最佳估算方法中不确定度评估关键问题分析

应用最佳估算+不确定度（BEPU）分析方法对核电厂进行事故分析或安全评审已成为国际发展趋势。本文对最佳估算分析中基于输入传递的统计类不确定度评估的流程进行了总结，并对其关键步骤进行了分析和研究。分析认为，评估流程可分为确定目标参数、确定重要输入参数及其分布、抽样、模型分析和目标参数分析5步，其中现象识别和重要度排序表（PIRT）是一种适用的重要输入参数确定方法，输入参数的分布需根据试验数据或专家判断确定；抽样方法上，可采用参数抽样或非参数抽样，后者可大幅减小抽样数量；不确定度评估所用模型须经过充分试验或分析证明其适用性；通过对目标参数进行统计，可获得不确定度范围及输入参数的敏感性。

Analysis on Key Issue of Uncertainty Evaluation in Best Estimate Method

The application of best estimate plus the uncertainty (BEPU ) analysis method becomes trend in the nuclear power plant accident analysis and safety review .The pro-cedure of the statistical BEPU method based on input parameters' propagation was sum-marized and the key steps were analyzed .The evaluation process can be divided into five major steps :Defining the target parameter (s) ,identifying important input parameters and their probability distribution ,sampling ,model analysis and target parameter analy-sis .It's believed that phenomenon identification and ranking table (PIRT ) is a suitable method to identify the important input parameters .The distribution of input parameters is usually obtained from the experimental data or expert judgment .The parametric or nonparametric sampling method can be used to determine the required sampling number , and the later greatly decreases sampling number .Its calculation model should be verified sufficiently to demonstrate its applicability .The statistical results of objective parameter provide uncertainty range and the sensitivity of the input parameters .