Abnormal Normal Accidents
The huge earthquake that hit northern Japan and the tsunami that followed it twenty minutes later created a terrible tragedy in that country. The loss of life added to the loss of property and communities creates a situation hard to imagine for those of us watching from a distance, particularly with additional concerns about a nuclear disaster and the spread of radiation.
Perrow coined the term to signify the inevitable situation that occurs when a tightly coupled and interactively complex system breaks down. “Tightly coupled” refers to a system in which the parts are highly interdependent, they have major impacts on each other when change occurs and those impacts happen quickly.
“Interactive complexity” refers to complicated systems in which processes lead to yet other processes, many of which may be unidentified and unplanned. The example he uses is primarily of the “high-risk technology” of nuclear plants. The many cases of normal accidents he recounts typically involve not only technological breakdowns but also operator error that exacerbates the problems.
Many organizations that use high-risk technologies have paid attention to Perrow’s work and the concept of “normal accident” so that they may perhaps design in some added safeguards against them. NASA is one such organization.
While it remains to be seen how the nuclear plant situation in Japan will resolve, let’s see if what has happened in the first few days can be explained by Perrow’s theory.
The problem began when the earthquake shut down the active reactors at the plant. (Three of the six had been shut down for maintenance.) Flooding from the tsunami wiped out the generators that kept the cooling pumps running. This situation resulted in the core melting down in three reactors with explosions and fires damaging the housing and other internal equipment. Continuing explosions within the chambers prompted evacuations of workers and there have been some evidence of radiation leaks in the surrounding areas.
After nine days, power was restored and repairs began. Radiation contaminated food grown in farms surrounding the area and traces of radioactive iodine were found in Tokyo’s water supply.
Does this fit the model of a “normal accident” even though the earthquake and tsunami were certainly not “normal” events?
Nuclear power plants are tightly coupled systems in which each process is linked to another as seen when the cooling pumps failed, a chain reaction of events resulted in fires, explosions, and radiation leaks. So we have interactive complexity. Although workers attempted to fix the systems, the chain reactions of failures happened quickly and at times unexpectedly from what we’ve seen reported.
We’ll need more information to see if operator error was involved, although part of the theory of “normal accidents” suggests that no human operator could possibly keep up with the accident chain of events since they happen so quickly and so unexpectedly. At this level, operator error isn’t an opportunity to blame the workers for the problem or lack of solution; it’s simply part of the complex puzzle that makes such a high tech system.
Perrow’s theory highlights the normality of accidents when we have such high-risk technologies. That accidents will happen is a given. What we do in response to knowing this will happen becomes important.
Interestingly, a number of countries have decided to close or scale down their own nuclear power plants as a response to this event. This is interesting considering they didn’t make such a decision after the accident at Three Mile Island in Pennsylvania in 1979 or even after the Chernobyl disaster in 1986.
The previous two events had operator errors involved that contributed to the events, while this one was prompted solely by a natural event and we’ve yet to hear about any operator errors. We may realize that we have trouble designing out natural events such as earthquakes. However, we continue to think we can design out operator error even though Perrow’s theory suggests otherwise.