The Institution of Civil Engineers – the UK counterpart to the American Society of Civil Engineers – issued a report on the whole-life risks for infrastructure (in its broad sense of the engineered built environment) some time ago that I finally found the time to read. And now I want to encourage every engineer to read it if they haven’t already. In short, the problems revealed by failures (the report discusses the Grenfell tower fire and the Morandi Bridge collapse) are ones that we see time and again: designs that barely meet design criteria and there fail to meet them as they age, designs driven by first-cost accounting without regard for maintenance, lack of maintenance, and designs that repeat old mistakes. The picture above is from a UK forensic report on a 1908 loft building fire in New York, and the problem of floor-to-floor flame spread is one that still kills people today.
One of the summary points in the introduction (on page 6) jumped out at me: “Furthermore, economic pressures, prioritisation of capital cost savings over whole life value, and narrowly-designed contract incentives can create unintended outcomes that increase risks.” This seems familiar because I was talking about it eight years ago. At that time I said “But what does efficiency actually mean? Any design is a response to multiple criteria including the physical properties of materials, code requirements, costs, clients’ stated desires, aesthetics, and conservation philosophy. When we pursue efficiency, we optimize our design to favor one or two – or on rare occasions three – of the optional criteria….We can provide maximum flexibility of use at the expense of greater intervention and cost.” Or, to look at it the other way around, if we pursue lowest first cost, we do so at the expense of other criteria, most often structural robustness and architectural performance.
The vast majority of failures have from more than one cause. The combination of poor design and materials degradation over time is an example of two causes that interact with each other. The report promotes the “Swiss cheese” model of risk, which is a useful mental model, despite the name. If you picture a stack of slices of Swiss cheese, with each slice representing a defense against failure and each hole in a slice representing a way in which that particular defense might not work, failure occurs when one or more holes line up all the way through the stack.
We can improve safety by improving design quality, by improving maintenance regularity, by adding resilience to basic design criteria, and so on. All are good. However, each of these ideas shrinks the size of the holes in only one slice of cheese: the holes are still present. So the more different ideas we implement, the better the odds that at least one line of defense will work in a given situation.
We – the AEC community – can’t guaranty safety. But we can improve the odds.
