System Fractures: A Summary

The causes of disasters fall into a small number of repeating patterns. These lessons apply not only to the operation of complex, high-energy systems, but even to daily lives and workplaces. As an analogy, consider how a piece of metal breaks over time. Under stress, cracks begin to grow out of tiny manufacturing flaws and damage during use; then at a critical point a crack spreads like a gunshot and the piece fails completely.

As with metal, weak points appear in all systems: these weak points are human errors and machine malfunctions. A good system is one in which people catch weak points early, before a string of them link up to a system fracture. Good systems have much redundancy, which usually takes the form of many alert people who are alert and empowered.

This is relevant because technological disasters hardly ever come like bolts from the blue ... from a single unexpected event. Nearly all have been preceded by early warning signs, called precursors. These occurred days or weeks ahead. These were indications that serious flaws existed and were starting to link up. The good news is that these precursors, if noticed and acted on, give people a chance to act before the day of disaster.

Companies and agencies that deal in high-energy, complex machines need to be reminded on a regular basis that techno-disasters have a high and long-lasting cost. Beside the obvious costs of deaths, damage, hikes in insurance costs, and months of business interruption, in some cases so much public mistrust follows that an entire segment of industry may be wiped out, once the public comes to see it as both risky and optional. At this point a disaster becomes a business catastrophe, literally, a turning point. Crashes of two Comet airliners in 1954 halted British airliner manufacturing for so long that the American industry took over the business of manufacturing jet airliners. And the scale can be enormous: A series of dam failures in China on the night of August 7, 1975, killed over 26,000 people. Incident costs ran well over four billion dollars at the Three Mile Island Unit 2 partial meltdown. When the final tally is in, response costs for the failed flood-control system in New Orleans and the Deepwater Horizon blowout will be higher still.