The acquisition of data for use in the probabilistic risk assessment of underground disposal of radioactive waste

In developing a methodology for assessing potential sites for the disposal of radioactive wastes, the UK Department of the Environment has conducted a series of trial assessments. For the trial assessments there are few measured data available and the input probability distributions to the SYVAC A/C framework of Monte Carlo simulations have to be assessed from expert opinion. A feasibility study was conducted on the use of a formal method for encoding probability distributions. Three parameters were successfully encoded as part of the study and helpful conclusions drawn. At the stage of a more detailed site assessment measured data will be available and these need to be incorporated into the subjective distributions obtained from experts. Two illustrative case studies were conducted in which a Bayesian approach was used to incorporate the measured data. The first case study illustrated how the measured data can substantially reduce the uncertainty about a parameters value whereas the second case study contained a large amount of measurement uncertainty