Early Leak and Burst Detection in Water Pipeline Networks Using Machine Learning Approaches - Research

Aug 3, 2026

Leakages in water distribution networks create major water wastage and higher operational costs, while conventional detection techniques struggle with complex or subtle patterns and often cannot reliably distinguish conditions. A comprehensive study compares fourteen machine learning algorithms using multi-class classification on data collected from an experimental site under leak, major leak, and no-leak scenarios. Performance is assessed with metrics including accuracy, precision, recall, and F1-score, showing strong results for Random Forest, K-Nearest Neighbours, and Decision Tree. The findings support scalable, automated detection of leaks and bursts for real-time monitoring, improved infrastructure management, and more sustainable water resource use.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages19

Reading Time48 min

AccessFree to download

Introduction

Challenges of traditional leak management

Move toward integrated proactive solutions

Early Leak and Burst Detection in Water Pipeline Networks Using Machine Learning Approaches This is the Published version of the following publication Joseph, Kiran, Shetty, Jyoti, Patnaik, Rahul, Matthew, Noel S, van Staden, Rudi, Liyanage, Wasantha P, Powell, Grant, Bennett, Nathan and Sharma, Ashok (2025) Early Leak and Burst Detection in Water Pipeline Networks Using Machine Learning Approaches. Water, 17 (14) . p. 2164. ISSN 2073- 4441 The publisher’s official version can be found at [https://doi.org/10.3390/w17142164](https://doi.org/10.3390/w17142164) Note that access to this version may require subscription. Downloaded from VU Research Repository [https://vuir.vu.edu.au/49536/](https://vuir.vu.edu.au/49536/) Article Early Leak and Burst Detection in Water Pipeline Networks Using Machine Learning Approaches Kiran Joseph 1,2, Jyoti Shetty 3, Rahul Patnaik 3, Noel S. Matthew 3, Rudi Van Staden 1, Wasantha P. Liyanage 1, Grant Powell 2, Nathan Bennett 2 and Ashok K. Sharma 1, * Academic Editors: Gui Jin, Jun Yang and Hongwei Zhang Received: 14 May 2025 Revised: 3 July 2025 Accepted: 11 July 2025 Published: 21 July 2025 Citation: Joseph, K.; Shetty, J.; Patnaik, R.; Matthew, N.S.; Van Staden, R.; Liyanage, W.P.; Powell, G.; Bennett, N.; Sharma, A.K. Early Leak and Burst Detection in Water Pipeline Networks Using Machine Learning Approaches. Water 2025, 17, 2164 . [https://doi.org/10.3390/](https://doi.org/10.3390/)[ ](https://doi.org/10.3390/)w17142164 Copyright: © 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license ([https://creativecommons.org/](https://creativecommons.org/)[ ](https://creativecommons.org/)[licenses/by/4.0/](licenses/by/4.0/)) . 1 Institute for Sustainable Industries and Liveable Cities, Victoria University, Melbourne, VIC 3011, Australia; [kiran.joseph2@live.vu.edu.au](kiran.joseph2@live.vu.edu.au) (K.J.); [rudi.vanstaden@vu.edu.au](rudi.vanstaden@vu.edu.au) (R.V.S.); [wasantha.pallewelaliyanage@vu.edu.au](wasantha.pallewelaliyanage@vu.edu.au) (W.P.L.) 2 Greater Western Water, Melbourne, VIC 3429, Australia; [grant.powell@gww.com.au](grant.powell@gww.com.au) (G.P.); [nathan.bennett@gww.com.au](nathan.bennett@gww.com.au) (N.B.) 3 RV College of Engineering, Mysore Road, Bengaluru 560059, India; [jyothis@rvce.edu.in](jyothis@rvce.edu.in) (J.S.); [rahulpatnaik.cd23@rvce.edu.in](rahulpatnaik.cd23@rvce.edu.in) (R.P.); [noelshajimathew.cd23@rvce.edu.in](noelshajimathew.cd23@rvce.edu.in) (N.S.M.) * Correspondence: [ashok.sharma@vu.edu.au](ashok.sharma@vu.edu.au) Abstract Leakages in water distribution networks pose a formidable challenge, often leading to substantial water wastage and escalating operational costs. Traditional methods for leak detection often fall short, particularly when dealing with complex or subtle data patterns. To address this, a comprehensive comparison of fourteen machine learning algorithms was conducted, with evaluation based on key performance metrics such as multi-class classification metrics, micro and macro averages, accuracy, precision, recall, and F1-score. The data, collected from an experimental site under leak, major leak, and no-leak scenarios, was used to perform multi-class classification. The results highlight the superiority of models such as Random Forest, K-Nearest Neighbours, and Decision Tree in detecting leaks with high accuracy and robustness. Multiple models effectively captured the nuances in the data and accurately predicted the presence of a leak, burst, or no leak, thus automating leak detection and contributing to water conservation efforts. This research demonstrates the practical benefits of applying machine learning models in water distribution systems, offering scalable solutions for real-time leak detection. Furthermore, it emphasises the role of machine learning in modernising i

FAQs

Why do traditional leak detection methods struggle in water pipeline networks?

Traditional approaches can be labor-intensive, limited by environmental noise, and less effective in complex or buried infrastructures, and they often do not precisely locate leaks.

How does the study evaluate machine learning models for leak and burst detection?

It compares fourteen machine learning algorithms using multi-class classification on experimental data and evaluates outcomes with metrics such as accuracy, precision, recall, micro/macro averages, and F1-score.

Which machine learning models showed the best performance in detecting leaks?

The results highlight superior performance from Random Forest, K-Nearest Neighbours, and Decision Tree, demonstrating high accuracy and robustness.

Aug 3, 2026

Upload by Seraphina