Artificial Intelligence Pilot Launches at U.K. Water Utility

An artificial intelligence (AI) pilot project recently launched in the United Kingdom to proactively monitor assets at wastewater treatment facilities. Anglian Water, one of the U.K.'s largest water utilities, began the year-long test as a way to predict, detect, and address issues, improving maintenance in their water recycling centers.

They're not alone. Some other early adopters have begun using AI for water treatment, water quality forecasting, and equipment failure prevention, although digitization adoption remains relatively slow in the water sector overall.

With droughts and flood events intensifying globally, the stakes keep getting higher. Here's a closer look at real-world smart tech initiatives to save water, lower costs, protect assets, and complement routine water testing.

Safeguarding Wastewater Treatment

Anglian Water treats water coming from more than 6 million people across the East of England. The utility also provides drinking water to 4.3 million customers. This year, the American nonprofit WaterStart brought Anglian Water together with the Norwegian company InfoTiles, which makes artificial intelligence software that connects to existing supervisory control and data acquisition control (SCADA) systems.

For the 12-month pilot project, UtilityWeek reported that Anglian Water is using InfoTiles software at 30 wastewater treatment sites to pull data from SCADA sensors. The software has real-time analytics and machine learning capabilities that can perform tasks such as monitoring biofilter arm rotation and sludge levels in primary settlement tanks.

This technology enables Anglian Water teams to better visualize equipment reliability, anticipate and head off potential problems, and prioritize asset maintenance, according to the Nordic company.

The water utility has embraced AI for other projects as well, working with Sydney-based Vapar to try out their algorithm on wastewater pipe inspection closed-circuit television footage. Trained to detect at least 2 million different pipe defects, the AI system automatically reviewed 10,000 meters (more than 32,800 feet) of footage in under four hours, the Australian company's case study said.

Another Anglian Water initiative called Safe Smart Systems received £7.5 million ($9.2 million) in funding last year from Ofwat, the water services regulation authority for England and Wales. Led by Anglian Water, the initiative included partners from other U.K. water companies as well as Imperial College of Science, Technology and Medicine, the University of Sheffield, Skanska, and Microsoft.

"Safe Smart Systems will develop a future-oriented, secure and self-regulating artificial intelligence decision engine, which will identify and learn the consequences of failure and then trigger proactive interventions to optimize the network," the utility's 2022 Innovation Acceleration plan explained.

Advanced Computing for Public Health

Flowing through Maryland to the District of Columbia, the Anacostia River has been through the ecological ringer. Its close proximity to agricultural, industrial, and urban developments led to significant historical pollution including runoff, heavy metals, sewage, and toxins, the U.S. Environmental Protection Agency noted. In early 2020, the nonprofit organization Anacostia Riverkeeper, which works to protect and restore the water, formed a partnership with Boston-based DataRobot through the company's AI for Good program.

Together, Anacostia Riverkeeper and DataRobot worked on a system to collect data from United States Geological Survey sensors along the river, use machine learning to predict E. coli levels, and share that information with the public. WAMU reported at the time that they aimed to augment volunteer water testing.

"This system helps shorten the delay between sample taking and results by making water quality predictions multiple times per day," the DataRobot case study said. "Although it's not a replacement for physical samples, it helps add another layer of information."

Another example is the Terminal Island Water Reclamation plant in Los Angeles, which treats around 15 million gallons of water each day, according to LA Sanitation. The plant has machines that use AI for wastewater treatment, including automatically taking samples, removing pipe buildup, and making adjustments to the purification systems, Mashable reported.

Environmental engineer Lance Thibodeaux told the online news outlet that this technology is critical. "Without computer-powered advanced wastewater treatment, 'the alternative is there won't be enough water when our kids are adults,' Thibodeaux said. 'So we need to look at reusing the water as much as possible.'"

AI Developed to Search for PFAS

Meanwhile, an interdisciplinary team led by the University of Waterloo in Ontario started focusing on per- and polyfluoroalkyl substances (PFAS) in Canadian water systems, CTV News reported in September. The group that's working to identify and treat PFAS, also called forever chemicals, includes artificial intelligence experts. Their plan is to take samples at points along the entire water treatment process and develop machine learning models that can identify the most effective ways to treat PFAS.

Last year, Chinese researchers from Tongji University in Shanghai published a review of AI and machine learning advancements for drinking water treatment in the Chemical Engineering Journal. Their review found that the most widely used AI technologies had clear advantages over mathematical or statistical methods.

"AI technologies have the ability to monitor the evolution of water quality, analyze and predict water quality, and reveal the process of pollutant migration and transformation, thereby shifting the focus from solving existing problems to identifying risks in advance and dynamically optimizing the facilities," the scientists wrote.

Read These Next

Alyssa Danigelis

Alyssa Danigelis is a professional freelance journalist who covers business, sustainability, energy, science, and technology. She received a bachelor’s degree from Mount Holyoke College and a master’s degree from Columbia University’s Graduate School of Journalism. Having grown up in Burlington, Vermont, she spent formative time in Boston and pounded the pavement for years in New York City before moving to sunny Colorado, where she currently resides.