Abstract:In addition to the issues of water quality and economic losses, leakages in pipe networks are a potential danger to public safety. Early detection of water leaks in pipe networks is crucial for implementing countermeasures to reduce structural damage and water losses. The method of data collection based on sensors installed in pipeline networks has gained a lot of attention due to its potential for application in real-time monitoring systems for leak detection. However, many models developed on data collected from simulations, engineered tests or field experiments, and few have been validated using real network data. To make further investigations in real water distribution systems (WDSs), machine learning (ML)-based leak detection models were developed and validated using signal data collected by piezoelectric accelerometers installed in WDSs over several cities of China. Nonlinear features describing the non-stationary characteristics contained in leak signals are extracted from the signals to build feature sets as inputs to the common classifiers, i.e., support vector machine (SVM), decision tree (DT), and K-nearest neighbor (KNN). In addition, the spectrograms obtained by the short-time Fourier transform processing are used for the convolutional neural network SqueezeNet. The effects of different features, different dataset compositions, and different models for leak detection are compared and analyzed. SqueezeNet has the best performance with 95.15% accuracy in identifying leaks, and KNN is the best of the three classifiers with superior sensitivity and 88.17% accuracy in identification. This paper demonstrates the effectiveness and practicability of using machine learning models to identify leaks in real pipe networks. The application of machine learning models can perform effective leak detection with minimal human intervention, thus providing a meaningful contribution to the study of leak detection in real WDSs. The proposed model aims to identify leakages under operational conditions in real-life pipe networks with higher efficiency and reduced manual involvement. This paper demonstrates the applicability of accelerometers based on microelectromechanical systems for leak detection and establishes real network-based machine learning models; thereby, providing meaningful contribution to the literature regarding leak detection in a real WDS.

Measurement of Buried Iron Water Pipe Leak Based on Acoustic Emission Sensors and ChiMerge Rough Set Theory

Measurement of Leak Location in Buried Iron Water Pipe with Acoustic Emission and EMD

Measurement Of Water Pipe Leak Based On Sensors And Empirical Mode Decomposition Of Power Feature

In-Situ Leak Detection in Urban Water Supply Pipelines Using Flexible Vibration Sensors

Evaluation of Pipe Axial Defeat Through the Use of Ultrasonic Guided Wave and Matching Pursuit

Implementation of Water Pipeline Leak Detecting and Location System Based on Virtual Instruments

Research on Leak Detection for Pipelines Based on Acoustic Multi-Sensor Date Fusion

Detection of Water Pipe Leak Location Using Wavelet Packet Decomposition and Power Feature Extraction

Voidage Measurement of Air-Water Two-phase Flow Based on ERT Sensor and Data Mining Technology

Design of a Wireless Intelligent Sensor for Pipeline Leak Detection Based on Acoustic Emission Technology

An Improved Identification Method of Pipeline Leak Using Acoustic Emission Signal

In-Line Acoustic Device Inspection of Leakage in Water Distribution Pipes Based on Wavelet and Neural Network

Remote Measurement Of The Pipe Thickness Reduction By Microwaves

1509 Microwave Measurement and Evaluation of Wall Thinning in Metal Pipes

Leak Detection in Water Distribution Systems by Classifying Vibration Signals

Application of acoustic intelligent leak detection in an urban water supply pipe network

Anomaly Detection in Targeted Pipe Sections in Water Pipe Systems Using Hydroacoustic Signal Deconvolution

Adaptive Acoustic Leak Location of Long Buried Pipe via Transient Optimization Statistical Analysis

An Improved Method for Pipeline Leakage Localization With a Single Sensor Based on Modal Acoustic Emission and Empirical Mode Decomposition With Hilbert Transform

Multi-source information fusion to identify water supply pipe leakage based on SVM and VMD

Experimental Investigation into the Acoustic Characteristics of the Ground Surface Response Due to Leakage in Buried Water-Filled Pipelines