticipate and model that sort of scenario? Do you have a computer programme that pretends its got the bridge sitting there and can work out all of the forces a bridge would be feeling in said hurricane, for example? Is that how you do this?

Allan - You do both. We do it on the computer and in the wind tunnel. You make scale models of it in the wind tunnel but we know have fairly sophisticated computational wind tunnels and the real difficulty is the wind. If you’re looking at the bridge, you’re looking at the wrong thing. We really need to understand the wind and all the swirls and vortices and the forces that it creates out of nowhere. It makes the bridge move a bit. There’s a feedback system. It’s really very complicated and, in fact, because I know how complicated it is I’m not all that certain that we should be quite so confident that we know what we’re doing. We are very confident but we want one in a million chance that this is going to go wrong. Whereas I sometimes struggle to put the figures quite so high because it is really difficult physics.

Chris - Because things do end up needing changes don’t they? If you look at the Tate Modern bridge in London, it was a bit of an oversight wasn’t it on the part of the engineers that people were actually going to walk across this bridge? And it’s once people began walking across it that it set up resonances on the bridge and the whole thing was swaying from side to side and they had to stabilise it. Did no-one actually think someone’s going to use this bridge when they were testing it and building the model?

Allan - No they did, and it’s quite an interesting one. The fundamental physics there is actually biology. It’s about the people and not about the bridge, it’s the way that people change the way they walk. There was a feedback and the people behaved like a fluid and as the bridge moved a bit they changed the way they walk. But that really demonstrated the day one phenomenon that the most telling thing was when they first put the crowds on it. That was a real load test. I told my family don’t go on it that day. It’s a load test, it’s an experiment.


Add a comment