Approaches to the regulation of risk from technological systems, such as nuclear power plants or chemical process plants, in which potential accidents may result in a broad range of adverse consequences must take into account several different aspects of risk. These include overall or average risk, accidents posing high relative risks, the rate at which accident probability decreases with increasing accident consequences, and the impact of high frequency, low consequence accidents. A hypothetical complementary cumulative distribution function (CCDF), with appropriately chosen parametric form, meets all these requirements. The Farmer limit line, by contrast, places limits on the risks due to individual accident sequences, and cannot adequately account for overall risk. This reduces its usefulness as a regulatory tool. In practice, the CCDF is used in the Canadian nuclear licensing process, while the Farmer limit line approach, supplemented by separate qualitative limits on overall risk, is employed in the United Kingdom.
Read full abstract