Uncertainty quantification application on the Phebus FPT1 using the coupling of MELCOR and DAKOTA in a Python environment/architecture

Abstract

During the last few years, the international nuclear scientific community involved in the development of deterministic safety analyses in the Severe Accident (SA) domain has focused its interest on the analysis of methodologies for the uncertainty quantification using the state-of-art integral SA codes (e.g. ASTEC, MAAP, MELCOR, etc.). Within this framework, the H2020 “Management and Uncertainty of Severe Accidents” (MUSA) project, coordinated by CIEMAT (Spain), aimed to establish a harmonized approach for the application of uncertainty quantification methodologies to SAs. Along with the MUSA Working Package named “Application of UQ Methods against Integral Experiments”, coordinated by ENEA (Italy), the uncertainty analysis methodology has been applied to the Phebus FPT1 experiment. To develop the reference base case, the integral code MELCOR 2.2 has been selected, and it has been coupled with the uncertainty tool DAKOTA to develop the uncertainty analysis. The study has been focused on the degradation and beginning of the aerosol phase of the test, considering the MELCOR aerosol miscellaneous constants as uncertain input parameters and the maximum value of the total mass of aerosol in suspension in the containment atmosphere as Figure of Merit (FOM). The uncertainty analysis permitted a first estimation of the uncertainty bandwidth of the FOM, comparing it with the experimental and reference case values. Among the uncertain input parameters, the dynamic shape factor and the agglomeration shape factor present, respectively, a moderate and significant statistical correlation with the FOM.