Complex Systems Are Very Likely to Experience Accidents :: Challenger Accidents Accident Papers

Complex Systems Are Very Likely to Experience Accidents Many people in today’s industrial countries have experienced the frustration and inconvenience of having their car break down. That event, while troublesome, often does not pose any significant danger to people. It is a useful microcosm, however, because cars, like other complex systems, will almost certainly malfunction at some point during use. While we cannot prove the following assertion for sure, empirical data and observations strongly suggest that complex systems that are made of unreliable components will inevitably experience accidents so long as there exist flaws in the system that have no reliable safeguards. Numerous studies have investigated such a ‘Normal Accident Theory,’ and two notable cases very strongly indicate its validity: petroleum refinery accidents and the space shuttle Challenger, both of which will be discussed in this paper. Normal Accident Theory (NAT) is the label for a school of thought that considers accidents in complex systems to be inevitable. Two characteristics of complex systems that are very important to NAT are the interactive complexity of a system and a system’s coupling. Coupling is determined primarily by the time between processes in a system, the independent or dependent progression of such processes, and the number of different ways that a system’s goal can be achieved (Piccard, 1999). Systems are classified as ‘tightly coupled,’ meaning that the time between processes is small, the processes are highly interdependent, and there are few paths, if not one, to the goal; or ‘loosely coupled,’ which is the opposite. These characteristics are particularly useful for comparing different complex systems and evaluating them to determine which are at the highest risk for accidents. The results can then be used to minimize, but not elimina te, the possibility that an accident will occur. Sociologist Charles Perrow is generally credited with developing NAT. In order to understand the principles of NAT, several definitions that it uses are essential. An accident is defined as â€Å"an event that is unintended, unfortunate, damages people or objects, affects the functioning of the system of interest, and is non-trivial.† (Perrow, 1994) There are two types of accidents: component failure accidents, which â€Å"involve one or more component failures (part, unit, or subsystem) that are linked in an anticipated sequence,† and ‘normal accidents,’ or system accidents as they can be called (Perrow, 1994).