Knowledge representation in HLW disposal facility safety assessment

A good understanding, by the public and licensing authorities, of the decision making process of radioactive waste disposal facilities safety assessment is crucial for confidence building and further discussions of the results. To this end, uncertainty analysis plays a very important role in such a multi and interdisciplinary study in which it is required the interaction of experts of several different fields of research including coupling of computer models with different standards and formats for input data and results. The complexity of a safety assessment, including the interaction between experts from several different fields of sciences, makes it very difficult to accommodate the uncertainties, that stem from several different sources, in a common framework. For example, data that support calculations for useful lifetime of concrete structures provide different levels of evidence than data that support selection of parameters values on simulations of radionuclides migration in the geosphere or prediction of redox front movement rates. Therefore, a methodology is needed that is capable of analyzing uncertainties for all data sets in the same framework, including propagation and quantification of degrees of support and evidence for results. In 2004 a technical co-operation project was established between the National Nuclear Energy Commission, Brazil, and the STUK- the Radiation and Nuclear Safety Authority, Finland, for the development of an alternative study of uncertainty through the application of fuzzy sets and evidence theory. The objective of this project is to study the various sources of uncertainties, objective and subjective, and their effects on decisions throughout. (author)