Flemish water services company De Watergroep provides clean tap water to clients through a 34,000-kilometer water main network. To meet its customers’ needs and tackle environmental challenges, the company is exploring the possibilities of innovation and digitalization. One of the most pressing issues today is the reduction of loss through leaks. The solution? An AI-powered system that enables quick detection and intervention.
“Leaks are a very real problem for every water supply company,” explains Cristina Orodel, team manager of asset information management at De Watergroep. “Around 20% of the water we transport is not consumed by end users. While this includes water used for pipe maintenance and to extinguish fires, the majority of the water is lost through leakages. Our aim was to significantly reduce this amount by implementing smart technology.”
The main cause of leaks is outdated infrastructure. Cristina: “Our network includes pipes that are more than 40 years old, and it’s impossible to renew everything at once. Depending on priority and planning, we replace around 1.5% of our piping per year. This means that there is still a lot of possibility for leaks to occur. In many cases, finding and fixing a leak can take several weeks, especially if they occur in less obvious places like under a river.”
Because reducing leakage is one of the strategic pillars of De Watergroep, the company decided to equip its network with flow sensors. These sensors send information about the waterflow to the data warehouse. Up until today, the interpretation of that data was performed manually.
“Our long-term relationship with delaware allows us to challenge each other and push our boundaries.”
Cristina Orodel, team manager of asset information management at De Watergroep.
“Our goal was to automate that data interpretation process to allow for rapid detection and intervention,” says Cristina. “For this purpose, we assembled a hybrid team of leak experts, internal IT people, and experts from delaware, with whom we’ve been working closely for the last 15 years. This long-term relationship allows us to challenge each other and push our boundaries. After a preliminary kick-off meeting, delaware helped us set up a clear roadmap for the project, with weekly meetings.”
The new solution combines historical data, data analytics and key algorithms to quickly detect anomalies in the waterflow. Leaks are now discovered up to two weeks faster, after which the right people are notified at the appropriate time. They take the necessary steps to reduce water waste and end the leakage, which has immediate benefits, both economically and ecologically. The solution runs on SAP Leonardo as a Proof of Concept, this platform makes it easy to integrate new technologies into De Watergroep. Cristina concludes: “While leakage was our primary concern, the data offers many more valuable insights to our business. In the near future, for example, it could allow us to warn our clients proactively when a leak occurs in their home, potentially saving them a lot of money. We’re also looking for ways to open up the data for creative start-ups to come up with new ideas.”
Leak detection becomes child's play with the AI-powered algorithm that delaware developed for De Watergroep. Just have a look at this fantastic video that shows how children would solve the water leakage problem.