Gustavo de Almeida (UoS), Sergio Maldonadfo (UoS)
To apply for this project please click here. Tick programme type – Research, tick Full-time or Part-time, select Academic year – ‘2024/25, Faculty Environmental and Life Sciences’, search text – ‘PhD Ocean & Earth Science (FLOOD CDT)’. In Section 2 of the application form you should insert the name of the project and supervisor(s) you are interested in applying for. if you have any problems please contact I.D.Haigh@soton.ac.uk.
To apply for this project please click here. Tick programme type – Research, tick Full-time or Part-time, select Academic year – ‘2024/25, Faculty Environmental and Life Sciences’, search text – ‘PhD Ocean & Earth Science (FLOOD CDT)’. In Section 2 of the application form you should insert the name of the project and supervisor(s) you are interested in applying for. if you have any problems please contact I.D.Haigh@soton.ac.uk.
Rationale:
Floods are the most devastating and costly among all natural hazards. The risk of flooding is expected to rise substantially in the coming decades as population growth increases the exposure of people and assets, and as the climate emergency changes the intensity and frequency of storms and also accelerates sea level rise. Flood inundation models are widely used to understand and design measures to mitigate the risk of flooding. Models currently available are based on the solution of the two-dimensional shallow water equations, a system of nonlinear partial differential equations expressing the principles of mass and momentum conservation. To simulate real-world problems accurately, these models need to be run using finely resolved topography. This translates into long computing times that often limits the size of the domains and/or the duration of the events possibly modelled. Given the growing need for simulations of large domains and for multiple simulations used in probabilistic forecast, available techniques are not fit for purpose.
Methodology:
Recently, new Artificial Intelligence (AI) techniques (e.g. deep learning) have started to find applications in flood inundation modelling. In particular, new research indicates that deep-learning algorithms have a huge potential to offer solutions that may outperform conventional techniques of numerical integration of the shallow-water equations. In this project you will work at the forefront of AI methods to develop and test the most advanced, purely AI-driven model to simulate the propagation of flood inundation at high-performance. The successful applicant will have an excellent degree in applied mathematics, physics or a relevant engineering subject. Ideally, the candidate should have some experience in fluid dynamics/hydraulics and machine learning. You will join a world-leading research team and environment at the University of Southampton, a Russell Group member ranked as one of the world’s top 100 universities. Of particular importance for this project is the access to outstanding supercomputing facilities at the University of Southampton.
Location:
University of Southampton
Background Reading:
- Shamkhalchian, A. and de Almeida, G. A. M (2020), Upscaling the shallow water equations for fast flood modelling, Journal of Hydraulic Research, Vol. 59, 2021, Issue
- Xia, X, Liang, Q, Ming, X (2019) A full-scale fluvial flood modelling framework based on a high-performance integrated hydrodynamic modelling system (HiPIMS), Advances in Water Resources, 132, 103392,
- Qi, X. de Almeida, G.A.M., Maldonado, S. (pre-print) Physics informed neural networks for solving flow problems modelled by the Shallow Water Equations;
Contact Email:
