天然气管道泄漏检测中的知识发现

摘要

As rapid development of foreign and domestic natural gas pipeline construction, the safe operation of pipeline is particularly important in the current. Based on analyzing the cause of gas pipeline leak and using modal acoustic emission method and negative pressure wave method to detect the pipeline,to get pipeline detection information, through the pretreatment leak signal data, according to data mining algorithm realize natural gas potential information mining. Based on that use decision tree classification, DBSCN cluster analysis, K-nearest neighbor algorithm to realize the knowledge discovery process, looking for internal relations, development trend and potential rules, controlling the natural gas transmission beforehand and intelligent monitoring.%随着国内外天然气管道建设的迅速发展,管道的安全运行在当前尤为重要.通过对天然气管线泄漏原因的分析,采用模态声发射法和负压波法对管线进行检测,目的是获取管道检测信息,通过预处理泄露信号数据,针对数据通过预处理以取消数据差异和冗余,采取挖掘算法实现天然气潜信息挖掘,在此采用了决策树分类、DBSCN聚类分析、K近邻算法完成知识发现,寻找管道传输的内在联系、发展趋势及潜在规则,实现天然气传输的事前控制和智能监测.