【摘要】 基于数据挖掘方法对管道腐蚀缺陷进行安全评估，主要内容包括：总结现有的腐蚀管道剩余强度评价方法，收集管道失效数据对各评价模型进行对比验证；基于管道失效评估模型及各变量参数分布规律，生成一个随机数据集，包括几何特性、腐蚀坑尺寸、管道年龄及管道材料等特征，并通过蒙特卡罗模拟管道的失效概率；根据管道失效概率进行风险分类，并通过机器学习算法训练有效的管道风险分类模型。更多还原

【Abstract】 The security evaluation of pipeline corrosion based on the data mining method was implemented.The main contents include summarizing the existing residual intensity evaluation methods of corroded pipeline and collecting the pipeline failure data to compare and verify the evaluation models; basing on pipeline failure evaluation models and various variable parameter distribution rules to generate a random data set, including geometric characteristics, corrosion pits, pipeline age, pipeline materials and the probability of simulating pipelines through Monte Carlo; risk classification according to the pipeline failure probability, and training effective pipeline risk classification models through the machine learning algorithm.更多还原