About the project
This project seeks to develop novel fluidic topology optimisation approaches for gas turbine component design by leveraging cutting edge machine and deep learning methods alongside computational fluid dynamics, ultimately aiming to revolutionise the solution of aerodynamic design problems.
Topology optimisation has grown in popularity over the past decade for structural problems but has yet to gain significant traction within the engineering community for aerodynamic problems. Although controlling the porosity over a control volume offers considerable design freedom, this approach tends to result in only a local optimum and struggles to effectively incorporate various constraints and objectives, such as manufacturability. This severely limits its application in gas turbine design.
This project aims to develop a novel approach to fluidic topology optimisation by building upon the cutting edge machine/deep learning approaches for optimisation and geometry generation developed by the Rolls-Royce University Technology Centre (UTC) for Computational Engineering.
By developing a unique feature-based fluidic topology optimisation process we aim to revolutionise aerodynamic design optimisation, producing more efficient designs in a greatly accelerated timeframe. While focused on gas turbine applications, for example, the optimisation of combustor, blade or seal features, this research has the potential to have a far reaching impact across engineering.
