The
DVV, Duisburger Versorgungs- und
Verkehrsgesellschaft mbH,
group includes over 30 companies providing services in the areas of energy,
transportation, technology and telecommunications. Netze
Duisburg GmbH, one
of the group companies, is responsible for gas, water, and district heating
networks in the city of Duisburg.
Before
implementing SIWA Leak Finder, the Duisburg water distribution network was
divided in two large areas: North (658 km of pipeline) and South (562
km).
Due
to the large size of each of the two areas, it was very difficult to locate
leaks within the network. Although large areas make detecting and pre-locating
leaks a challenge, they also offer advantages, including greater network
resilience, reduced risk of running out of water for fire fighting and lower
risk of water quality problems.
Netze
Duisburg wanted a system that would allow them to detect and pre-locate leaks
without any need to close valves to reduce the size of the areas.
Virtual
DMAs
(virtual district metered areas) combine Artificial Intelligence with hydraulic
simulation and
permit the creation of small DMAs without compromising network performance. The
concept of virtual DMAs was therefore very suitable for the city of
Duisburg.
SIWA Leak Finder software was implemented throughout the entire network (1,220 km).
German engineering company RBS Wave created the hydraulic model and established the positions where flow meters should be installed to create both virtual and physical DMAs. Nine physical DMAs were created, the smallest with a pipe length of 11 km and the largest with a pipe length of 562 km.
108 flow meters were installed.
Additionally, 83 smart meters were installed on the premises of strategic customers, allowing the real time calibration of the hydraulic model by the BuntBrain LeakFinder software.
The result was an enormous reduction in the search area.
Subsequently, network monitoring has continued to be improved during the last months with the installation of 15 new flowmeters and 17 smart meters. This has improved the accuracy of leak positioning and the leak detectability.
The
main benefit which the solution has provided has been the ability to accurately
detect and pre-locate leaks, reducing the search area required to find a leak
by over 10 times without compromising the performance of the distribution
network, which has remained open and resilient.
Since
the installation of SIWA Leak Finder, 490 leak events have been detected and
pre-located. The event detection accuracy rate was 80% in 2023 and accuracy is
constantly improving thanks to the self-learning algorithms.
56
leaks have been detected in the largest physical DMA (562 km), which is
equipped with 38 flow meters and 29 smart meters. By the end of 2023, the
delivered water had decreased by 9% compared to the level in the previous year,
going from 3.24 m3/(km*h) in 2022 to 3.16 m3/(km*h) in 2023. For DMA with 562
km of pipe network, this means a reduction of 393.850 m3 a year.
The
results demonstrate the potential for the SIWA Leak Finder software to make an
extremely valuable contribution to reducing water losses in large network
areas.
Netze Duisburg - AI-based leak detection in water distribution network
Fast and accurate leak detection with SIWA Leak Finder
Netze Duisburg - AI-based leak detection in water distribution network
Fast and accurate leak detection with SIWA Leak Finder
Completion: ongoing
Contact
Contact our sales for more information and to discuss directly how our products and solutions could help you to best solve your problem.
Related Links
Please note that some of the links may be only available in Intranet.
