To create a diagnostic platform built on the understanding of the propagation and reflection of low and high frequency wave transmission in pipeline systems and their associated boundaries; technologies for defect detection, device and pipe characterization, early diagnosis and warning; algorithms for processing wave signals in the for acoustic communication; comprehensive water pipe acoustic channel models with coupled communications and a world leading test bed to prove and showcase the novel predictive methods and technologies.

Task 1: Creating an experimental test bed for Smart UWSS research

Leader(s): Lee PJ, Lee JHW, Brunone B, Murch RD

Member(s): Karney BW, Zhang X, Meniconi S, Bermak A, Ghidaoui MS

Motivation: The aim of this task is to develop a comprehensive HK based laboratory and field validation program. We have secured support from the HK WSD to use an existing DMA in Quarry Bay, HK, as a field test bed. Key research issues related to the use of both LFW and HFW hydraulic transients for fault detection will be investigated in targeted experiments. Our vision is to develop a hitherto unavailable in-depth understanding of hydraulic transients in the HK water supply system, and to create new paradigms for the detection and identification of faults and pipe conditions.

Objective(s): The objectives are to: (i) develop the first Smart UWSS pipe network test bed in HK; (ii) study the use of LFW and HFW for active generation of controlled transients and identification of system defects; (iii) calibrate, verify and refine the models developed for LFW (1D & quasi-2D) and HFW (2D & 3D) wave propagation, defect detection, device characterization and air-water interaction in Tasks 2, 3, 4; (iv) test the signal design, detection and processing strategies in Task 3; (v) transfer cutting-edge technology on pipeline condition diagnostics to HK.

(Left) Field test in Ngau Tau Kok, Hong Kong; (right) Proposed field test bed in Beacon Hill training facility

Task 2: Physics of Waves in Pipe Systems

Leader(s): Ghidaoui MS, Karney BW

Member(s): Xu K, Lee PJ, Brunone B, Duan HF, Murch RD, Zou J, Zhang X, Dimitrakopoulos I, Youcef-Toumi K

Motivation: The proposed research seeks to create a diagnostic framework where LFW provide the initial system reconnaissance (i.e., to identify problematic sections or zones within the water network). HFW are then applied as needed to zoom into regions and devices to produce high resolution images that can identify anomalies and infer system state. A necessary requirement for this inverse methodology is to develop accurate and efficient models and algorithms of the probing waves which entail better understanding of wave physics in pipes and development of robust solution techniques.

Objective(s): To develop, test, refine and prove mathematical models for LFW and HFW that can be used to provide images of various resolutions of pipe systems and devices in real-time.

High Frequency Wave (HFW) propagation in a pipe with a junction (adopted from Louati 2016)

Task 3: Acoustic Signal Processing for Water Columns

Leader(s): Murch RD

Members(s): Zheng YR, McKay MR, Palomar DP, Lee PJ, Yang TC, Ghidaoui MS

Motivation: Signal processing provides a framework for the estimation of essential parameters needed for the detection of blockages, leaks and other forms of pipe deterioration in the presence of interference and noise. For example signal processing provides estimates of resonant frequencies and transfer functions that are directly related to techniques for detecting pipe defects in Task 4. Signal processing also provides methods for increasing the accuracy of those estimates in the presence of interference/noise and when limits on the probing wave amplitude are imposed (Task 2). Furthermore, if probing wave signals with lower amplitudes than currently required can be designed, whilst maintaining parameter estimation accuracy, then acoustic transducers could be considered for generating the probing wave signals instead of valves. This will vastly increase the signal processing and bandwidth possible and hence the spatial resolution of the analyses. Signal processing can also lead to novel approaches for signal communication along the water column and reduce the number of external measurement access points required in the Smart UWSS

Objective(s): To (i) develop channel signal models for water pipes that incorporate the channel, noise and interference; (ii) use these models to develop advanced signal processing techniques to optimize the probing wave signal while minimizing its transmission power; (iii) develop approaches for estimating key parameters needed in detecting pipe defects and enhance their accuracy and; (iv) use signal processing models to devise novel underwater communication techniques.

Analysis of in-pipe pressure signals

Task 4: Defect Detection & Device Characterization by LFW and HFW

Leader(s):Ghidaoui MS

Member(s): Zou J, Duan HF, Brunone B, Meniconi S, Xu K, Lee PJ, Palomar DP, McKay MR, Karney BW, Katafygiotis LS

Motivation: Task 2 asserts that leaks, blockages, bursts, deteriorated pipes, and malfunctioning devices (pumps, valves) are not only ubiquitous and harm system operation, but distort wave signals. This task specifically seeks to use measured wave signals in an inverse sense to deduce the system characteristics that caused the measured response. The key motivation is to develop algorithms actually able to de-convolute measured signals into a physical understanding of the system – that is, into the constituents responsible for causing the observed response. Expertise of team members at recognizing such signatures in LFW and using them for defect detection is foundational for this task.

Objective(s): To use the knowledge, facilities and models developed in Tasks 1, 2 and signal processing methodologies in Task 3 to create long-range, non-intrusive and reliable techniques to diagnose pipe systems on the basis of measured signals of LFW and HFW.

In-pipe blockage detection by inverse analysis