基于不同抽样方法的核电厂大破口失水事故BEPU分析OA

The use of the best estimate plus uncertainty(BEPU)analysis method for nuclear power plant accident analysis has become an international trend.Random sampling based on specific sampling methods for uncertain input parameters is an important part of the uncertainty evaluation process.As the input of the optimal estimation program,random sampling parameters will directly affect the uncertainty quantification results of response parameters.At present,there is no systematic work on uncertainty quantification and sensitivity analysis based on different sampling methods;uncertainty quantification and sensitivity analysis using different sampling methods have not been systematically carried out.Taking the large break loss of coolant accident(LBLOCA)of a million kilowatt level pressurized water reactor nuclear power plant as the analysis object,important input parameters were randomly sampled based on simple random sampling(SRS)and Latin hypercube sampling(LHS).Uncertainty quantification and sensitivity analysis were conducted on key safety parameters,and the differences in the calculation results between the two random sampling methods were evaluated.The calculation results indicate that Latin hypercube sampling is more efficient in reproducing the distribution characteristics of uncertain input parameters than simple random sampling;the upper limit of one-sided statistical tolerance for peak cladding temperature obtained using two sampling methods meets the limit specified in the acceptance criteria;Based on the two sampling methods,the peak cladding temperature calculated using Wilks'non-parametric statistical theory is relatively close;using the global sensitivity analysis method,the main influence parameters of the cladding peak temperature identified by the two sampling methods are the same.