Thanks to a machine learning platform from French data science company Dataiku, the water and sanitation services manager can identify problematic water pipes.
Almost 25% of the drinking water produced is lost in France due to water leaks. “This waste is due to the age of aging pipes,” explains Christophe Orceau, CTO of Kurrant, a consultancy company specializing in the digitalization of cities and public services that supports many communities and water stakeholders. It is in this context that the group Saur, an actor in the management of water and sanitation services in France and internationally, decided in 2021 to develop an algorithm capable of defining a criticality index for drinking water distribution pipes. Its objective: “allows the identification of pipes responsible for water leaks and contributes to improving network performance”, says Grégory Denis, data director at Saur and responsible for developments in artificial intelligence.
To develop this index and determine how AI can add to its skills and services, Saur convened in late 2021 Data, a French company specializing in data science and AI. “The advantage of this partner is to facilitate collaboration between different commercial contacts and allow end-to-end completion of projects, from the exploratory phase to production of the product. algorithms without breaking the software environment. Furthermore, the intellectual property of what is developed remains the exclusive property of Saur”, explains Grégory Denis.
The project was implemented at the beginning of 2022 and was gradually extended to all of France. “The difficulty is often access to quality data and standardized formats. Fortunately, the datahub previously assembled by the Saur group accelerated the project”, testifies Amaury Delplancq, Vice President of Dataiku for Southern Europe. Saur can now consider deployment of the solution on a group scale.
Millions of rows of data examined by AI
To identify water leaks, Dataiku’s supervised machine learning solution analyzes different types of data and intervenes at multiple levels. It is already based on the completeness of data from asset analyzes of the network managed by Saur, which are based on the age of the pipes, the presence of progressive flow opening equipment to avoid hydraulic shocks or data from connected equipment, such as smart meters, especially at night, when flows are expected to be practically zero.
The algorithms also analyze water volume data on distribution, consumption, soil nature and geohazard data to understand soil movements. Satellite data and operational intervention data are also integrated. This represents a total of millions of lines digitized by AI. The network operator is in the process of adding the visualization of sectoral returns to this service thanks to the mass of volumetric information reprocessed by big data to automatically complete missing or aberrant data.
250 thousand linear kilometers of drinking water analyzed
AI is also involved in quickly locating identified leaks to guide the operational agent in the field. Another use of this algorithm: determining the best way to invest in the renewal of hydraulic assets. “The idea is to direct investments towards shelves that improve the network’s performance”, celebrates Grégory Denis. Saur manages 250,000 linear kilometers of drinking water for around 20 million people. AI results integrated into mobile apps show recommendations from Saur teams based on ten years of repair history. The first results made possible by AI: “Identifying the 10% of pipes responsible for a third of leaks saves a lot of time in locating”, celebrates Grégory Denis.
“AI and big data integrate very well into water business processes and add value to hydraulic skills”
AI combined withIoT It thus asserts itself as a true added value in water management. “The technology is obvious to anticipate any problem, such as water scarcity”, says Christophe Orceau, CTO of Kurrant, who is not involved in this Saur group project but has, as an expert, a good view of the market. An opinion shared by Grégory Denis: “AI and big data integrate very well into the processes of water professions and add value to hydraulic skills.”
This leak detection project is not the only one carried out by the Saur group. The drinking water supplier had already developed, through its subsidiary ImaGeau, an AI to predict the level of groundwater and surface water. Currently, teams cross-reference drinking water consumption forecasts with groundwater recharge forecasts. Another example, Saur is working on detecting deviations in tank filling. The next step will be to developGenerative AI for responding to competitions and monitoring contractual commitments in a 100% secure LLM environment.
Through usage linked to network performance, Saur demonstrates the usefulness of AI in its business. And there is still much to be done across the water cycle. “After groundwater, the second step is to monitor pumping, an activity that consumes a lot of energy and can be optimized. And after detecting leaks in the distribution network, AI and IoT can intervene in the treatment of used water and its recycling”, lists Christophe Orceau. The group has expertise, through its subsidiary Sterau, in engineering for the design and construction of wastewater treatment plants. Further expand the uses of AI, an idea that comes naturally.