Bridge Inspection and Testing

Ed Morrissey had two interesting posts over the weekend on bridges, first on inspection techniques and next on nondestructive testing. It appears that inspecting bridges is as much an art as a science. You can’t tear down a bridge to test it; you have to make inferences from observations that are in some cases not at all metric. A bridge inspector has to search for signs that would indicate, under current engineering knowledge and theory, that a bridge is in danger of collapse.

It seems to me that there are two questions that the investigation of the I-35W bridge must address. The first is, Did the inspectors fail to notice things that would indicate, under current engineering knowledge and theory, that the bridge was in danger of collapse? It seems to me likely, though not certain, that the answer will turn out to be no. Or that it will be impossible to determine that the answer is yes. I just think it’s very unlikely that Minnesota bridge inspectors will turn out to have been incompetent.

The second question is raised if the answer is no or cannot be determined: What caused the bridge to collapse that current engineering knowledge and theory did not predict? Remember that the designers of the bridge back in 1967 did not understand the potential of metal fatigue. That was discovered by engineers not long after 1967, and presumably bridge inspectors have looked for signs of dangerous metal fatigue ever since. Any answers to this question will enable bridge inspectors to do their job better in the future.

It may be determined that bridges built circa 1967 have a common characteristic that puts them in danger of collapse 40 years later. If so, that suggests that a massive program of bridge replacement may be necessary and that much more intensive bridge inspection is needed immediately. I am prompted here by my rereading of Henry Petroski’s 1995 book Engineers of Dreams. Here is an interesting passage near the end:

Must we thus expect, if not allow, a bridge failure to occur now and then? The history and promise of bridges suggest we must, for reasons that have to do with neglect of the past and its relevance for the future. Neglect of the past is often embodied in a short-term historical memory, thinking, with hubris, that one’s own generation’s engineering science and technology have progressed so far beyond what they were a generation or two earlier that the bridges of one’s professional progenitors, and even one’s mentors, make pretty pictures but not examples or models for modern engineering. A historical perspective on bridges and their engineers reveals not only that such shortsightedness is nothing new but also that it has led to disaster time and again.

A close reading of the history of major bridge failures is contained in a remarkable piece of scholarship by Paul Sibly and his adviser, then at University College London, Alastair C. Walker. Among the conclusions of their work, published in 1977, was the strong temporal pattern that bridge failure had followed from the middle of the nineteenth century. What Sibly and Walker noted was that the collapse of the Tay, Quebec, and Tacoma Narrows bridges, which occurred in 1879, 1907, and 1940, respectively, were very nearly thirty years apart. A less commonly remembered incident, but one that was equally dramatic and in its own time the subject of investigation by a royal commission, was the collapse of Robert Stephenson’s Dee Bridge in 1847—further reinforcing the observation that a thirty-year cycle was associated with bridge failures. To test their hypothesis, which pointed to a major bridge failure about the year 1970, Sibly and Walker looked at incidents from around that time and found that, indeed, in 1970 there were two significant failures of a new type of steel bridge, known as a box girder, then under construction in Milford Haven, Wales, and in Melbourne, Australia.

The Tay, Quebec, and Tacoma Narrows bridges as described earlier in the book by Petroski were innovative structures, at the cutting edge of engineering, that collapsed while under construction or shortly after coming into use. The I-35W bridge collapsed 40 years after construction. But the date of its construction—1967—is very near Petroski’s 30-year interval. This suggests to me that if it is discovered that there was something in the design and engineering of the I-35W bridge that made it liable to collapse 40 years later for reasons not understood by today’s engineering knowledge and theory, that factor could also be present in many other bridges of the same vintage. Of which, I suspect, there are very many around the country and the world. It will be interesting to see the results of the investigation of this tragic event.