Probabilistic accident consequence uncertainty analysis: Dispersion and deposition uncertainty assessment, appendices A and B

Abstract

This report was prepared as an account of work sponsored by an agency of the United States Government.Neitherthe United States Government nor any agency thereof, nor any of their employees, makes any warranty, expressed cf implied, Or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any information, apparatus, product, or process disclosed in this report, or represents that its use by such third party would not infringe privately owned rights. DISCLAIMERPortions of this document may be illegible electronic image products. Images are produced from the best available original document. AbstractThe development of two new probabilistic acQdent consequence codes, MACCS and COSYMA, was completed in 1990. These codes estimate the risks presented by nuclear installations based on postulated frequencies and magnitudes of potential accidents. In 1991, the US Nuclear Regulatory Commission (NRC) and the Commission of the European Communities (CEC) began a joint uncertainty analysis bf the two codes. The ultimate objective of the joint effort was to develop credible and traceable uncertainty distributions for the input variables of the codes. As a first step, a feasibility study was conducted to determine the efficacy of evaluating a limited phenomenological area bf consequence calculations (atmospheric dispersion and deposition parameters) and to determine whether the technology exists to develop credible uncertainty distributions on the input variables for the codes. Expert elicitation was identified as the best technology available for developing a library of uncertainty distributions for the selected consequence parameters.The study was formulated jointly and was limited to the current code models and to physical quantities that could be measured in experiments. The elicitation procedure was devised from previous US and EC studies with refinements based on recent experience. Elicitation questions were developed, tested, and clarified. Sixteen internationally recognized experts from nine countries were selected using a common set of selection criteria. Probability training exercises were conducted to establish ground rules and set the initial boundary conditions. Experts developed their distributions independently. Results were processed with an equal weighting aggregation method, and the aggregated distributions were processed into code input variables.To validate the distributions generated for the wet deposition input variables, samples were taken from these distributions and fully to develop and implement a unified process for the elaboration of uncertainty distributions on consequence code input parameters. Formal expert judgment elicitation proved valuable for synthesizing the best available information. Distributions on measurable atmospheric dispersion and deposition parameters were successfully elicited from experts involved in the many phenomenological areas of consequence analysis.