Skip to main navigationSkip to main content
The University of Southampton
Intelligent & Resilient Ocean Engineering – Royal Academy of Engineering Chair in Emerging TechnologiesNews and Events

IROE Early Career Researcher, Dr Jared Charles, presents his work at STEM for Britain Finals 2023

Published: 24 March 2023
Dr Jared Charles at S4B
Dr Jared Charles presenting his research at Westminster

Jared Charles, Research Fellow in Offshore Renewable Energy (Geotechnics), supported by the Royal Academy of Engineering Chair in Emerging Technologies Centre of Excellence for Intelligent and Resilient Ocean Engineering and the Supergen ORE Hub presented his research on deriving geotechnical parameters for offshore infrastructure foundations using machine learning at STEM for BRITAIN 2023.

STEM for Britain is an annual exhibition of posters by early-career research scientists, engineers and mathematicians held in Westminster by the Parliamentary and Scientific Committee. The event fosters engagement between early-career researchers and parliamentarians. Jared presented his poster (available here) in the engineering category of the STEM for Britain finals on the 6th of March 2023 at the House of Commons.

Jared’s poster titled “Application of machine learning to derive geotechnical engineering parameters for safe and efficient design” first introduced the challenges facing the offshore wind industry: the very ambitious government targets for future turbine installation, the ever-increasing size and complexity of offshore sites, and the current lack of industry capacity in terms of vessels and trained personnel. Jared then proposed a solution – the use of machine learning (in this case artificial neural networks) to recover key but difficult to measure design parameters from a range of simpler soil properties. Shear stiffness degradation – a measure of how soils become less stiff with deformation – is key for the design of offshore wind turbines but is challenging to measure at the scales needed for sites that can be as large as 200km2. Jared has developed an open access software implementing a neural network that can predict shear stiffness degradation for any number and combination of 8 simple geotechnical soil properties. The outputs are more accurate than pre-existing empirical approaches. The software has been published as a graphical application that practising engineers can freely download.

Notes for editors

Next >>
Privacy Settings