Introduction
Composite materials hold tremendous potential in enabling a low-carbon and sustainable future, particularly in renewable energy production, sustainable energy storage, and lightweight solutions for transportation.
Yet, conventional approaches to designing and deploying composites are not inherently sustainable and often lack the efficiency needed to meet rising performance demands under extreme conditions.
Hybrid composites—combinations of different fibres and matrix systems—offer unique opportunities to achieve superior balances of mechanical performance, cost-effectiveness, and sustainability. For instance, recent studies have shown that glass–carbon hybrid composites can improve energy absorption by up to 78% compared with single-fibre carbon systems under impact loading for an Aero-engine application, therefore improve the design efficiency and reduce fuel consumption [1] . Despite this promise, the design space for hybrid composites is huge. Variations in fibre type, stacking sequence and volume fraction etc produce complex constitutional structure–property relationships and pose significant challenges for recyclability and circularity at end-of-life for these multi-material systems. Traditional trial-and-error methods and purely physics-based simulations are both time-intensive and costly, limiting the exploration of this high-dimensional space.
Advances in artificial intelligence (AI), particularly machine learning (ML) and generative design, now offer powerful tools to address these challenges in conventional composites [2]. By integrating experimental datasets, computational simulations, and AI-driven optimisation methods, it becomes possible to rapidly identify and design novel hybrid composite configurations with tailored multifunctional properties, while also considering recyclability and circularity.
This project will develop an AI-driven framework to optimise the design and circularity of hybrid composite materials. The framework will integrate physics-based modelling, data-driven approaches, and experimental validation to accelerate the discovery of sustainable, high-performance hybrid composites with enhanced lifecycle performance. The student will:
collect data from experiment, published database, and also data that generated from pre-established 2D axisymmetric and multi-scale model,
apply ML methods to predict mechanical properties from compositional and other parameters,
employ multi-objective optimisation techniques (e.g., Bayesian optimisation) to identify trade-offs between mechanical performance, cost, and recyclability, incorporating lifecycle assessment (LCA) tools and circularity indicators,
perform experimental validation for these AI-optimised design.
References:
[1] Wu X, Finlayson J, Wisnom MR, Hallett SR, Improved energy absorption of novel hybrid configurations under static indentation, proceedings of the 20th European Conference on Composite Materials ECCM20. 26-30 June, 2022, Lausanne, Switzerland.
[2] Wang YF, Wang K, Zhang C, Applications of artificial intelligence/machine learning to high-performance composites, Compos B Eng, 285 (2024) 111740.