Abstract
One of key issues in assessing a safety of a deep geological disposal system for high-level nuclear waste is to evaluate the effects of model uncertainties and parameter value uncertainties on the results of the radionuclide migration analysis. In general, models and parameter values used for the analysis in geosphere are obtained from in-situ tests data, that have great uncertainty because of a limited number of tests conducted in a heterogeneous rock formation. Therefore, it is necessary to evaluate the uncertainties of model and parameter value obtained from in-situ tests quantitatively.
From this point of view, we investigate uncertainties due to model and parameter values obtained from inversion analysis of in-situ tracer test. This report describes the inversion analysis of Nagra MI tracer test and discusses the model uncertainties due to the number of migration pathways between injection well and extraction well and parameter value uncertainties due to data fluctuation of tracer breakthrough curve. From the results, it was found that the results of inversion analysis showed good agreement with the results of the in-situ tracer test, that the parameter value uncertainties due to data fluctuation were much smaller than the model uncertainties due to the number of migration pathways, and that the information criteria could be useful for choosing the best model among several prospective models.
