SenSiteUQ: digiLab’s Machine Learning Sensor Placement Capability is Transforming the Water Sector

digiLab has developed SenSiteUQ, a sensor placement optimiser in the Uncertainty Engine platform designed to tackle one of the most pressing environmental challenges: wastewater management. Effective monitoring of wastewater—containing pollutants such as human waste, chemicals, and food scraps—is critical for water suppliers to ensure and maintain water quality and to reduce the frequency of pollution events.

Recognising this need, UK water suppliers partnered with digiLab to leverage machine learning for optimising sensor placement in sewer networks. The goal is to improve the efficiency of sewage level measurements and recommend optimal water sensor placements. To support this innovation, OfWat awarded £436,000 in funding to help develop and commercialise the technology - Yorkshire Water served as the pilot partner for the project.

Looking ahead, UK water companies plan to deploy over 300,000 sensors by 2030. These sensors will transform wastewater management by forecasting and preventing combined sewer overflow (CSO) spills, identifying blockages, pinpointing sources of contamination, and enabling predictive maintenance. With SenSiteUQ, digiLab is at the forefront of this transformation, ensuring a smarter and more sustainable approach to wastewater monitoring.

What is the key challenge that the water sector is facing?

The biggest challenge in wastewater management is placing sensors efficiently—not just for coverage, but for cost-effectiveness and reliability. Existing software often focuses solely on network coverage, overlooking critical factors like risk, maintenance, and deployment costs. This does not give all the relevant information needed to effectively measure sewage levels. digiLab provides the solution by offering a scientific approach that ensures that each sensor is placed in a location from which it will contribute the most valuable information possible, to reduce uncertainty in the network.

How is digiLab utilising probabilistic machine learning to solve this issue?

SenSiteUQ offers a simple and innovative solution. It is a sensor placement software framework, which employs a rigorous, scientific approach to ensure that each sensor is placed to contribute the maximum possible information for the parameters selected.

SenSiteUQ utilises APIs from digiLab's probabilistic ML platform - the Uncertainty Engine - in order to reduce the number of sensors required. It optimises sensor placement in water networks to maximise information about complex water systems and provide users with confidence in their risk exposure. Using machine learning, it allows rapid and accurate sensor deployment, providing value for the expected installation of 300,000+ sensors for the 2030 plan.

digiLab is currently advising Yorkshire Water on optimal sensor placement for wastewater networks in Bradford, UK. The sensor placements were determined using senSiteUQ’s unique drainage network analysis techniques, prioritising network nodes with the highest overall importance to the drainage network’s hydraulic dynamics. The presented sensor placement layout represents the most information-efficient locations to place sensors – i.e. the sensor locations that provide the most information about the state of the entire urban drainage network – while considering the practicality of placing and maintaining the sensors. digiLab’s convenience factor uses multiple layers of geospatial data to identify high-access deployment locations, minimising failed placements.

Tom Ogden, Innovation Technical Specialist at Yorkshire Water says: "It has been a pleasure to work with digilab to trial the SenSiteUQ software to develop our knowledge on optimum placement of sensors on our wastewater network, so that we can prevent pollution incidents before they happen and provide an enhanced wastewater network for our customers”.

Key Values of SenSiteUQ

1. Cost Reduction: reduces cost by avoiding placing unnecessary sensors

2. Risk Reduction: reduces risk and increases confidence by ensuring optimum sensor coverage in a network, thereby improving observability along the network

3. Faster Deployment: speeds up sensor deployment, thereby reducing time and cost, by minimising traffic disruption

4. Reduced Maintanence: reduces time and cost of sensor maintenance by taking into account mobile network signals and therefore preserving sensor battery life

About digiLab

Having grown from a small team of expert mathematicians and data scientists, digiLab is now a leading AI company that empowers organisations in highly regulated or safety-critical industries to solve their complex engineering, infrastructure or data challenges.

As world leaders in uncertainty quantification, based on years of cutting-edge academic research, our team of 30+ experts specialise in areas with uncertain or sparse data, teaching teams how to solve their grand challenges using digiLab’s platform.

digiLab’s platform is a no-code agentic AI platform that combines data, models and workflows to connect a trustworthy AI digital thread through an organisation. digiLab enables you and your teams to solve complex problems in a secure and auditable platform to build decision intelligence and shorten time to value.