基于样本相关性及SVM的管道泄漏检测方法研究Research on pipeline leakage detection method based on sample correlation and SVM
何健安,高炜欣,袁鹏程
摘要(Abstract)：
分析当前煤层气生产现场的实际情况,针对生产数据不全的问题,提出采用模式识别的方法对煤层气管道微量泄漏进行判断。通过分析煤层气生产现场数据的相关性曲线,明确模式识别时样本的数量范围;分析已有的模式识别方法,提出一种基于SVM的泄漏识别方法;根据煤层气生产的实际情况,分析确定适合于管道泄漏检测的核函数,并给出完整的泄漏检测算法。通过实例对所提算法进行验证,实验表明该算法对已有煤层气长输管道SCADA系统是一个有益的补充。
关键词(KeyWords)： 泄漏检测;相关性;样本数量;模式识别;支持向量机;核函数
基金项目(Foundation): 西安石油大学研究生创新与实践能力培养计划资助项目(YCS18113048)~~
作者(Author): 何健安,高炜欣,袁鹏程
DOI: 10.16652/j.issn.1004-373x.2018.23.025
参考文献(References)：
- [1]FUMESS I L A.Development in pipeline instrumentation[J]Measurement and control,1987,20(1):7-15.
- [2]张红兵,李长俊.不等温输气管道泄漏监测技术[J].石油工程建设,2005,31(1):25-27.ZHANG Hongbing,LI Changjun.Leakage monitoring technology for non-isothermal gas transmission pipelines[J].Petroleum engineering construction,2005,31(1):25-27.
- [3]夏海波,张来斌,土朝辉.国内外油气管道泄漏检测技术发展现状[J].油气储运,2001,20(1):1-5.XIA Haibo,ZHANG Laibin,TU Zhaohui.Detection technology o oil and gas pipeline leaks at home and abroad development status[J].Oil&gas storage and transportation,2001,20(1):1-5.
- [4]王桂增,董东.基于Kullback信息测度的长输管线的泄漏检测[J].信息与控制,1989(2):14-17.WANG Guizeng,DONG Dong.Leak detection of long-distance pipeline based on Kullback information measure[J].Information and control,1989(2):14-17.
- [5]唐秀家,颜大春.基于神经网络的管道泄漏检测仪器及方法[J].北京大学学报,1997,33(3):319-326.TANG Xiujia,YAN Dachun.Instrument and method for pipeline leakage detection based on neural network[J].Journal of Beijing University,1997,33(3):319-326.
- [6]NASH G,KARNEY B.Efficient inverse transient analysis in series pipe systems[J].Journal of hydraulic engineering,1999,125(7):761-764.
- [7]张学工.模式识别[M].北京:清华大学出版社,2010.ZHANG Xuegong.Pattern recognition[M].Beijing:Tsinghua University Press,2010.
- [8]CRISTIANINI Nello.支持向量机导论[M].北京:电子工业出版社,2004:82-106.CRISTIANINI Nello.An introduction of support vector machines and other kernel-based learning methods[M].Beijing:Electronic Industry Press,2004:82-106.
- [9]邓蕊,马永军,刘尧猛.基于改进交叉验证算法的支持向量机多类识别[J].天津科技大学学报,2007,22(2):58-61.DENG Rui,MA Yongjun,LIU Yaomeng.Support vector based on improved cross validation algorithm multi-machine recognition[J].Journal of Tianjin University of Science&Technology,2007,22(2):58-61.
- [10]王行甫,陈家伟.基于高斯核的SVM的参数选择[J].计算机系统应用,2014,23(7):242-245.WANG Xingfu,CHEN Jiawei.Parameter selection of Gaussian kernel based SVM[J].Computer systems&applications,2014,23(7):242-245.
- [11]孙即祥.现代模式识别[M].北京:高等教育出版社,2008.SUN Jixiang.Modern pattern recognition[M].Beijing:Higher Education Press,2008.
- [12]连雷,姜守达,刘晓东.基于微分进化算法的SVM参数选择[J].哈尔滨工程大学学报,2009,30(2):138-141.LIAN Lei,JIANG Shouda,LIU Xiaodong.SVM parameter selection based on differential evolution algorithm[J].Journal of Harbin Engineering University,2009,30(2):138-141.