In the current project, a new systematic approach has been developed, called SAPIUM (systematic approach for input uncertainty quantification methodology) for transparent and rigorous model IUQ. This approach is based on inverse propagation of the information associated with the discrepancy between simulation results and experimental data, using verified and validated simulation models. The starting point of the work performed in SAPIUM is the available state of knowledge from previous related NEA projects and current practices in regulation, industries and research.
The SAPIUM project was organised as a writing group with contributions from ten organisations, including technical support organisations, industry and universities. By clearly compiling the existing approaches for IUQ, as well as the methodological tools to handle each step, the SAPIUM report offers a shared and better understanding about appropriate “good practices” for model input uncertainty quantification. This should improve the reliability of the BEPU analysis and facilitate progress on the validity of extrapolation of its results to the nuclear power plant case. The SAPIUM report does not describe a ready-to-use methodology for model input uncertainty quantification, but rather identifies a number of generic steps and requirements that need to be covered in order to successfully address the key issues identified in previous benchmark studies for IUQ.