UNCERTAINTY / SENSITIVITY METHODOLOGIES FOR SAFETY ASSESSMENTS OF LOW-LEVEL WASTE DISPOSAL FACILITIES

摘要

Safety assessment for low-level waste facilities requires the interaction of a largernnumber of disciplines in order to model environmental phenomena necessary to evaluate safetyrnof disposal. The physical systems involved can often be very complex. The initial purpose ofrnthe safety analysis is to better understand the system under study. Eventually, as the systemrnbehavior becomes understood more fully, the assessment is used to support regulatoryrndecisions. Corresponding to the specific goals of the project the objectives for uncertainty andrnsensitivity analysis will also vary, depending on the stage of the analysis.rnTypically, the safety analyst has to simplify the physical system into a conceptual modelrnthat can be modeled mathematically. The first step in this process involves defining anrnexposure scenario and this is often a significant source of uncertainty (future climaternassumptions or individual habits). Simplification of the physical system to a mathematicalrnmodel is another source of uncertainty, commonly called model uncertainty. Other sources ofrnuncertainties include parameter estimation and random variability in parameters measurements.rnIn assessments the analyst may need to rely on expert judgment due to lack of data, lackrnof knowledge concerning future conditions and parameter values (and distributions), or anyrnaspects of the system under study that are not well understood by current science. Thisrngenerates another kind of uncertainty, “subjective uncertainty”.rnThis paper is part of the work being developed by the IAEA sponsored project:rnImprovement of Long Term Safety Assessment methodologies for Near Surface radioactivernWaste Disposal Facilities (ISAM Project). As part of the ISAM Project a safety assessmentrnprocess has been proposed and it is the purpose of this paper to examine the application ofrnuncertainty and sensitivity analysis within the proposed safety assessment process.