摘要
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。
The actual situation of current coalbed methane(CBM)production site is analyzed.For the problem of incomplete production data,the method of pattern recognition is used to judge the minimum leakage of CBM pipeline.The range of sample number for pattern recognition is determined by analyzing the correlation curve of production site data of CBM.A leakage recognition method based on SVM is proposed by analyzing the existing pattern recognition methods.According to the actual situation of CBM production,the kernel function suitable for pipeline leakage detection is analyzed and determined,and a complete leakage detection algorithm is given.The algorithm is verified with an example.The experimental result shows that the algorithm is a helpful supplement for SCADA system of CBM long?distance pipeline.
作者
何健安
高炜欣
袁鹏程
HE Jian’an;GAO Weixin;YUAN Pengcheng(Shaanxi Key Laboratory of Measurement and Control Technology for Oil and Gas Wells,Xi’an Shiyou University,Xi’an 710065,China;Ministry of Education Key Laboratory of Photoelectric Logging and Detecting of Gas and Oil,Xi’an Shiyou University,Xi’an 710065,China)
出处
《现代电子技术》
北大核心
2018年第23期118-122,共5页
Modern Electronics Technique
基金
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)~~
关键词
泄漏检测
相关性
样本数量
模式识别
支持向量机
核函数
leakage detection
correlation
sample number
pattern recognition
support vector machine
kernel function
