I’ve been excited and nervous about the debate on nuclear power safety both at the same time. It’s made me think about the nature of infrastructure failures.
In short, our lifeline systems are becoming so safe that all failures will be catastrophic.
Today on one of the morning radio shows I listen to, an interviewee was explaining to her interviewer the difference in the nature of design and construction of infrastructure systems at the beginning of the 20th century (build-test) and the beginning of the 21st century (model-prototype-scale-model-prototype-test-build).
While my description may be a bit exaggerated, our advances have indeed been that substantial, especially when it comes to the level of safety required. I loved the interviewee’s description of the risk management portion of the design, operation, and management of infrastructure: “You tell the engineer you want it to be safe… The engineer asks you ‘How safe?’ then figures out how to get there with your budget and other constraints…”
There are several thoughts that come to mind when thinking about this statement. Let me share just a couple. First, safety is a function of not only the engineers’ ability to imagine the most outlandishly improbable catastrophic events, but also of the clients’ desire to protect themselves and their customers against such outlandishly improbable events. Second, due to the lengthy and sophisticated design process, including several stages of modeling, prototyping, and testing, small-scale failures are almost unheard of (if you will omit cases of water main breaks or power outages attributable to tree branches falling on lines). We only hear about the spectacularly tragic events.
The spectacularly tragic events are frightening because risk management suffers from remarkably steep diminishing returns. Protecting assets and local populations against the most 1% extreme events may cost more than the other 99% in some cases. And this is when the engineers have been sufficiently creative to dream up an ambitious enough set during the design stages. Perhaps we should be training engineers in “predictive precautionary pessimism.”
The events unfolding after the Japanese tsunami and earthquakes last month make me nervous because they show us clearly that the overwhelming majority of infrastructure failures from here on out will be spectacular failures. These same events make me excited because they show that we as engineers have been remarkably resourceful in designing for the most formidable circumstances.