IoT and Data-Driven Acoustic Leak Detection Using Microcontroller and Fast Fourier Transform

Ensuring access to clean and safe drinking water is essential for public health, environmental sustainability, and the functioning of modern society. However, water leakage in municipal pipeline networks continues to be a persistent and costly issue worldwide. This research explores both established and emerging technological approaches to mitigating water loss, with a specific focus on the water distribution infrastructure of Gjovik Municipality in Norway. The study combines theoretical analysis with the development of a functional prototype designed to detect pipeline leaks using acoustic monitoring techniques. The prototype employs an ESP32 microcontroller and a digital micro-electro-mechanical system-based acoustic sensor, utilizing Fast Fourier Transform to analyze sound signals associated with potential leak events. Controlled laboratory tests demonstrated the system's ability to detect anomalies within defined frequency ranges indicative of leakage. While promising observations have been noticed, the prototype has limitations - particularly in its sensitivity to environmental noise and its untested performance in field conditions. Nevertheless, the findings highlight the potential of a low-cost, modular system to serve as a complementary tool alongside professional-grade municipal solution. With future enhancements - such as integration of wireless communication, machine learning algorithms for signal classification, and rigorous field deployment - this approach could significantly contribute to more efficient and sustainable water management practices.