Investigation on Uncertainty Quantification Method in Realistic LOCA Analysis

In this paper, we introduce many kinds of uncertainty quantification methods used in realistic LBLOCA analysis, and quantify the uncertainties of the output of large break loss of coolant accident based on the best estimate thermal hydraulic system code CATHARE GB LBLOCA model of CPR1000 nuclear power plant. Furthermore, we compare and identify the differences in many areas such as the process of input and output parameters and analysis procedure, and recognize that the sensitivity study method is most conservative because of the existing of the artificial conservative assumption, and the normal distribution test is required in the conventional parameter statistic method, Owen factor method and Bootstrap method, but not in Wilks method, and Wilks method can achieve more realistic results.