Thermocouple layout optimisation towards divertor digital twins using Bayesian experimental design
doi: 10.2139/ssrn.7454942
Mike Battye
Data scientist and PhD candidate working on fusion digital twins, Bayesian experimental design, and applied ML for high-assurance systems.
University of York (Fusion CDT) · previously Assystem and Mirada Medical.
I build data and model systems that have to be trusted. Medical imaging first, then nuclear engineering, now tokamak divertors.
About
I spent the first part of my career as an industry data scientist: deep-learning contouring for radiotherapy at Mirada Medical, then data platforms and analytics in nuclear and engineering contexts at Assystem.
I am now a PhD candidate at York, working on divertor digital twins for through-life management of the heat-exhaust component of a tokamak. That means diagnostic design, reconstructing fields from sparse sensors, and prognostics, not only a snapshot of the present.
In 2025 I published a first-author review on digital twins in fusion energy research in IEEE Access.
I take a small number of part-time remote engagements. Separately, I am exploring a 2026–27 research-translation venture around this divertor work.
Selected work
Fusion digital twins
First-author review of digital twins in fusion (IEEE Access, 2025). The PhD is diagnostic design, sparse-sensor reconstruction, and prognostics for divertor twins; public pieces so far are the review, a submitted diagnostic-layout preprint, and IAEA contributions. Publications & talks.
Mirada DLCExpert
Shipped deep-learning contouring models for radiation oncology. Co-author on the 2023 Physica Medica evaluation of CT- and MR-based automatic contouring for the EPTN neuro-oncology atlas. Read the case.
Assystem
Data science for nuclear clients: platforms, analytics, and digital-twin-adjacent modelling support. Read the case.
How to work together
Contract work
I take a limited number of part-time remote contracts for data and modelling work in energy, nuclear, medical, or other regulated settings.
Typical range £65–£80/hr depending on scope and IR35. Max commitment compatible with the PhD; serious scopes only.
Research translation
Incubating a fusion-tech spin-out in 2026–27 around divertor digital twins and diagnostic design. Open to conversations with technical partners; investor intros welcome once there is a fit. Useful conversations right now are technical: data, sensors, simulation, and what a productised designer would need to look like.
Publications
Digital Twins in Fusion Energy Research: Current State and Future Directions
More papers and talks
Geometric and dosimetric analysis of CT- and MR-based automatic contouring for the EPTN contouring atlas in neuro-oncology
Bayesian Experimental Design for Divertor Heat Loads in Fusion Power Plants
A Divertor Diagnostic Designer for a Digital Twin Environment: producing inputs for Bayesian experimental design and other optimisation workflows
Contact
Industry / contract: [email protected]
Academic: [email protected]