The Interpretation of Risk and Sensitivity Under the Peak-of-the-Mean Concept

摘要

Although the calculation of the peak risk for the nominal case is straightforward, calculations for the disruptive case are more complicated when the event probability is small, event times are random, and effects of the releases are of relatively short duration. Analysts must guard against risk dilution caused by inappropriately increasing the bounds of the distribution of poorly known parameters, especially if these parameters affect the timing of peak dose within a realization. Finally, it is possible to conduct sensitivity analyses using doses calculated at the time corresponding to the maximum risk or using the peak dose from each Monte Carlo realization. Quantitative comparisons suggest that using the peak dose from each realization can provide greater discriminatory power. However, some parameters, especially those that determine timing of the releases, should be interpreted carefully for the impact of their associated uncertainties on system performance. NOTE: The NRC staff views expressed herein are preliminary and do not constitute a final judgment or determination of the acceptability of a license application for a geologic repository at Yucca Mountain.