Risk from Supernovae compared with Risk Standards for Nuclear Reactors

Abstract

FARMER¹ has argued that any nuclear power reactor is in principle capable of offering a range of risk to life and property and that the plant safety standards should be controlled so that the larger consequences have an associated lower risk. Discussing appropriate standards, he observes that the permissible frequency for the largest accidents seems to be in the range 10⁻⁶ to 10⁻⁸ (reactor years)⁻¹, depending on the system of extrapolation for risk versus accident size.