About the Project
A fully funded PhD opportunity to participate in the world-leading research undertaken by the EPSRC Doctoral Landscape Award at the University of Sheffield.
Supervised by Dr Aidan Hughes and Prof Keith Worden.
The University of Sheffield is one of the leading Russell Group universities in the UK. We carry out cutting-edge research with strong links to industry. When you enrol to do a PhD with us, you will be working with world-leading academics and have access to top of the range facilities. As a PhD student you will have the opportunity to gain skills not only to conduct research, but also to take your career to the next level, whether you want to stay in academia, go into industry or the public sector, or set up your own company. You will have access to a range of training and support services to help you excel in your studies and beyond.
Project Description
It is well established that medical treatments can vary in their effectiveness from patient to patient, based on factors such as their age, socio-economic background, and the other pathologies that may be affecting an individual. In much the same way, the interventions that engineers take to maintain and operate structures (e.g. repairs, component replacements) will vary in effectiveness from system to system, depending on factors such as the age, environment, material, and design.
Properly quantifying the effects of interventions over populations of structures is vitally important for decision-making as, without accurate models of these effects, an operator cannot be confident that a given intervention will have the desired effect, or have any effect at all. Moreover, in some situations, inaccurate intervention models will lead to suboptimal decisions being made, potentially jeopardising the safety of a structure and introducing resource inefficiencies.
Traditional approaches to machine learning establish correlations between some inputs and some outputs. In the context of learning intervention models, this would mean finding correlations between interventions (the input) and their effects (the output). In some situations, correlations can be useful for making predictions, however, these predictions will be dependent on idiosyncrasies and biases associated with the training data. This limits the ability to generalise and transfer these predictions to previously unseen scenarios and structures. To circumvent the limitations associated with learning correlations, we can instead learn causal models in which causal relationships between inputs and outputs are established. To effectively apply traditional machine-learning methods to learn causal models, we would require randomised controlled trials for a population of engineering systems. This scenario is impractical for a number of reasons, for example, the structures we are interested in managing do not exist within controllable environments.
To circumvent the limitations associated with traditional machine learning techniques, we can turn to causal machine learning. Causal machine learning allows for the discovery of causal relationships from purely observational data. Causal machine learning methods have recently been identified as an approach for determining treatment effects in the medical field, in situations where it is impossible to control for all variables associated with a population [1].
This project will apply causal machine learning to learn intervention models for engineering populations, accounting for potential confounding variables such as differing environments, and sources of damage. The work will be highly applicable in a broad range of contexts including bioengineering, and nuclear energy. While the application areas are broad, the current project will focus on three main applications that each exemplify an engineering population; specifically, offshore wind farms, networks of bridges, and an inventory of manufacturing machines. Each of these applications stand to benefit greatly from the enhanced decision-support capabilities that would be provided by data-informed intervention models, driving value in areas that are critical to strategic aims in the UK in terms of economy and sustainability.
How to apply
Interested candidates are strongly encouraged to contact the project supervisors to discuss your interest in and suitability for the project prior to submitting your application.
Please refer to the EPSRC DLA webpage for detailed information about the EPSRC DLA and how to apply. Applicants must ensure that they use the correct application form linked to the EPSRC DLA webpage and must submit an application for October 2027 entry. The form is unique to the EPSRC DLA. It contains the list of advertised projects from which to select, and three questions specific to the DLA.
The application deadline is 17:00 GMT 13 January 2027.
The majority of our DLA studentships will be allocated to candidates who qualify for Home fee status, in accordance with EPSRC funding regulations which place strict limits on international recruitment.
Funding Notes
The award will fund the full tuition fee and a maintenance stipend at the UKRI rate (currently £21,805 per annum) for 3.5 years, as well as a research grant to support costs associated with the project.
References
[1] S. Feuerriegel et al., Causal machine learning for predicting treatment outcomes, Nature Medicine 30, 958-968, 2024.