输气管道泄漏音波传声特性及泄漏识别  

Acoustic Transmission Property of Gas Pipeline Leakage

作　　者：忻盛婷 梁伟[1] 张来斌[1] 

机构地区：[1]中国石油大学(北京)机械与储运工程学院

出　　处：《石油机械》2013年第6期110-114,共5页China Petroleum Machinery

基　　金：国家自然科学基金项目(51005247);国家油气重大专项课题(2011ZX05055);北京市科技新星计划项目(2010B068)

摘　　要：介绍了音波泄漏检测与定位的基本原理,研究了管道泄漏音波的传声特性,利用输气管道泄漏检测平台进行了不同输气压力下不同泄漏孔径的泄漏检测试验,分析了不同条件下的声谱特征。根据管道泄漏音波不同信号特征量建立不同条件下的隐马尔可夫模型,对管道泄漏进行了分类识别。试验分析结果表明,该方法有较好的识别率,隐马尔可夫模型应用于输气管道泄漏识别是可行的,且比支持向量机具有更高的识别准确率。该方法为天然气管道泄漏的进一步研究提供了理论基础。The basic principle of sound wave leakage detection and positioning was introduced in the paper. The acoustic transmission property of pipeline leakage sound wave was studied. The leakage detection platform of gas pipeline was applied to conduct the leakage detection test with different leaking hole diameters under different gas transmission pressures. The acoustic spectrum feature in different conditions was analyzedAccording to dift^rent characteristic quantities of signals, the hidden Markov model （HMM） was established in different conditions. The classification identification of pipeline leakage was conducted. The test analysis shows that the method has a desira- ble identification, proving that it is feasible to apply the HMM to the identification of gas pipeline leakage. Moreover, it has higher identification accuracy than support vector machine （SVM） . The method provides the theoretical basis for further study of natural gas pipeline leakage.

关 键 词：天然气管道 音波 隐马尔可夫模型 泄漏 识别 

分 类 号：TE973.6[石油与天然气工程—石油机械设备]