An Intelligent Data-Driven Framework for Water Leakage Detection and Maintenance Decision Support in Water Distribution Networks

Sunanth H; Akila K

doi:10.36948/ijfmr.2026.v08i05.88957

An Intelligent Data-Driven Framework for Water Leakage Detection and Maintenance Decision Support in Water Distribution Networks

Author(s)	Mr. Sunanth H, Dr. Akila K
Country	India
Abstract	Water leakage in distribution networks results in significant water loss and makes maintenance more difficult, especially when the leakage occurs in underground pipelines. Pressure measurements provide useful information about changes in network behaviour, but interpreting measurements from multiple locations can be difficult when demand conditions also vary. This paper presents an integrated data-driven framework for leakage detection, leak-location prediction, severity classification, water-loss estimation, and maintenance decision support. The framework uses the EPANET NET3 benchmark network and the Water Network Tool for Resilience (WNTR) to generate simulated leakage scenarios and corresponding pressure measurements. Pressure values from ten selected monitoring nodes, together with a network demand factor, are used as inputs to three Random Forest models. The first model predicts the probable leak location, the second predicts leak severity, and the third estimates daily water loss. A rule-based layer converts the prediction results into maintenance priority and recommendation. To demonstrate sensor-to-application communication, an ESP32 with virtual sensor data is simulated using Wokwi and connected through MQTT to a Python prediction application. A Streamlit dashboard is used to display the final results. The final dataset contains 1000 simulated records with 16 variables. Using the saved models and the evaluation procedure described in this work, the leak-location model achieved 97.50% accuracy, the severity model achieved 98.50% accuracy, and the water-loss regression model achieved an R² value of 0.9979. The results show that the proposed prototype can connect hydraulic simulation, machine learning, simulated IoT communication, and maintenance-oriented decision support in a single workflow. Since the current dataset and sensor layer are simulated, further validation using physical sensors and real water-distribution data is required before practical deployment.
Keywords	Water Leakage Detection, Water Distribution Network, Leak Localization, Random Forest, EPANET, WNTR, MQTT, IoT, Water Loss Estimation, Maintenance Decision Support
Field	Engineering
Published In	Volume 8, Issue 5, September-October 2026
Published On	2026-10-03
DOI	https://doi.org/10.36948/ijfmr.2026.v08i05.88957

View / Download PDF File

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research

A Widely Indexed Open Access Peer Reviewed Multidisciplinary Bi-monthly Scholarly International Journal

An Intelligent Data-Driven Framework for Water Leakage Detection and Maintenance Decision Support in Water Distribution Networks

Share this