Sensitivity and Uncertainty Analyses Applied to One-Dimensional Radionuclide Transport in a Layered Fractured Rock. Evaluation of the Limit State Approach. Iterative Performance Assessment. Phase 2

摘要

In assessing high-level waste (HLW) repository performance, mathematical models will be used to estimate system behavior in response to future perturbations, including the release of radionuclides into the geosphere. Performance assessment models are based on a combination of scientific principles and empirical data. It can be expected that, because of inherent variabilities in the model inputs due to the difficulties in obtaining relevant field data and in validating the models, uncertainties will exist in the model input parameters as well as in the description of processes included in the models. Therefore, the predicted response or performance will also be uncertain. In the context of this report, uncertainty analysis means probabilistic analysis in which the uncertain input parameters are treated as random variables with assigned probability distributions. The goal of the uncertainty analysis is to compute the cumulative probability (or cumulative distribution function, (CDF)) of one or more responses or performance measures. Sensitivity analysis, on the other hand, refers to the estimation of change in the performance measure if any one of the input variables is varied. Probabilistic sensitivities are defined as variations in the performance CDF because of a change in an input variable.