Uncertainty Assessment for PA Models

A mathematical model comprises input variables, output variables and equations relating these quantities. The input variables may vary within some ranges, reflecting either our incomplete knowledge about them (epistemic uncertainty) or their intrinsic variability (aleatory uncertainty). Moreover when solving numerically the equations of the model, numerical errors are also arising. The effects of such errors and variations of the inputs have to be quantified in order to asses the model s range of validity. The goal of uncertainty analysis is to asses the effects of parameter uncertainties on the uncertainties in computed results. The purpose of this report is to give an overview of the most useful probabilistic and statistic techniques and methods to characterize uncertainty propagation. Some examples of application of these techniques for PA applied to radioactive waste disposal are given.