Funding providers: Swansea University and TWI Ltd.
Subject areas: Aerospace engineering, computer science, mechanical engineering
Project start date:
- 1 January 2024 (Enrolment open from mid-December)
Project supervisor: Professor Chenfeng Li ([Email Address Removed])
Aligned programme of study: PhD in Aerospace Engineering
Mode of study: Full-time
Project description:
Non-destructive testing (NDT) is a testing and analysis technique used by industry to evaluate the properties of a material, component, structure or system for characteristic differences or welding defects and discontinuities without causing damage to the original part. Two techniques offered by TWI are advanced Ultrasound Testing (UT) and X-ray Computed Tomography (XCT.)
However, both technique can produce large 3D datasets that can be challenging to interrogate for human experts to fully analyse. Small defects or anomalies may be missed when manually examining through the acquired volumetric data. Artificial intelligence (AI) has emerged as a promising solution to automate and enhance defect detection in these complex structural data. The aim of this study is to improve diagnostic accuracy through development an AI system that can perform rapid processing of large 3D data volumes, and identify defects that are difficult to discern with the naked eye and provide a second opinion with a level of certainty to assist human experts.
The PhD work will be delivered through the following measurable objectives:
- Literature review of the challenges associated with different NDT disciplines, with a particular focus on volumetric data from XCT and UT.
- Investigate various AI techniques to improve the analysis of NDT data.
- Develop novel algorithms and models specifically designed for limited available data scenarios.
- (Option) Explore data augmentation techniques and semi-supervised learning methods to enhance the training process.
- Investigate the integration of different NDE data types and develop a comprehensive analysis framework.
- Validate the AI models on real-world NDE datasets.
The nature of the research work will be largely on data science and AI, with the support of material characterization using the advanced material analysis suites at the Faculty of Science and Engineering of Swansea University.
Eligibility
Candidates must hold an Upper Second Class (2.1) honours degree in Engineering or similar relevant science discipline. If you are eligible to apply for the scholarship (i.e. a student who is eligible to pay the UK rate of tuition fees) but do not hold a UK degree, you can check our comparison entry requirements (see country specific qualifications). Please note that you may need to provide evidence of your English Language proficiency.
Due to funding restrictions, this scholarship is not open to applications from international students (unless eligible to pay UK tuition fee rates as defined by UKCISA regulations).