核级管道腐蚀数据的概率统计分析

针对核级管道腐蚀环境的复杂性及腐蚀过程的随机性,提出了基于概率统计方法的最大腐蚀深度预测模型。首先对核级管道进行腐蚀失效分析;其次采用广义极值分布模型（GEV）拟合管道最大腐蚀深度数据,用L-矩法计算模型的参数值,分析核级管道腐蚀深度的统计规律;最后引用回归期的概念预测管道最大腐蚀深度。以某核级管道为例,预测其最大腐蚀深度为4.575 1 mm,超过最大腐蚀深度的概率为0.75%。计算结果证明：应用极值理论作统计分析时,广义极值分布模型具有更广的适用性,该研究对分析腐蚀管道的可靠性和安全性具有一定的意义。

Probability Statistical Analysis of Corrosion Data of Nuclear Pipeline

As for the complexity of nuclear pipeline corrosion environment and the randomness of the corrosion process, a prediction model of corrosion depth was established based on the probability and statistics method. Firstly, the corrosion failure of the nuclear pipeline-rupture and perforation leakage caused by corrosion was analyzed. Secondly, the generalized extreme value distribution （GEV） model was introduced, and the model parameter was calculated by L - moments method. Finally, return period was used to compute the maximum corrosion depth of nuclear pipeline. The example indicated that the maximum corrosion depth of some nuclear pipeline was 4. 575 1 mm, and the probability was 0.75%. The result shows that the generalized extreme value distribution model has a wider applicability, which is helpful for analyzing the reliability and security of corrosion pipeline.