供水管道泄漏信号机理及弱信号检测与处理方法研究

The safe operation of water supply pipeline system is important for our life and has received extensive attention. In this project, we propose the research on detection and processing method of pipeline leak signal mechanism and weak signal background of city water supply, industrial water supply and slurry transportation underground pipeline. According to the different operating conditions and environmental conditions, we will search more accurate propagation rule to characterize the water hammer wave and sound signal of pipeline leakage. Then, overall consider the leak signal character and the even and uneven noise, establish a suitable method to detect and deal with the weak signals. This process will realize by using wavelet transform, the sequential probability ratio test, multistage hypothesis testing, adaptive linear enhancer and adaptive linear predictor methods. We will distill the Characteristics of weak leakage signal and provide a reliable theory to fix the leak location. For the sake of pipeline leakage prediction, we propose the predictive method of pipeline leakage based on data mining. In order to solve the problems such as mass information, schema integration, data redundancy and value conflict, we have design the data mining rules which is gradually deepen. These rules can help us finding the internal relations from mass leakage data, development trend and potential rules. This method will predict and avoid some of the normal leakage.

供水管道系统的安全运行问题已受到广泛关注。本课题以城市供水和工业供水及矿浆输送地下管道为背景，对管道泄漏信号机理及弱信号检测与处理方法进行研究。针对不同的运行工况和环境条件，探寻更为准确且典型的管道泄漏水击波和声信号的产生机理及传播规律。以此为基础，根据泄漏信号特征及管道内外存在的各种平稳和非平稳噪声，采用小波变换、多级假设检验、模糊神经网络、多传感器数据融合以及自适应线性增强器与自适应线性预测器相结合等方法，研究与之相适应的微弱信号检测与处理方法，实现管道泄漏微弱信号特征量的有效提取，为漏点定位提供可靠依据。为了对管道泄漏进行预测预报，本课题探索性地研究基于数据挖掘的管道泄漏预测方法，针对提取到的用于数据挖掘的海量信息，解决模式集成、数据冗余和数值冲突等问题，并设计由浅入深的数据挖掘规则，从大量的泄漏数据中挖掘管道传输的内在联系、发展趋势及潜在规则，进而预测并避免一些常态泄漏事故的发生。

供水管道系统的安全运行问题已受到业界的广泛关注。本课题以实验管道系统和工业管道系统为背景，对管道泄漏信号机理、微弱信号检测与处理方法及泄漏预测预报方法等进行了研究。利用压力和声发射等相关检测设备，进行了大量的实验测试及分析，研究了多种工况下的管道泄漏瞬变压力波和声发射信号特征，得到了比较典型的瞬变压力波传播过程和变化规律、泄漏声信号的产生机理、频率特性及传播特性，为检测设备研制及泄漏特征信号提取奠定了基础。针对不同的工况和环境条件，研究了管道泄漏微弱信号特征量的提取方法，包括傅里叶变换法、小波变换法、相关检测法、自适应滤波法及随机共振法等，并研制了相关的软硬件系统，实验测试结果验证了理论和方法的正确性，达到了预期目标。为了实现管道泄漏的预测预报，本项目重点开展了管道系统缺陷检测与评价及管道泄漏预测方法研究，并探索性地研究了基于知识的管道泄漏预测方法，具体研究内容涉及电涡流检测、超声导波检测、分布式光纤检测、机器视觉、三维可视化、数据融合、缺陷检测及管道系统完整性评价方法等，为后续的深入研究奠定了基础。本项目按任务书计划进度推进，解决了供水管道泄漏信号机理和微弱信号检测及处理方法等问题，为相关的关键技术研究和工程应用提供了理论依据。