Available student project - Machine learning approaches for nuclear fusion reactions

Research fields

Project details

Low energy nuclear scattering reactions invloving protons and neutrons are critical to the astrophysical processes that create the elements in stars and stellar explosions. They can be used on Earth to characterise materials, and can be used to produce power through fusion and fission. They are, however, delightfully complex. The reaction must proceed through quantum states of the combined nucleus, and these states overlap, such that disentangling them to understand the reaction and make predictions is very challenging.

One reaction of partcular interest is the fusion with a proton with Boron-11. This reaction has enormous potential as an energy source, as it produces three alpha particles and releases a large amount of energy. Crucially, no neutrons are produced, meaning that fusion power reactors could be more compat and require much less maintinence. However, the probability of this reaction occuring as a function of energy is not well understood.

Using cutting-edge nuclear models and machine learning, this project will involve the evaluation of low-energy nuclear scattering and reaction data for proton-Boron-11 collisions to enable us to predict the reaction probability at energies where there is currently no information.

Project suitability

This research project can be tailored to suit students of the following type(s)

Contact supervisor

Simpson, Edward profile