传感技术学报
傳感技術學報
전감기술학보
Journal of Transduction Technology
2015年
7期
1097-1102
,共6页
韩玲娟%王强%杨其华%范昕炜
韓玲娟%王彊%楊其華%範昕煒
한령연%왕강%양기화%범흔위
管道泄漏检测与定位%分布式光纤传感%小波包分解%欧氏距离法%虚警率
管道洩漏檢測與定位%分佈式光纖傳感%小波包分解%歐氏距離法%虛警率
관도설루검측여정위%분포식광섬전감%소파포분해%구씨거리법%허경솔
pipeline leakage detection and localization%distributed optical fiber sensor%wavelet packet decomposi-tion%Euclidean distance method%false alarm rate
介绍了一种基于Sagnac/Mach-Zehnder混合干涉仪原理的分布式光纤传感水下输气管道泄漏检测及定位系统。在强干扰微小泄漏情况下,通过分布式光纤传感器获取由管道泄漏引起的光干涉信号,采用小波包方法对信号进行多层分解,提取小波包能量特征,用欧式距离法对管道泄漏与否进行聚类分析;对泄漏系统利用零点频率进行精确定位,并将基于零点频率的传统检测定位、经小波去噪和最小二乘曲线拟合处理后的检测定位及本文提出的检测定位三种方法进行对比分析,统计了系统虚警率(FAR)。实验结果表明,该系统能准确识别管道泄漏,本文提出的泄漏检测定位法虚警率最低,与经小波去噪和最小二乘曲线拟合处理后的检测定位法相比,虚警率降低了6.6%,提高了系统泄漏检测定位的可靠性。
介紹瞭一種基于Sagnac/Mach-Zehnder混閤榦涉儀原理的分佈式光纖傳感水下輸氣管道洩漏檢測及定位繫統。在彊榦擾微小洩漏情況下,通過分佈式光纖傳感器穫取由管道洩漏引起的光榦涉信號,採用小波包方法對信號進行多層分解,提取小波包能量特徵,用歐式距離法對管道洩漏與否進行聚類分析;對洩漏繫統利用零點頻率進行精確定位,併將基于零點頻率的傳統檢測定位、經小波去譟和最小二乘麯線擬閤處理後的檢測定位及本文提齣的檢測定位三種方法進行對比分析,統計瞭繫統虛警率(FAR)。實驗結果錶明,該繫統能準確識彆管道洩漏,本文提齣的洩漏檢測定位法虛警率最低,與經小波去譟和最小二乘麯線擬閤處理後的檢測定位法相比,虛警率降低瞭6.6%,提高瞭繫統洩漏檢測定位的可靠性。
개소료일충기우Sagnac/Mach-Zehnder혼합간섭의원리적분포식광섬전감수하수기관도설루검측급정위계통。재강간우미소설루정황하,통과분포식광섬전감기획취유관도설루인기적광간섭신호,채용소파포방법대신호진행다층분해,제취소파포능량특정,용구식거리법대관도설루여부진행취류분석;대설루계통이용영점빈솔진행정학정위,병장기우영점빈솔적전통검측정위、경소파거조화최소이승곡선의합처리후적검측정위급본문제출적검측정위삼충방법진행대비분석,통계료계통허경솔(FAR)。실험결과표명,해계통능준학식별관도설루,본문제출적설루검측정위법허경솔최저,여경소파거조화최소이승곡선의합처리후적검측정위법상비,허경솔강저료6.6%,제고료계통설루검측정위적가고성。
For underwater gas pipeline leakage point detection and localization,a distributed optical fiber sensor system was proposed based on a hybrid configuration of Sagnac/Mach-Zehnder interferometer. Tiny leakage signals were covered by strong interferences,and optical interferometric signals caused by pipeline leakage can be obtained by the distributed optical fiber system. Wavelet packet is used to deal with interferometric signals,and signals were decomposed into different frequency bands to extract energy features. Those features were clustered with Euclidean distance to discriminate pipeline leakage. The leakage point can be captured from null frequency spectrum. Three leakage localization methods (the traditional detection based on null frequency,the detection of wavelet de-noising and least squares curve fitting process,the proposed method) performances were compared,and the false alarm rate (FAR) is calculated. The experimental results show that the system can identify leakage point accurately. The pro?posed method owns the lowest FAR in three methods. Compare to the wavelet de-noising and least squares curve fit?ting method,the FAR of proposed method is reduced about 6.6%,and it improves the reliability of leakage detec?tion and localization system.