WAVELET-MONTE CARLO HYBRID SYSTEM FOR HLW NUCLIDEMIGRATION MODELING AND SENSITIVITY AND UNCERTAINTY ANALYSIS

摘要

This paper presents results of an uncertainty and sensitivity analysis for performance of the different barriers of high-levelradioactive waste repositories. SUA is a tool to perform the uncertainty and sensitivity on the output of WaveletIntegrated Repository System model (WIRS), which is developed to solve a system of nonlinear partial differentialequations arising from the model formulation of radionuclide transport through repository.SUA performs sensitivity analysis (SA) and uncertainty analysis (UA) on a sample output from Monte Carlosimulation. The sample is generated by WIRS and contains the values of the output values of the maximum releaserate in the form of time series and values of the input variables for a set of different simulations (runs), which arerealized by varying the model input parameters. The Monte Carlo sample is generated with SUA as a pure randomsample or using Latin Hypercube sampling technique. Tchebycheff and Kolmogrov confidence bounds are computedon the maximum release rate for UA and effective non-parametric statistics to rank the influence of the model inputparameters SA.Based on the results, we point out parameters that have primary influences on the performance of the engineeredbarrier system of a repository. The parameters found to be key contributor to the release rate are selenium andCesium distribution coefficients in both geosphere and major water conducting fault (MWCF), the diffusion depth andwater flow rate in the excavation-disturbed zone (EDZ).