EchoUrbanNet objectives
In France, twenty to fifty percent of drinking water is lost within urban Water Distribution Networks (WDN) distributions from leakages. Several techniques already exist for detecting leakages within WDN. Passive acoustic or hydro-acoustic techniques have focused on the study of noise induced by the presence of an anomaly from analysing the resulting noise emitted by the source. Nevertheless, these techniques are costly and of limited area in WDN. Other techniques based on recording transient water-hammer waves modifications resulting from the presence of leakages have been proposed. Even promising because they are not limited to specific configurations and less sensitive to noise, these methods suffer from several bottlenecks :
i. Transient analysis computations are limited to few hundred pipes units only with classical methods (e.g. method of characteristics, finite volume);
ii. The high-frequency pressure sensors numbers to deploy within a WDN is too large and non- affordable in practice;
iii. The WDN model quality albeit good is nevertheless uncertain so that many parameters might have to be considered in addition to possible fault/leakages locations.
The goal of this project is to overcome these three bottlenecks from proposing new methods/approaches in order to improve WDN knowledge, leak detection and analysing tools and reliability.
News
Previous papers
A. Bayle, F. Plouraboué, Eur. J. Mech. – B/Fluids, 2023, 101, pp.59-88.
A. Bayle, F. Plouraboué, Wave Motion, 116, 103081, 2022.
A. Bayle, F. Rein, F. Plouraboué, J. Sound and Vib., 562, 117824, 2023.
A. Bayle, F. Plouraboué, J. Hyd. Eng, 150, 2, 04023062, 2024.
K. Bertling, J. Perchoux, et al., Opt. Expr., 22, 24, 30356-30356, 2014.
R. Guibert, A. Bayle, F. Plouraboué, 230, 119538, Water Research, 2022.
S. Maquéda, J. Perchoux, C. Tronche, et al., Sensors, 23(7), 3720
O. Piller, G. Gancel and M. Propato, Exeter, GBR, pp. 263-268, 2005.
F. Plouraboué, P. Uszes & R. Guibert, IEEE Trans. Netw. Sci. Eng, 9, 3, 1437-1450, 2022.
F. Plouraboué, Eur. J. .Mech. B/Fluids, 108, 237-271 ,2024.
F. Plouraboué, Phys. Rev. E, 109, 054310,2024.
D. Steffelbauer, et al., J. Water Resour. Plan., 148(3), 2022.
Methods
IMFT partner : The development of new approaches for fast-transient dynamic in WDN will be pursued based upon generalizing recent studies from (Bayle & Plouraboué 2022, 2023). More precisely, the wave dynamics solution can be decomposed into the eigenfunctions base (Cf left figure with a 3-star graph illustration) of the continuous Laplacian operator of the metric’s graph called the »quantum-graph » operator (Plouraboué 2024). This spectral solution can be quasi-analytically computed, over the graphs nodes, without the need for discretizing the graph links in space, hence saving a huge numerical cost for direct transient analysis. One objective of the project is to evaluate this spectral solution over huge networks. From this direct approach, several strategies can be proposed to infer the leakage location from the transient solutions.
INRAE partner : Furthermore, for slow transient dynamics, the rigid water collumn model will be considered within the same feature permiting to avoid the networks link’s discretization. Sensitivity equations for optimal sensor placement will be used and solved.
The sensors developed in LAAS is based on the laser feedback interferometry (LFI) principle (Taimre et al., 2015) and uses the change in the refractive index of transparent media due to the pressure variation to observe pressure waves (Bertling et al., 2014).Thanks to the extreme simplicity of their optical arrangement and to the compactness associated to micro-electronics technologies, the LFI sensors are easy to embed in harsh environment such as water pipes while their unit cost can level-down to a few tens of euros. Recent work for the study of shock waves have demonstrated the ability of LFI sensors to observe extremely fast pressure blast in air (Maqueda et al., 2023) in the ten MHz range. In the frame of the EchoUrbanNet project major challenges as regarding the LFI sensor are the improvement of the acquisition electronics that will lead to frequencies up to several GHz, the design of waterproof and size reduced sensing system and a proper modelling of the acousto-optic effect induced by the water-hammer wave.
The water network of the International Office of Water (OIEAU) illustrated here has been designed for the training of leak detection operators. Each leak is bypassed from an intact pipe, with a set of 53 valves allowing for the sectioning or connection of any part of the network. Each material departure point isequipped with a flowmeter, and 2 air valves also provide access to the pipeline through a manhole. Several simulated house connections allow for simultaneous consumption, with one including a leak. A looped area contains also the three materials, with each part of the loop having a calibrated leak of 2 mm in diameter. 4 fire hydrants are also installed.
This network provides an ideal playground for testing any type of equipment in real-world conditions. We plan to install 2-3 sensors in the manholes, with their positions evolving according to the needs. The public network inlet is in a large chamber with two flowmeters and a pressure reducing valve, allowing for a 1 to 4 bars test pressure range. ETTIS can also supply the network with up to 8 bars pressure using its mobile leak quantification van trailer (1 m³ onboard with pressure control and measurement system). ETTIS will provide some tests on their network bed and in OIEau to validate the developed
sensors. This will allow a significant validation component at low cost. The high-frequency sensor conceived by LAAS will be deployed, tested and validated in ETTIS platform.
Results
The project kick-off is 01/02/2025