Uncertainty Analysis for a Depressurised Loss of Forced Cooling Event of the PBMR Reactor

This paper presents an uncertainty analysis for a Depressurised Loss of Forced Cooling (DLOFC) event that was performed with the systems CFD (Computational Fluid Dynamics) code Flownex for the PBMR reactor. An uncertainty analysis was performed to determine the variation in maximum fuel, core barrel and reactor pressure vessel (RPV) temperature due to variations in model input parameters. Some of the input parameters that were varied are: thermo-physical properties of helium and the various solid materials, decay heat, neutron and gamma heating, pebble bed pressure loss, pebble bed Nusselt number and pebble bed bypass flows. The Flownex model of the PBMR reactor is a 2-dimensional axi-symmetrical model. It is simplified in terms of geometry and some other input values. However, it is believed that the model adequately indicates the effect of changes in certain input parameters on the fuel temperature and other components during a DLOFC event. Firstly, a sensitivity study was performed where input variables were varied individually according to predefined uncertainty ranges and the results were sorted according to the effect on maximum fuel temperature. In the sensitivity study, only seven variables had a significant effect on the maximum fuel temperature (greater that 5°C). The most significant are power distribution profile, decay heat, reflector properties and effective pebble bed conductivity. Secondly, Monte Carlo analyses were performed in which twenty variables were varied simultaneously within predefined uncertainty ranges. For a one-tailed 95% confidence level, the conservatism that should be added to the best estimate calculation of the maximum fuel temperature for a DLOFC was determined as 53°C. This value will probably increase after some model refinements in the future. Flownex was found to be a valuable tool for uncertainly analyses, facilitating both sensitivity studies and Monte Carlo analyses.