Of Disasters and Dragon Kings: A Statistical Analysis of Nuclear Power Incidents & Accidents
|Author||S.Wheatley, B.Sovacool, D.Sornette|
|Classification||6.01.3.70/80 (NUCLEAR SAFETY - RISK ANALYSES / RISK PERCEPTION)|
From the publication:
Of Disasters and Dragon Kings: A Statistical Analysis of Nuclear Power Incidents & Accidents Spencer Wheatley1, Benjamin Sovacool2 and Didier Sornette1 1 ETH Zurich, Department of Management, Technology and Economics, Switzerland 2 Center for Energy Technologies and Department of Business and Technology, Aarhus University, Denmark e-mails: email@example.com, BenjaminSo@hih.au.dk and firstname.lastname@example.org April 10, 2015 Abstract We provide, and perform a risk theoretic statistical analysis of, a dataset that is 75 percent larger than the previous best dataset on nuclear incidents and accidents, comparing the three measures of severity: INES (International Nuclear Event Scale), NAMS (Nuclear Accident Magnitude Scale) and dollar losses. The rate of nuclear accidents with damage above 20 MM 2013 USD (normalized by the number of reactors in operation) has decreased from the 1970s until the present time. Along the way, the rate dropped significantly after Chernobyl (April, 1986) and is expected to be roughly stable around a current level (in 2015) of 0.002 to 0.003 events per plant per year. The distribution of damage sizes appears to have undergone a regime change shortly after the Three Mile Island major accident (March, 1979). The median damage size became approximately 3.5 times smaller, but the tail became much heavier, such that it is well described by a Pareto distribution with parameter α 0.55. In fact, the damage of the largest event (Fukushima, 11 March, 2011) is equal to near 60 percent of the total damage of all 174 accidents in our database since 1946. We also document a statistically significant runaway disaster regime in NAMS (radiation release) data as well as a related runaway disaster regime in damage sizes, which we associate with the “dragonking” phenomenon. With the current model and in terms of dollar losses, there is a 50% chance that (i) a Fukushima event (or larger) occurs in the next 50 years, (ii) a Chernobyl event (or larger) occurs in the next 27 years and (iii) a TMI event (or larger) occurs in the next 10 years. Further, smaller but still expensive ( 20 MM 2013 USD) incidents will occur with a frequency of about one per year. Finally, we find that the INES scale is inconsistent in terms of both damage and NAMS (radiation release) values. For the damage values to be consistent, the Fukushima disaster would need to be between an INES level of 10 and 11, rather than the maximum level of 7.