Accelerating Materials Innovation: Generative AI for Wear-Resistant Alloy Design

About the Project

This project is part of cohort 3 of the EPSRC CDT in Developing National Capability for Materials 4.0, with the Henry Royce Institute.

Friction and wear account for an estimated 23% of global energy consumption [1], driven by both energy losses and the need to replace degraded components. Current solutions rely on alloying with elements such as chromium and tungsten, but these come with significant environmental costs and increasingly fragile supply chains. As ore quality declines and critical materials become harder to secure, there is an urgent need for sustainable, high-performance alternatives. However, traditional materials discovery remains slow and reliant on trial-and-error, limiting progress. This project aims to accelerate the development of next-generation wear-resistant alloys that are both durable and environmentally responsible.

This project will employ a diffusion-based, structure-aware generative AI (GenAI) framework to develop novel wear-resistant alloys. Using MatterGen [2] as a basis for crystalline design, adaptive modules will be developed to train the GenAI model and generate new alloys through the use of targeted material properties as proxies for wear resistance. These will be verified through high-fidelity checks to ensure they are stable, processable compositions suitable for industrial use.

This vision will be achieved through the following project stages:

1. Curating a materials dataset of alloys with measured or computed mechanical properties relevant to wear resistance.
2. Adapting and retraining the MatterGen architecture within a constrained chemical design space aligned, including the development of an alloy-specific structural encoding suitable for diffusion-based generation.
3. Implementing property conditioning so MatterGen preferentially generates candidates with high predicted wear resistance (via proxies) while maintaining thermodynamic stability.
4. Developing a screening pipeline that combines surrogate models, DFT calculations, and simple processability filters to down select promising alloys with top candidates going forward for fabrication and experimental wear testing.
5. Comparison of the best GenAI designed alloys against established wear resistant materials and refining the workflow based on discrepancies between predicted and measured performance.

Through the EPSRC CDT in Developing National Capability for Materials 4.0, this research will establish a digital design loop in which GenAI and targeted testing are combined to accelerate the development of sustainable, wear-resistant alloys for demanding applications. By embedding data management, GenAI, simulation and experiment in a unified framework, the project exemplifies the Materials 4.0 CDT ethos: using digital methods not just to understand existing materials, but to codesign new compositions that address real world challenges around supply chain risk, resource efficiency and net zero goals.

Funding Notes

This is a fully funded project, part of cohort 3 of the EPSRC CDT in Developing National Capabilities in Materials 4.0. The studentship covers home fees, a tax-free stipend of at least £20,780 plus London allowance if applicable, and a research training support grant.

Enquiries

For general enquiries, please contact doctoral-training@royce.ac.uk.

For application-related queries, please contact Dr Dorothy Evans (dorothy.evans@strath.ac.uk).

If you have specific technical or scientific queries about this PhD, we encourage you to contact the lead supervisor, Dr Ashlee Espinoza (ashlee.espinoza@strath.ac.uk)

Application Process

Please note that each partner of the CDT in Materials 4.0 will have its own application process.

For Strathclyde, please email Dr Ashlee Espinoza, with Dr Dorothy Evans in cc (both email addresses above), using the subject line ‘Application to the Materials 4.0 CDT’.

The Materials 4.0 CDT is committed to Equality, Diversity and Inclusion. We strongly encourage applications from underrepresented groups.