About the project
This PhD project will develop advanced ultrasonic array techniques for hydrogen leak detection, localisation and characterisation in complex, noisy environments. Combining mathematical modelling and physics-informed signal processing with AI-driven methods (including PINNs), the research aims to enhance robust, cost-effective leak detection with industrial applications on complex sites.
This PhD project offers the opportunity to work at the cutting edge of signal processing, ultrasonic sensing, and AI-driven leak detection—an area of increasing importance for hydrogen safety and clean energy transition.
With part-funding from Shell, this research has strong industrial relevance, ensuring real-world impact:
- you will develop expertise in cutting-edge mathematical foundations of Deep Learning and AI;
- you will learn ultrasonics and array signal processing, studying how leak-generated signals propagate in complex environments to improve the localisation and characterisation of gas leaks;
- you will integrate fundamental wave physics modelling with advanced data-driven methods to enhance detection accuracy and develop new approaches to help shift to autonomous monitoring systems;
- your project will involve working with complex, noisy datasets where multiple leaks, scattering surfaces, and environmental noise present real-world challenges;
- a key part of the project is exploring novel approaches, such as physics-informed neural networks (PINNs), which combine deep learning with physical modelling to improve leak localisation;
- you will also investigate different architectures and develop multi-array configurations, balancing performance with practical deployment constraints.
This interdisciplinary project will place you at the forefront of developing safer, more effective gas leak detection systems, contributing directly to the future of hydrogen infrastructure.
