ISSN 2756-3251
Advanced Journal of Environmental Science and Technology | Vol. 17 (6), pp. 672-684, June, 2026
Research Article
Acoustic Tomography and Machine Learning Detect Subsurface Water Leaks in Aging Urban Distribution Networks to Reduce Non-Revenue Water
Kofi Asante¹*, Beatriz Silva-Santos², Jonas Lindstrom³
¹Department of Civil Engineering, Kwame Nkrumah University of Science and Technology, Kumasi, Ghana
²Department of Hydraulic and Sanitary Engineering, Federal University of Minas Gerais, Belo Horizonte, Brazil
³Department of Water Resources Engineering, Lund University, Lund, Sweden
Corresponding Author: Kofi Asante — [email protected]
Abstract
Aging pressurized water distribution pipelines lose up to 40% of treated freshwater through undetected background leaks, exacerbating urban water scarcity and pumping energy consumption. We deployed hydrophone sensor arrays along 42 kilometers of ductile-iron distribution mains and trained a gradient-boosted decision tree classifier on cross-spectral acoustic features. The system localized pinhole leaks as small as 0.15 liters per second within a 1.8-meter margin of error while filtering out ambient vehicular traffic noise. Field repair of 29 identified leaks saved an estimated 1.1 million cubic meters of treated water annually.
Keywords: acoustic leak detection; non-revenue water; urban water infrastructure; machine learning; hydrophone sensors; water conservation
Manuscript Timeline: Received 15 May, 2026; Revised 12 June, 2026; Accepted 23 June, 2026; Published 30 June, 2026.
Citation: Asante, K., Silva-Santos, B., and Lindstrom, J. (2026). Acoustic Tomography and Machine Learning Detect Subsurface Water Leaks in Aging Urban Distribution Networks to Reduce Non-Revenue Water. Advanced Journal of Environmental Science and Technology, 17(6), pp. 672-684.
Subscribe to read the full article: https://www.internationalscholarsjournals.org/subscribe-to-read