2023-24-project-catalogue

###Graph-based methodologies for water networks monitoring and control

Project ID: 2228bd1083 (You will need this ID for your application)

UCL Lead department: Electronic and Electrical Engineering (EEE)

Project Summary:

Why this research is important

Water transmission pipelines lose 25% of the water transmitted through them, on average. Since water treatment and distribution contribute to 4% of global energy consumption, lost water represents an impressive economic and environmental cost. Moreover, there are concerns that leakages might affect water quality. It is therefore imperative to prevent the threat of leaks and minimize their damages, by promptly detecting and locating them. This project represents a timely opportunity as droughts have become a reality in the UK too, while taking advantage of the development of IoT technologies.

Who you will be working with

The project is interdisciplinary and will include working with the teams of the supervisors in the Control Engineering and Machine Learning areas, as well as collaboration with existing research and industrial partners, both at UCL (ICCS and CEGE), in the UK (MorrisonWS) and internationally (KIOS, Cyprus and Veitur, Iceland).

What you will be doing

Project’s objectives are:

to model water networks and water quality based on data,
to design novel methodologies to detect and locate water leakages,
to optimise sensors and actuators deployment on the networks,
to design novel distributed methods for water quality monitoring and control,
to test and validate the developed methodologies on testbeds and real-world data, provided by collaborators. A major challenge for leak detection is the lack and sparsity of measurements, as currently sensors are placed only at some points of the network and meter readings are available only in some cases. The candidate will investigate the opportunity to integrate graph-based learning techniques with model-based control techniques to overcome this challenge.

Who we are looking for

A motivated candidate, loving interdisciplinary research and collaborations, with excellent programming skills (Matlab/Python), great knowledge of control theory and principles of signal processing/machine learning.