Machine learning (ML)-assisted detection of micromechanical fracture in bioinspired composites
Collaborating with the National Physical Laboratory (NPL), our project pioneers machine learning to enhance micromechanical fracture detection in bioinspired composites, advancing materials science and engineering.
Start date1 January 2024
- We are seeking a self-funded candidate
This collaborative research project, conducted in partnership with the National Physical Laboratory (NPL), focuses on revolutionising micromechanical fracture detection within bioinspired composites through the innovative use of machine learning techniques. These bioinspired composites exhibit remarkable mechanical properties with diverse industrial applications. However, ensuring their structural reliability necessitates precise detection and assessment of cracks during micromechanical fracture toughness testing.
The intricate microstructure and heterogeneous characteristics of these composites pose unique challenges to automated crack detection. These challenges are further compounded by the resource-intensive processes of data acquisition and labelling. This pioneering initiative aims to harness the power of transfer learning by adapting pre-trained deep learning models, initially designed for general image analysis, to the intricacies of crack detection within bioinspired composites.
Our objectives encompass both technological advancement and practical applicability. We aim to:
- Investigate the effectiveness of transfer learning in elevating the accuracy of crack detection during micromechanical fracture toughness testing in bioinspired composites.
- Develop a robust transfer learning framework that integrates pre-trained deep learning architectures and tailors them for precise crack detection within these challenging materials.
- Identify optimal strategies for fine-tuning pre-trained models, addressing the unique challenges of crack detection within bioinspired composites.
- Conduct comprehensive evaluations to quantitatively measure the framework's performance, employing metrics such as Intersection over Union (IoU) and F1 score. A comparative analysis against traditional crack detection methods will highlight the superiority of transfer learning.
The potential impact of this research is substantial. Successful implementation could significantly advance the field of materials science and engineering by enabling more accurate and efficient assessments of fracture toughness and structural reliability in bioinspired composites. Moreover, the transfer learning framework developed here could pave the way for analogous advancements in other materials with complex microstructures, amplifying its influence across diverse industries.
We are seeking an enthusiastic and motivated candidate to join our groundbreaking research project focused on leveraging transfer learning techniques for crack detection in bioinspired composites. The ideal candidate should possess a strong foundation in materials science, image analysis, and deep learning methods. This is a unique opportunity to contribute to cutting-edge research that bridges the gap between materials engineering and artificial intelligence.
Qualifications and skills:
- A bachelor’s or master’s degree (or equivalent) in Computer Science, Physics, Engineering, or a related field.
- Strong background in image analysis, computer vision, or deep learning techniques.
- Proficiency in programming languages such as Python and experience with deep learning libraries (TensorFlow, PyTorch, etc.).
- Familiarity with transfer learning concepts and applications.
- Ability to work with large datasets and preprocess image data for analysis.
- Strong problem-solving skills and a critical mindset to address challenges in crack detection.
- Excellent communication skills for collaboration within an interdisciplinary research team.
- Collaborate with research partners to develop machine learning frameworks for detection of micromechanical crack in bioinspired composites.
- Collect and preprocess image datasets of bioinspired composites for training and evaluation.
- Implement and adapt deep learning models for crack detection accuracy enhancement.
- Analyse and interpret results, applying relevant performance metrics.
- Contribute to research publications, presentations, and potentially patent applications.
This is an exciting opportunity to engage in transformative research with real-world implications. The successful candidate will work closely with an interdisciplinary team of experts and contribute to advancing both materials science and deep learning applications.
Open to any self-funded UK or international candidates. You will need to meet the minimum entry requirements for our PhD programme.
How to apply
Application should be made through the Engineering Materials PhD programme page. In place of a research proposal you should upload a document stating the title of the project that you wish to apply for and the name of the relevant supervisor.
Read our studentship FAQs to find out more about applying and funding.