159 Techniques for Sensitivity Analyses on Non-Monotonic Functions (PSAM-0378)

The ability of sensitivity techniques developed at the Center for Nuclear Waste Regulatory Analyses (CNWRA) to highlight non-monotonic trends is studied in this paper. The capability of the parameter tree, the mean-based and standard deviation-based sensitivity, and the partitioning methods to identify sensitivities of non-monotonic functions is explored. Sensitivity indices and significance tests for the various techniques are discussed in the paper. It is concluded that the standard deviation-based index can consistently identify non-monotonic trends. The parameter tree and the partitioning methods provide some capability to detect non-monotonic trends; however, they lack resolution if the performance metric reaches a maximum near the median value of the input parameter distribution. At the core of the standard deviation-based index is a mapping of an input parameter distribution into square values of a standard normal distribution (zero mean and unit standard deviation). This mapping de-emphasizes input parameter extremes, and highlights middle values of an input parameter distribution. Simple mathematical functions, as well as a complex environmental stochastic model, are used to present the techniques. Acknowledgments: This paper was prepared to document work performed by the CNWRA for the Nuclear Regulatory Commission (NRC) under Contract No. NRC-02-02-012. The activities reported here were performed on behalf of the NRC Office of Nuclear Material Safety and Safeguards, Division of High Level Waste Repository Safety. This paper is an independent product of the CNWRA and does not necessarily reflect the view or regulatory position of the NRC.