AI-Driven IOT-Based Underground Water Pipeline Leak Detection Using Acoustic Signal Analysis

Abstract - Water pipeline leakage results in significant water loss, infrastructure damage, and economic burden. Traditional leak detection techniques rely on manual inspection or expensive monitoring equipment, limiting scalability and real-time responsiveness. This paper presents an AI-driven IoT-based underground pipeline leak detection system that integrates acoustic sensing, environmental compensation, and deep learning-based classification. The system utilizes an ESP32 microcontroller for edge data acquisition, combined with acoustic, temperature, and wet sensors. Extracted acoustic features are processed using a one-dimensional DenseNet deep learning model to classify normal flow and leak conditions. The system transmits processed results to a cloud-based IoT dashboard for real- time monitoring and predictive analytics. Experimental validation demonstrates improved detection accuracy, reduced false alarms, and scalable deployment capability for municipal and industrial water distribution systems. Keywords: Leak detection, IoT, Deep learning, DenseNet, Acoustic sensing, Edge AI, Smart water management