Research | TR Imaging

Background

Our recent TRS project—Smart UWSS (Urban Water Supply Systems)—pioneered the use of classical time-reversal (TR) for high-resolution diagnosis of fully-fledged defects. RGC’s Monitoring and Assessment panel rated the progress “exemplary” and the achievements “exceptional”. This project moves the research to a proactive mode, where millimetre-scale anomalies are detected and imaged using centimetre-long and centimetres-amplitude waves over a 100–300 m range. Theoretical barriers confronting us are twofold:

The wave-flow physics governing millimetre-scale imaging belong to a nonlinear regime where classical TR is invalid.
The theoretical foundation for separating small amplitude signals scattered by anomalies from those scattered by flow and discontinuities in a topologically complex, dynamic, uncertain, and lossy UWSS waveguide, does not exist.

Adding measurement noise and modelling uncertainty make subsequent development of stable high-resolution images a challenging ill-posed problem.

We will study processes violating classical TR and the imaging degradation they induce, develop theories and methodologies to counter or cancel such degradation, and explore novel non-reciprocal TR methodologies. We will examine signatures of small defects and anomalies in measured signals, design stable and high-resolution TR imaging for incipient defects, explore the use of metamaterial concepts and develop physics-informed machine-learning techniques for long-range high-precision imaging.

Novel physical insights, imaging systems, better models, and refined designs arising from this research will improve all stages of UWSS system management—from better first design to superior performance and longer system life—achievements that will help HK meet 2030 water-loss-reduction and 2050 net-zero targets, and develop environmental services into a pillar global industry (~HK$10B and growing). The three tasks of the project are:

Methodology

Task 1

Develop HKUST-GZ and Q-leak facilities
Enable the experimental step of imaging
Test and refine theories developed in Tasks 2 & 3
Establish baseline images of the new HKUST-GZ and Q-leak systems needed for TR imaging research (Task 3)

Task 2

Study processes that violate TR and how they interact with waves and degrade the fidelity of TR focusing
Develop theories and methodologies to counter or cancel such degradations
Develop and explore the non-reciprocal TR approach

Task 3

Investigate the signatures of incipient defects and anomalies
Design optimal (stable and robust) TR Imaging for incipient defects
Develop physics-informed machine-learning techniques to enhance TR imaging
Explore metamaterial concepts for super-resolution & long-range imaging

The research and system levels – Realization of the Time Reversal Imaging for Sustainable & Smart Urban Water Supply Systems vision.

We are transforming the new 6.5 km potable water supply network at HKUST’s Guangzhou Campus into the world’s first fully accessible living-lab for UWSS imaging. Similar transformations are planned for the newly opened and well-instrumented Q-leak system owned by our partner, the Water Supplies Department (WSD). We will leverage not only these new testbeds but also the experience, facilities, knowledge, collaboration, and industry partnerships developed in the previous TRS along with the expertise of our internationally recognized, cross-disciplinary research team, now extended to include pioneers in TR, metamaterials, super-resolution, artificial intelligence, in-pipe sensor and robotics technologies, and inverse mathematics as well as TR-imaging inventors.

Figure 2: An illustration of fully automated HF-TR imaging system

The HKUST-GZ living lab will be fully equipped to enable a fully automated TR-imaging technology that will provide a comprehensive 3D virtual image of an underground UWSS. The automated, distributed, cloud-based monitoring and diagnostic system—an evolution of SCADA systems—will comprise of multiple high-sampling rate (2 M Samples/s) data-acquisition systems, transducers, and control devices deployed at each access point of the HKUST-GZ water distribution network, and will be accessed and controlled remotely via an interactive web-application. The installed devices will perform synchronized measurements between access points with previously unattainable accuracy (microseconds) thanks to integrated GPS technologies. Data will be uploaded at high rates to a cloud server via 5G connections. The system will also remotely control robotic devices, wave generators, high-frequency transducers, and hydrophones. Hence, it will be able to perform diagnosis and imaging tests on demand. The data obtained will be stored in the cloud and processed in real time with our novel TR imaging algorithms to evaluate the integrity of the water network. Potential defects and pipe wall deterioration will be shown graphically on a map of the underground UWSS in the form of a comprehensive 3D virtual image. The designed diagnostic and imaging system will be inherently scalable; thus, it will be possible to add multiple measurement and control stations according to our needs and the spatial extent of the water distribution network.