Across Europe, more than 30,000 wastewater treatment plants work around the clock to turn sewage into clean water. However, climate change, increasing pollution and stricter environmental regulations are making this task increasingly complex. Heavy rainfall, droughts and emerging contaminants require operators to constantly adapt. At the same time, modern treatment plants generate enormous amounts of data that are impossible to analyse manually in real time. Could AI help plant operators overcome these challenges?
This question has been at the heart of the DARROW project. Over the past four years, researchers, engineers, technology developers and wastewater utilities have worked together to explore how artificial intelligence can enable smart wastewater treatment through better data, smarter decision-making and more adaptive process control.
Towards smart wastewater treatment
DARROW developed a modular platform of AI tools that help wastewater treatment plants make better use of their operational data. The DARROW platform combines technologies for data quality improvement, process modelling, decision support and intelligent process control. Together, these tools enable plant operators to move from reactive operation towards more predictive and adaptive plant management.
At the Tilburg wastewater resource recovery facility in the Netherlands, project partners tested DARROW’s tools under real operating conditions.
“We have a wastewater treatment plant with an energy factory and resource recovery factory. There’s lots of process units, and they all need to be coordinated as best as we can,” explains Ruud Peeters, Senior Policy & Innovation Adviser at Water Board De Dommel.
Modern treatment plants like the one in Tilburg rely on hundreds of sensors measuring parameters such as pH, ammonium, nitrate and temperature. As a result, they generate continuous streams of data. DARROW investigated how artificial intelligence can transform this information into practical support for operators. The goal is to predict plant behaviour, detect anomalies and recommend operational settings.
“AI can check the system continuously […] and make small changes continuously,” says Bart Rommens, Process Operator at the Tilburg plant.
Trust remains essential
One of DARROW’s key findings is that successful AI adoption depends on transparency and trust. Rather than replacing operators, AI should complement their expertise by providing reliable recommendations while leaving final decisions in human hands.
“We need to see that the system makes sensible decisions,” says Peeters. “The operators always need to be involved because there are always exceptions. You can’t leave that to an AI.”
Looking back on DARROW
As the DARROW project comes to an end, these insights and results come together in the project’s final video “AI in the Water – How Artificial Intelligence is Changing Wastewater Treatment”. Filmed at the Tilburg wastewater resource recovery facility, the video follows researchers and plant operators. They show how artificial intelligence is being developed, tested and evaluated in a real wastewater treatment plant.
The film offers a behind-the-scenes look at the collaboration that has shaped DARROW. It also highlights the opportunities and challenges of bringing AI into one of society’s most essential infrastructures, advancing the transition towards smart wastewater treatment.