作者：马云栋1, 董绍华1, 徐晴晴1, 魏昊天1, 彭东华2, 宋顶1
作者单位：1. 中国石油大学(北京)安全与海洋工程学院, 北京 102249;
2. 国家石油天然气管网集团北京管道有限公司, 北京 100000
关键词：管道微泄漏;声学内检测;改进EMD-小波阈值降噪;网格参数寻优SVM
摘要：
针对液体管道微泄漏（小于1 L/min）难以检测和识别的问题，建立基于被动声学的液体管道微泄漏内检测方法，研究管道微泄漏的声学机理，并搭建实验平台。采用管道内置高灵敏度声学传感器的方式，进行不同泄漏孔径和和泄漏内压下微泄漏声学实验，微泄漏声信号幅值随泄漏孔径和泄漏内压呈正相关变化。利用改进EMD（IEMD）-小波阈值降噪算法对微泄漏声信号进行降噪处理，以减少噪声对真实泄漏信号的影响。提取并定义不同信号处理领域（时域、频域）的特征参数，如均值、标准差、均方根、峰峰值等，以表示微泄漏的复杂性。将特征参数作为下一步管道泄漏识别的数据库，并将数据分为实验组和验证组。采用网格参数寻优支持向量机（SVM）构建自动分类模型对管道微泄漏进行识别，降低人工误判概率。研究结果表明：管道微泄漏识别准确率达到97.87%，可以实现对管道微泄漏的准确识别。研究结果可为声学内检测器的研发提供理论基础。
Study on internal detection method of small leakage in liquid pipeline based on passive acoustics
MA Yundong, DONG Shaohua, XU Qingqing, WEI Haotian, PENG Donghua, SONG Ding
1. College of Safety and Ocean Engineering, China University of Petroleum, Beijing 102249, China;
2. National Petroleum and Natural Gas Pipeline Network Group Beijing Pipeline Co., Ltd., Beijing 100000, China
Abstract: Aiming at the problem that it is difficult to detect and identity the small leakage (less than 1 L/min) of liquid pipeline, a method of internal detecting the small leakage of pipeline based on passive acoustic is established. This paper studies the acoustic mechanism of pipeline micro leakage, the experimental platform of micro leakage detection in pipeline based on passive acoustics is built. The acoustic experiments of small leakage under different leakage aperture and internal pressure are carried out by using a high-sensitivity acoustic sensor built in the pipeline, the amplitude of micro leakage acoustic signal changes positively with the leakage aperture and leakage internal pressure. The IEMD wavelet threshold denoising algorithm is used to denoise the micro leakage acoustic signal to reduce the influence of noise on the real leakage signal . The characteristic parameters of different signal processing fields (time domain and frequency domain), such as mean, standard deviation, root mean square, peak and peak, are extracted and defined to represent the complexity of small leakage. The characteristic parameters are used as the database of pipeline leakage identification in the next step, and the data are divided into experimental group and verification group. The grid parameter optimization support vector mechanism (SVM) is used to build an automatic classification model to identify the small leakage of pipeline and reduce the probability of manual misjudgment. The results show that the accuracy of pipeline small leakage identification is 97.87%, which can realize the accurate identification of pipeline small leakage. The results of this paper can provide a theoretical basis for the development of acoustic internal detector.
