模糊聚类法在油气管道风险评价管段划分中的应用  

Application of fuzzy clustering method to pipeline segmentation process of oil & gas pipeline risk assessment

作　　者：李大全[1,2] 张鹏[1] 艾慕阳[2] 王玉彬[2] 陈利琼[1] 

机构地区：[1]西南石油大学 [2]中国石油管道科技研究中心

出　　处：《天然气工业》2012年第7期63-67,107,共5页Natural Gas Industry

基　　金：国家自然科学基金(编号:50974105);高等学校博士学科点专项科研基金(编号:20105121110003)

摘　　要：在油气管道风险评价的管段划分过程中,若选用作为分段标准的管道属性对象太少,则管道的分段数目少,无法详细描述管道沿线各点的风险变化状况,也无法利用所得风险评价结果数据准确定位高风险管段并进行针对性的维护;相反,管道的分段数目过多,管道风险评价、结果分析的工作量则将成倍增加。为此,提出了利用模糊聚类法对油气管道进行动态分段的方法,并介绍了该方法的基本原理和一般步骤。实际应用结果表明:模糊聚类法通过选用不同的截距阵,可动态选取分段标准,使得管段划分的细密程度始终满足风险评价人员对目标的期望,使管段划分更加高效、合理,并且运用该方法选取管道属性对象作为划分标准及模糊聚类的计算过程都可以通过程序实现,提高了管段划分过程中计算的准确性和计算效率,从而提高了管道风险评价的准确性,降低了管道风险评价的成本。In the pipeline segmentation process of oil & gas pipeline risk assessment,if the pipeline property objects are too few as reference for the segmentation standard,the number of segments will be too small to help describe thoroughly the changing risk status along the pipeline.Thus,the obtained results of pipeline risk assessment can not be used to accurately locate the high-risk segments and thereby the targeted maintenance can not be performed either.On the contrary,if too many objects are selected,the workload of pipeline risk assessment and analysis will significantly mount up.Accordingly,this paper puts forward the fuzzy clustering method(FCM) and introduces its basic principles and general steps.The field application shows that the pipeline segmentation can be conducted dynamically by choosing different λ-cut matrices with the FCM,making the number of segments always meet the expectation.In conclusion,the FCM is efficient and reasonable in this case not only to improve the accuracy but to reduce the cost of pipeline risk assessment.

关 键 词：油气管道 风险评价 分段 模糊聚类 截距阵 管道属性 准确性 成本 效率 

分 类 号：TE88[石油与天然气工程—油气储运工程]