Why is it appropriate to use the criterion of a “one-in-a-million chance” per year to select accidents for analysis?
Using such a criterion allows the NRC to concentrate its resources and detailed analyses on those events that, while only remotely probable, are more likely to realistically contribute to public risk. Realistic and risk-informed regulatory decisionmaking focuses on the value of preventive and mitigative features for the more likely, albeit remotely probable, scenarios. To be analyzed in SOARCA, an accident scenario had to have a probability of occurring more than once in a million reactor years (or more than once in ten million reactor years for accidents that may bypass containment features). In addition, the staff considered scenarios that may have lower probabilities than the selection criteria, but potentially higher consequences. Scientific knowledge, combined with theoretical projections, allows probabilities to be assigned to extremely unlikely events. However, the estimated probabilities of such events are highly suspect because there is no human experience from which to judge
