Machine Learning Methods for Monitoring of Complex Water and Sewer Network Infrastructure 


Funding Body: EPSRC/Thames Water

Student: George Crowley

In recent years, water utility companies have been under increasing pressure from the general public and governmental bodies to address fundamental issues such as combined sewer overflow (CSO) spills into clean water sources and pollution events which essentially place foul water into clean water sources. To try to combat this issue, water utilities have heavily invested in sensor-based technologies to place into various parts of the networks, including sewer networks to proactively monitor and prevent pollution events from happening.  However, a fundamental question faced by the utilities when investing in this technology was: where do we now place these sensors in our networks? For sewer networks, it is widely regarded by the utilities that the use of sensors is there to monitor blockage events caused by issues such as fats, oils, and greases (FOG) and various other objects such as wet wipes.