Uncertainty and sensitivity analysis of the LOFT L2-5 test: Results of the BEMUSE programme

This paper presents the results and the main lessons learnt from the phase 3 of BEMUSE, an international benchmark activity sponsored by the Committee on the Safety of Nuclear Installations [CSNI: Committee on the Safety of Nuclear Installations (NEA, OECD), 2007. BEMUSE Phase III Report. NEA/CSNI R(2007) 4, October 2007] of the OECD/NEA. The phase 3 of BEMUSE aimed at performing Uncertainty and Sensitivity Analyses of thermal-hydraulic codes used for the calculation of LOFT L2-5 experiment, which simulated a Large-Break Loss-of-Coolant-Accident (LB-LOCA). Eleven participants coming from ten organisations and eight countries took part in this benchmark. In the first section of this paper, the context of BEMUSE is described as well as the methods used by the participants. In the second section, the results of the benchmark are presented. The majority of the participants find uncertainty bands which envelop the experimental data fairly well, however the width of these bands is much diverged. A synthesis of the sensitivity analysis results has been made and is expected to provide a useful basis for further uncertainty analysis dealing with LB-LOCA. Finally, recommendations are given both for uncertainty and sensitivity analysis.