摘要
在基于混沌关联维数分析的管道泄漏检测与定位方法的基础上,进一步从曲线特征上定量对基于负压波的泄漏检测和定位方法给予了证明.利用Grassberger-Procaccia算法计算压力信号的关联维数,然后根据关联维数的计算结果判断压力信号中是否存在负压波来进行泄漏诊断,并从理论上证明了关联维数可以区分不同类型的压力曲线以及在压力下降沿的突变点附近关联维数达到最大的结论.现场历史数据的测试结果表明,基于关联维数分析的泄漏诊断方法在灵敏检测故障的同时,极大地降低了扰动引起的误报率.
Based on the current method of correlation dimension analysis for pipeline leak detection and location,quantitative proof based on the curve characteristic was further given.The Grassberger-Procaccia algorithm was used to calculate the correlation dimension of pressure signals.Then the resulted correlation dimension was further used as a criterion to recognize the negative pressure wave which is an indication of leak faults.And also,one conclusion could be drawn from theory that the correlation dimension can distinguish different pressure waves and the correlation dimension would reach the maximum at the starting point of the negative pressure wave(NPW′s) falling edge.The test results based on real historical data show that the leak diagnosis method based on correlation dimension analysis can sensitively detect faults,and greatly decrease false alarm rate caused by disturbances.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第S1期126-130,共5页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(60574085)
国家高技术研究发展计划资助项目(2006AA04Z428)
关键词
混沌
关联维数
输油管道
泄漏诊断
负压波
检测
定位
chaos
correlation dimension
oil pipeline
leak diagnosis
negative pressure wave
detection
location
