摘要
对模式识别燃气管网泄漏检测方法进行了系统的阐述,利用天津城建大学燃气管网实验平台,对基于支持向量机模式识别燃气管网泄漏检测方法进行验证,并提出该方法泄漏检测的敏感程度问题.通过实验,对不同泄漏量的情况下模式识别燃气泄漏检测方法所能达到的准确率进行统计,发现在泄漏量小到一定程度时,模式识别泄漏检测将无法正确区分管网是否发生泄漏;分析了泄漏点位置对节点压力的影响,发现泄漏点位置影响泄漏检测准确率.
In this paper,the detection method of pattern recognition for gas pipeline leakage is systematically expounded. Based on the experiment platform of gas pipeline network in Tianjin Chengjian University,the detection method of gas pipeline network leakage based on support vector machine pattern recognition is verified, and the question of the sensitivity of the method is put forward. Through the experiment,the accuracy of the gas leakage detection method for pattern recognition with different leakage rate is calculated through statistics. The research finds that in a small amount of leakage to a certain extent, the pattern recognition leakage detection will not be able to correctly identify the leakage of the pipe network. It also analyzes the influence of the leakage location on the node pressure,and finds the influence of the leakage location on the leakage detection accuracy.
出处
《天津城建大学学报》
2017年第6期448-452,共5页
Journal of Tianjin Chengjian University
关键词
模式识别
燃气管网
燃气泄漏
检测敏感度
pattern recognition
gas pipeline
gas leakage
detection sensitivity
