埋地供水管道由于管龄增加、年久失修、腐蚀变形、不当施工破坏等原因经常出现泄漏，不仅加重了水资源短缺状况，其持续的冲刷对地表路面、地下其它城市生命线设施及周围建筑结构等造成巨大威胁，给城市公共安全造成诸多隐患和风险。本文以检测埋地管道泄漏源为主要研究目标，开展了泄漏信号传播规律、泄漏源水平与纵深定位、泄漏信号降噪等方面的研究。主要工作如下：研究了泄漏信号在管内、管壁、地表的传播特性，仿真模拟了泄漏信号从管道泄漏口向周围空间辐射传播过程。构建了基于直角坐标系与极坐标系的时延定位数值分析模型。设计了全尺寸室外实验平台，分析了传感器不同轴向、阵列阵型、管道压力、埋深、水平间隔、土质等要素对于泄漏信号检测与定位的影响。研究了基于地表传感器阵列信号采集和成像的泄漏源平面定位方法。提出了基于矩形传感器阵列的整体成像和分析方法。根据不同水平与高度间隔下泄漏信号的衰减规律，提出了泄漏信号有效频带筛选方法和频域切片序列成像方法。基于序列图像规模、能量分布规律等实现了序列图像特征提取、衰减和融合，构建了泄漏源识别和在地表平面（x、y坐标）的定位方法。研究了基于主动激励的增强剪切波埋地管道泄漏源纵深定位技术。设计了模态激振器和分时分段激励模式，并进行成像和序列图像分割。针对序列图像中纵深目标所在的区域特征，研究了纵深泄漏源的多种特征模态，构建了特征提取和融合的序列切片成像方法，并获得泄漏源纵深z坐标。通过对增强剪切波噪声的分析，发现非固定噪声源对增强剪切波成像的影响等效于在成像中“加雾”效果。提出了时频二维滤波和增强独立分量分析（TFF-EICA）算法及通道信号相似度匹配融合算法。其中，基于TFF-EICA算法对地表阵列传感器中多通道信号进行差分增强和分离降噪。使用通道相似度匹配融合算法对降噪后输出通道进行重定序。揭示了泄漏信号能量累和分布规律及低频能量占比规律，通过对提取的可疑泄漏信号通道进行对比校验和相关性分析，验证了该方法的有效性。结合已获得的水平x、y坐标、纵深z坐标及各空间坐标系相对位置关系，对两次定位结果进行空间重定位融合，映射出埋地管道泄漏源x、y、z轴三维坐标。实现了埋地供水管道泄漏源检测和三维定位。本文研究工作对城市供水管网泄漏检测、城市公共安全突发事件预警预防具有一定指导意义和实践价值。
Buried water supply pipes are susceptible to leakage, due to aging, disrepair, corrosion, deformation, improper construction damage. This not only aggravates the water shortage situation, but also poses a great threat to surface roads, underground lifeline facilities and surrounding building structures, causing many hidden dangers and risks to urban public safety. The main objective of the paper is to take the surface detection of buried pipeline leakage sources, and investigate on the propagation law of leakage signals, the horizontal and depth localization of leakage sources, and the noise reduction of leakage signals. The main work is as follows.The propagation characteristics of the leak signal in the pipeline system are studied, and the radiation propagation of the leakage signal from the pipe leakage hole to the surrounding space is simulated. A numerical analysis model based on the right-angle and polar coordinate systems for time delay localization is established. A full-scale outdoor experimental analysis platform is designed to analyze the effects of different factors, including the axial direction, array type, pipe pressure, burial depth, horizontal spacing and soil type on the detection and localization of leakage signals.A method for locating the leak source is investigated based on signal acquisition and imaging of surface sensor arrays. An overall imaging method is proposed using rectangular sensor arrays. Based on the decay law of the leakage signal at different horizontal and depth intervals, the effective frequency band selection method and the frequency domain slicing sequence imaging method are proposed. In the method, the sequence image features extraction, attenuation and fusion are carried out in terms of the scale of the sequence image and energy distribution law.The enhanced excitation method of shear wave is investigated for depth localization of buried pipeline leak source. Modal exciters and time-divisional excitation modes are designed, and imaging and sequence image segmentation are performed. Multiple feature modes of the leak source are studied for the feature analysis of the region where the depth target is located in the sequence image, and the sequence slice imaging is constructed for feature extraction and fusion, and the depth z-coordinate of the leak source is generated. It is found that the effect of non-fixed noise sources is equivalent to the effect of "fogging" in the imaging, which can be eliminated effectively in the proposed method.The time-frequency two-dimensional filtering and enhanced independent component analysis (TFF-EICA) algorithm and the channel signal similarity matching fusion algorithm are proposed. The TFF-EICA algorithm is used to differential enhance and separate the noise reduction of multi-channel signals from the surface array sensors. The channel similarity matching fusion algorithm is used to re-sequence the output channels after noise reduction. The energy accumulation and distribution law of the leakage signal and the low frequency energy proportion law are revealed. The effectiveness of the method is verified through the comparative verification and correlation analysis of the extracted suspected leakage signal channels.Combining the obtained horizontal x-y coordinates, depth z coordinates and the relative position relationship of each spatial coordinate system, the two localization results are spatially relocated and fused to map the buried pipeline leakage source in the three-dimensional coordinates. The buried water supply pipeline leakage source detection and three-dimensional positioning are achieved. The research work presented in the thesis has certain guiding significance and practical values for urban water supply pipeline network leakage detection and urban public safety emergencies early warning and prevention.
