The bridge failure stats from the previous post were updated through 2006, but I originally started looking at the stats when I was in grad school in 2000, and the data itself was no more recent than 1991. (True story, back then 42.3% of all failures were dinosaur related.) When I got my hands on the 2006 update I thought it might be interesting to look for changes over time.
I was looking for a change in the distribution of the failure causes. As the statistics from the last post point out, the majority of bridge failures are caused by hydraulic issues, but hydraulic factors have always been one of the least understood parts of the bridge design process. Through the 1980’s it was mostly an approximate process. That all changed due to two major bridge failures, one in New York state and one in Tennessee. Lives were lost during both failures, and both drew a lot of attention nationally from the public and from engineers. As a result, a lot of research has been done on bridge hydraulics and scour and state DOTs developed design standards in these areas. So I was curious to see if that shifted the failure statistics away from hydraulics.
Nope. Sure didn’t. The statistics from the 2006 dataset were surprisingly consistent with the 1991 data. Hydraulic failures went from 60% to 58% with the 2% drop being offset with a 2% rise in overload failures. Since the 2006 data was merely an updated version of the 1991 data I decided to separate it into pre 1991 and post 1991 data sets.
So as one final check of the data, I separated the New York state data from the rest of the country. The assumption being the New York data would be more comprehensive since New York State DOT was collecting the data. The resulting analysis did show a significant shift. Hydraulic failures dropped to 48% while collision failures went up to 26%. I think this is because collision failures happen more often on smaller bridges and aren’t as widely reported in the national media.
I don’t think it’s a good idea to make any grand conclusions due to the nature of the data collection, but I suspect the analysis of the New York state failures shows the distribution between hydraulic and collission failure causes more accurately. The most surprising part of the analysis was how consistent all the numbers were. There was very little change from 1991 to 2006, and even breaking out just one state showed almost no change in any of the failure causes other than hydraulic and collision. (And earthquake failures were not represented in the New York data.