Don't miss our weekly PhD newsletter | Sign up now Don't miss our weekly PhD newsletter | Sign up now

  PhD Studentship: Beyond data-driven: physics-informed and trustworthy machine learning for monitoring and optimization of machines and processes in advanced manufacturing

   School of Engineering

This project is no longer listed on and may not be available.

Click here to search for PhD studentship opportunities
  Dr Min Xia  No more applications being accepted  Funded PhD Project (Students Worldwide)

About the Project


A full PhD studentship (3.5-year stipend and tuition fees at the UK rate) is offered at the Department of Engineering, Lancaster University, UK to undertake research in exploring how physics-informed machine learning can increase the generalization and trustworthiness of AI-based methods in monitoring and optimization with applications in complex machines and processes in advanced manufacturing. 

Effective monitoring and optimization of machines and the processes are crucial in advanced manufacturing to improve manufacturing efficiency, avoid defects, reduce downtime of machines, eliminate the need for post-process inspection, increase productivity and reduce cost. In the past decade, machine learning (ML) has been widely applied in many areas including monitoring and optimization. However, the application of ML in these areas still faces significant hurdles and challenges due to the limitation in generalization of ML models (may not work well in reality), shortage in interpretability of current ML methods (not fully understand why the models work), and the high cost of collecting and labeling sufficient sensory data for training. In this research, revolutionary strategies and novel methodologies will be developed with interpretable and trustworthy ML.


This research will largely foster the effective and confident utilization of ML to achieve effective monitoring and optimization of machines and manufacturing processes. The main objectives include:

1. Develop novel hybrid methodologies for fault and defect detection that effectively integrate physics-based modeling with ML approaches, referred to as physics-informed machine learning (PIML) techniques. Build a physics-informed neural network (PINN) framework that fuses both data and first physical principles into the neural network to inform the learning processes.

2. Develop transfer learning-based methodologies that can fully utilize well-trained deep learning models to achieve strong generalization in small data regime.

3. Develop uncertainty quantification methods to measure the reliability of the ML models when out of distribution data is observed e.g. new machine or new material is used or the sensory data is corrupted with heavy noise.  

4. Develop an incremental learning framework to achieve continuous and adaptive learning from continuous incoming data of the practical system.

5. Develop an ML model for manufacturing quality prediction and achieve process parameter optimization with evolutionary computing techniques. 

About University/Department

Lancaster University has a world ranking of 146th out of more than 1,000 universities in the QS World University Rankings 2023. It is a strong and dynamic university with a very highly regarded Engineering Department. In the 2021 Research Excellence Framework, 95% of its research rated as world-leading or internationally excellent. Lancaster’s approach to interdisciplinary collaboration means that it has pre-eminent capacity and capability for the integration of Engineering with expertise in the areas of data science, autonomous and learning systems, intelligent automation, materials science and cyber security. The University is developing an ambitious growth plan for Engineering, including investment in staff, doctoral students, equipment and a new building focused on research themes including Digital and Advanced Manufacturing. Lancaster University has been rated top in the region in the Times and The Sunday Times Good University Guide 2023 and sits 12th on the table nationally.

Qualifications and experience:

·        Candidates should have a relevant degree at 2.1 minimum or an equivalent overseas degree in Mechanical Engineering Mechatronics/ Electrical Engineering/ Computer Science / Industrial Engineering.

·        A good background in machine learning and computer programming is desirable.

·        Excellent oral and written communication skills with ability to prepare presentations, reports and journal papers to the highest levels of quality.

·        Excellent interpersonal skill to work effectively in a team consisting of PhD students and postdoctoral researchers.

Non-UK students are welcomed to apply. Overseas applicants should submit IELTS results (minimum 6.5) if applicable.

How to Apply:

For further information, please contact Dr Min Xia ([Email Address Removed]). Candidates interested in applying should send a copy of their CV to Dr Min Xia. Closing date: 30 November 2022.

Computer Science (8) Engineering (12) Mathematics (25)

Funding Notes

This project is funded by Lancaster University. The funding covers the cost of tuition fees and a standard tax-free RCUK stipend for 3.5 years for UK applicants. Non-UK students are welcome to apply, but overseas tuition fees will be charged. The successful candidate can start in January 2023 or as early as possible.