WESSEX WATER TO DEPLOY AI ACROSS WASTEWATER NETWORK

UK WATER COMPANY WESSEX WATER is joining forces with artificial intelligence (AI) company StormHarvester, to expand the use of AI to detect blockages across its entire wastewater network.

The deployment follows a successful trial in May 2020 and is said to be the most extensive to be rolled out in any wastewater network in the world.

Following the trial, Wessex Water has confirmed it will deploy StormHarvester’s AI technology across its entire network over the next three years, as part of its commitment to continue providing high standards of sewerage services to its customers and the environment.

The expansion will cover nearly 35,000 km of sewers and wastewater generated by 2.8 million people.

“We are very pleased to win the network rollout,” said Brian Moloney, Managing Director at StormHarvester. “This is the biggest commitment ever to deploy AI in wastewater networks, not just in the UK but also globally.

The trial in the city of Bath tested the scope of AI to see whether it was possible to use machine learning to identify early-forming sewer blockages, mute unnecessary control room alarms and establish an operational basis for a shift towards condition-based maintenance. During the trial, StormHarvester’s Intelligent Sewer Suite detected over 60 early blockage formations in real-time.

“Our pilot showed 92% accuracy in identifying early forming blockages with zero missed and control room alarm rationalisation of 97%. Our technology also identified at least two incidents that we are fairly confident would have resulted in Category 3 spillages or worse. Numbers like these make for a strong business case for utilities,” said Brian.

Sewer blockages can lead to costly service failures, including pollution and flooding events, but they can be quickly remedied if spotted early enough. Wet weather makes it difficult to differentiate expected high sewer levels caused by heavy rainfall volumes from those arising from restrictions, such as partial or total blockages.

 By deploying AI with the capacity to differentiate between these different events, both an improvement in alarm quality and alarm rationalisation is possible.