Keywords: pipeline small leakage;acoustic internal detection;IEMD wavelet threshold denoising;grid parameter optimization SVM
2023, 49(9):19-26 收稿日期: 2022-5-3;收到修改稿日期: 2022-6-24
基金项目: 国家重点研发计划资助项目(ZX20170128)；中国工程院资助项目(ZX20200001)；中石油战略合作科技专项(ZX20190224)
作者简介: 马云栋（1990-），男，山东日照市人，博士研究生，研究方向为无损检测、安全监测与智能诊断。
参考文献
[1] 董绍华, 田中山, 赖少川, 等. 新一代超高清亚毫米级管道内检测技术的研发与应用[J]. 油气储运, 2022, 41(1): 34-41
[2] MUJTABA M, LEMMA T A, TAQVI S V V, et al. Leak detection in gas mixture pipelines under transient conditions using hammerstein model and adaptive thresholds[J]. Processes, 2020, 8(4): 474-494
[3] KHULIEF Y A, KHALIFA A, MANSOUR R B, et al. Acoustic detection of leaks in water pipelines using measurements inside pipe[J]. Journal of Pipeline Systems Engineering and Practice, 2012, 2: 47-54
[4] CHATZIGEORGIOU D, YOUCEF-TOUMI K, BEN-MANSOUR R, et al. Design of a novel in-pipe reliable leak detector[J]. IEEE/ASME transactions on mechatronics: A Joint Publication of the IEEE Industrial Electronics Society and the ASME Dynamic Systems and Control Division, 2015, 20(2): 824-833
[5] LAVEN K, MERGELAS B, INGLEHART L, et al. Leak detection in large diameter fiberglass pipe[C]//Pipeline Division Specialty Conference, 2006.
[6] 王少峰, 刘朋真, 王建国, 等. 基于小波包熵与Gabor小波变换的管道连续型泄漏源定位[J]. 仪表技术与传感器, 2017(9): 98-102
[7] 严欣明, 郝永梅, 顾玉明, 等. Matlab小波变换在燃气管道泄漏检测中的试验研究[J]. 中国特种设备安全, 2016, 32(6): 14-17+83
[8] 王建国, 刘朋真, 王少锋, 等. 基于EMD和小波包降噪的压力管道微泄漏源定位研究[J]. 河南理工大学学报(自然科学版), 2017, 36(4): 83-88
[9] 刘文芳, 韩军, 刘艳锋, 等. 基于机器学习的金属近表面缺陷超声检测方法[J]. 中国测试, 2022, 48(1): 46-52
[10] 杨雅彬, 刘晴, 武志成, 等. 机器学习分布式网络传输异常数据智能检测方法[J]. 中国测试, 2021, 47(3): 104-109
[11] 王新颖, 江志伟, 陈海群, 等. SVM与DS结合的管道泄漏诊断方法研究[J]. 计算机工程与应用, 2016, 52(5): 262-265+270
[12] 毛兴翔, 吴世德, 王文明, 等. 基于支持向量机的燃气管道泄漏识别方法研究[J]. 石油机械, 2021, 49(7): 147-154
[13] 马大猷. 现代声学理论基础[M]. 科学出版社, 2004: 293-297.
[14] ZHENG X, HAO Z Y. Diagnosis of valve-slap of diesel of value-slap of diesel engine with eemd-emd- agst approach[J]. Transactions of Tianjin University, 2012, 18(1): 26-32
[15] ZHAO X, PATEL T H, ZUO M J. Multivariate EMD and full spectrum based condition monitoring for rotating machine[J]. Mechanical Systems & Signal Processing, 2012, 27: 712-728
[16] WANG X, LIU C, BI F, et al. Fault diagnosis of diesel engine based on adaptive wavelet packets and EEMD-fractal dimension[J]. Mechanical Systems & Signal Processing, 2013, 41(1–2): 581-597
[17] DONOHO D L. De-noising by soft-thresholding[J]. IEEE Transactions on Information Theory, 2002, 41(3): 613-627
[18] 李茂, 杨录, 张艳花. 基于EMD及主成分分析的缺陷超声信号特征提取研究[J]. 中国测试, 2018, 44(2): 118-121+133
[19] SAIN S R. The nature of statistical learning theory[J]. Technometrics, 1997, 38(4): 409-409
[20] 罗小燕, 陈慧明, 卢小江, 等. 基于网格搜索与交叉验证的SVM磨机负荷预测[J]. 中国测试, 2017, 43(1): 132-135+144
