High level radioactive waste repository total system performance uncertainty and sensitivity analysis

A probabilistic performance assessment (PA) is one of the key tools that will be used to assess the acceptability of the proposed repository for high-level radioactive waste (HLW) in Japan. We present tool that demonstrate the impact of this uncertainty on the PA using Monte Carlo sampling approach for the system Performance. PA for Geological repositories are based on conceptual models of physical processes imposed in computer codes and parameters derived from field and laboratory data or from expert elicitation. Because of the variabilities and sparsity of measured data and the underlying uncertainty involved with modeling physical processes for many thousands of years, the results of any PA are uncertain. The code provides information delineating which input parameters most affect the model results and give a better understanding of the parameters that have the most influence on the model results, which can be used to improve the code. We applied this tool to the radionuclide migration through different barriers of the repository and heat transport processes. For carrying out the analysis we need a sampling space not less than 500 to ensure the analysis stability. (author)