Pressure Management and Machine Learning Application for Leakage Reduction in Urban Water Distribution Networks

IAHR Document Library

« Back to Library Homepage « Proceedings of the 5th IAHR Young Professionals Congress

Pressure Management and Machine Learning Application for Leakage Reduction in Urban Water Distribution Networks

Download

Author(s): Amir Noori; Ehsan Roshani; Hossein Bonakdari

Linked Author(s):

Keywords: EPANET; Water Distribution Network; Pressure Management; Leakage; Machine Learning

Abstract: This research uses EPANET 2.2 to simulate water distribution networks and assess the impact of pressure reduction on leakage rates. Machine Learning (ML) models, including Artificial Neural Networks (ANN), Support Vector Machines (SVM), and Extreme Learning Machines (ELM), are applied to predict leakage in a real-world WDN case study. Developed with nodal pressure management, these models optimize head loss and velocity from hydraulic simulations. Findings reveal that nodal pressure management reduces leakage by over 35% while ensuring uniform pressure distribution. Among the ML models, ELM proves to be the most effective for predicting pressure in WDS with minimal error and high accuracy.

DOI:

Year: 2024