Proper maintenance of underground infrastructure is crucial for a city's sustainable development. However, with its high-density underground utilities, such maintenance work is particularly challenging in Hong Kong. A research team from The Hong Kong Polytechnic University (PolyU) has leveraged advanced underground exploration technologies to develop underground utilities inspection systems that support early detection of urban infrastructure anomalies, including voids and pipe leakages, for enhanced urban management.
Underground utilities are essential for providing water, energy and communication services. As the infrastructure ages and deteriorates, it becomes prone to cracks, leakages and even road subsidence, leading to service disruptions and road accidents. Developed by Prof. Wallace Wai Lok LAI, Associate Head and Professor of the PolyU Department of Land Surveying and Geo-informatics, and his research team, their technologies help accurately pinpoint the source of leakages and indicate their severity through analysis of underground images and leak noises. Addressing the complexity of Hong Kong's underground pipeline network, these technologies can serve as safeguards against related urban risks.
Multi-channel and vehicle-towed GPR technology supports large-scale inspection
In the construction sector, ground-penetrating radar (GPR) technology is often used to investigate underground anomalies by scanning and imaging underground structures. The researchers utilised advanced multi-channel and vehicle-towed GPR that allows large-area scanning. From the images generated of underground pipes, they successfully decoded water leakage signatures in utilities surrounded by soil, and established a set of quantitative benchmarks for determining where there is leakage and assessing how serious it is.
With this technology, researchers can uncover potential underground cavities and pipeline leakages before they actually occur, and examine changes in time-lapse radar data for ongoing detection. One of the critical aspects of the project is the introduction of a unified framework for producing consistent and quantitatively interpretable GPR images. Prof. Lai said, "Traditionally, GPR technology is used for subjective near-surface geophysical mapping and prospecting. Our research presents a significant advancement in using it as an objective measurement and a diagnostic tool to identify and locate hazards, and assess their severity, further advancing the application of GPR."
Another side of the coin: Leak noise analysis also helps locate leakage source
When pipe leakage is detected in a particular region by GPR, it is important to locate the leakage for subsequent repair. Repair work relies on precise positioning for excavation, and this is where another technology comes into play—distinction of leak noise and its positioning. The researchers conducted analysis to understand the characteristics of such sounds for years, specifically examining the amplitude and magnitude of sounds distant from and at the leakage point. They further found that leakage caused by different factors, such as pipe cracks or valve leaks, and on different levels of severity produces noise with different patterns. Supported by these findings, through studying the sound data the researchers are able to discover the source of the leakage and distinguish between different leakage scenarios.
Currently, with the help of ground microphones and leak noise correlators, technicians in the industry collect leak noise at fixed points, including suspected leak points and high-risk locations like areas near valves. These tools are, however, prone to interference from environmental noise like traffic, making it hard to accurately identify the source and condition of the leak in many occasions. The team is now exploring the use of robots equipped with acoustic hydrophones that can go deep into underground pipelines to collect sound data directly for more precise locating of the leak source and arrangement for immediate repair.
Integrating AI and robotics technologies for future application
At the forefront of research on underground pipeline inspection for decades, Prof. Lai's projects have received support from the government and industrial institutions. Among these is the Water Supplies Department (WSD), which collaborated with Prof. Lai's team to launch the underground water mains leak detection training centre, Q-Leak, in 2021 to advance leak detection technology. The two parties earlier signed a Memorandum of Understanding with Shenzhen Bwell Technology Co. Ltd to jointly establish the Pipeline Robots Joint Laboratory, focusing on developing pipeline robotics technologies.
In addition, making use of the GPR images and leak noise previously collected, the research team is working with the Government and industry partners to establish a database and develop an AI model that enables efficient comparison and analysis of substantial underground pipeline images and sound data, while also generating more accurate and reliable assessment results. The team envisions that this initiative will facilitate large-scale inspection of underground pipelines in Hong Kong and beyond.
Prof. Lai remarked, "WSD aims to reduce the rate of water leakage from 13.4% to less than 10% before 2030. Meanwhile, the Highways Department reported 52 cases of road subsidence between 2021 and 2023, many caused by leakage in high-pressure underground water pipelines. By harnessing a range of advanced technologies, we aim to develop a data-driven warning system and surveillance plan, along with a risk-based asset management strategy, for detecting underground leakage and voids with improved accuracy and efficiency, and providing scientific support to relevant policy decisions."
