The application of Artificial Intelligence & Machine Learning (AL/ML) to deliver enhanced radioprotection at nuclear plants

About the Project

This project is funded by the Nuclear Doctoral Focal Award in Radiation Protection, Nuclear Safety and Environmental Sustainability (RAPTOR) which offers students a coordinated, industry-engaged route to develop cutting edge expertise. The goal is to train at least 60 PhD students to deliver practical solutions across the civil and defence nuclear sectors with work shaped by national priorities, scientific curiosity and real world needs from industry.

The Department of Physics at the University of Liverpool is seeking to recruit an outstanding individual to join the RAPTOR doctoral training programme as a PhD student. The research project, a collaboration between the University of Liverpool and Mirion will utilise Artificial Intelligence / Machine Learning (AI/ML) approaches coupled with health physics (HP) instrument data to materially strengthen radioprotection (RP) and ALARP in nuclear plants. This will be achieved by turning today’s point‑based measurements into predictive, physics‑consistent intelligence without replacing established RP practices or regulatory frameworks.

Mirion currently provides a range of HP instrumentation to nuclear plants including Fixed area gamma/neutron monitors, Airborne activity monitors (particulate, iodine, noble gas), Portal and contamination monitors, Portable survey instruments and Personal dosimeters. These are point measurements with limited special coverage, and the data are generally not well integrated. The goal would be to use AI/ML to augment these instrument data, connecting, and extrapolating under physical constraints.  Applications could include:

1. Continuous dose-rate field reconstruction producing a continuous 3D dose rate map to support radioprotection decision making for improved plant access control and worker dose minimisation.
2. Enhanced task-based worker dose planning, predicting dose prior to tasks to support staffing levels, work sequencing and shielding.
3. Plant contamination mapping to support access control, shielding and targeted clean-up

These use cases are particularly important for real-time outage ALARP support where the source term can change dynamically. The project will involve researching the state of the art in the emerging field of AI/ML to select the most effective approach. It will be important for regulatory approval for the models to be physics-constrained, respecting radiation transport physics and the relevant radioprotection physics. A relevant data strategy will be defined and deployed, including both modelled and actual Mirion plant instrumentation and health physics data. The student will build and train the models and test against traditional methods.

The student recruited to this project will be part of RAPTOR a new EPSRC doctoral training programme focused on radiation protection, environmental sustainability and nuclear safety. You will be based at the University of Liverpool and will undertake an intensive training programme in year 1 which leads to a PgCert, co-designed and co-delivered by partner institutions and Industry. This PhD programme offers a unique opportunity to join a vibrant, inclusive cohort focused on addressing major interdisciplinary challenges in nuclear science and engineering. It unites leading academic expertise with strong partnerships across the nuclear industry and will train students in nationally critical skills in cutting edge nuclear technologies.