Smart Sensing Technologies for Monitoring and Detecting Leaks in Underground Pipelines A Data-Driven Approach

Undetected leaks in underground pipelines cause significant financial losses, environmental degradation, and safety hazards in modern infrastructure. This study investigates the effectiveness of smart sensing technologies in improving leak detection and monitoring for subsurface pipelines, employing a data-driven framework. Utilizing acoustic, pressure, and temperature sensors, along with real-time analytics, the approach accurately identifies leaks, with the highest detection accuracy achieved by the LSTM neural network at 96.2% and a low false positive rate of 2.5%. Acoustic sensors detected leaks as small as 2.0 mm with calculated sound pressures up to 18.3 Pa, while pressure sensors identified leaks with pressure drops reaching 15.13 Pa for 4.0 mm openings. Temperature sensors measured heat transfer rates up to 1344 J for larger leaks. Machine learning algorithms applied to this sensor data enable predictive maintenance, allowing proactive responses to leak before they worsen. This paper discusses the advantages and limitations of these technologies, emphasizing their potential to enhance the reliability and sustainability of pipeline networks. Findings underscore the value of data-driven smart sensing solutions in managing pipeline integrity, offering a forward-looking strategy for resilient urban infrastructure.