Spent Fuel Performance Assessment for a Repository System Located in a Salt Dome

摘要

The results of the uncertainty analysis for the Brine Intrusion Scenario are expressed in terms of subjective probability distributions and statistical tolerance limits. Due to the generic nature of the repository, no findings from exploratory work were available to update the subjective probability distributions at the level of scenario and parameter uncertainties. The outcome of this analysis can be regarded as the first step of an iterative process to enhance the performance of the repository. The statistical tolerance limits (e.g. for the released dose rate) can be tightened by increasing the sample size and hence by increasing the number of model runs to be performed. Despite the relatively small sample size, the sensitivity analysis identified the main contributors to the uncertainty of the results of the consequence analysis (e.g. dose rate). This way it permits useful insights and recommendations as to where the state of knowledge should be improved in order to reduce the uncertainty of the results most effectively. The model results are subject to numerous uncertainties and can therefore only be given in form of subjective probability distributions. The model results are to be compared to protection goals. These goals are, however, formulated deterministically, i.e. they ignore the inevitable uncertainty of the computed results, which they are to be compared to, and are therefore incomplete. They need to be supplemented by additional requirements. An example of such an additional requirement for the maximum annual dose rate would be. for instance: "At a confidence level of at least 95%, the maximum annual dose rate needs to be below the limit value with at least 95% subjective probability".