Algorithm for better nuclear fusion developed with Google researchers
Image credit: Tri Alpha Energy
Researchers at Google have been working with a nuclear fusion start-up, Tri Alpha Energy, to develop an algorithm which combines human and computer intelligence to accelerate nuclear fusion research.
Nuclear fusion – the process which powers stars – holds enormous promise as a means of acquiring clean, renewable energy. The processes of nuclear fusion, smashing together atomic nuclei with enough energy for them to overcome their opposing forces and remain together, releases far more energy than nuclear fission, only requires hydrogen as fuel, and leaves behind no dangerous products.
However, the sheer amount of energy required for nuclear fusion means that nobody has ever managed to sustain fusion in a reactor for longer than a brief moment.
The researchers at Google and Tri Alpha Energy, which is backed by Goldman Sachs, the Wellcome Trust and Microsoft co-founder Paul Allen, collaborated to create an algorithm to improve the stability of these nuclear fusion experiments.
The ‘Optometrist Algorithm’ is a machine-learning tool that helps investigators choose between various parameters for an experiment. These experiments can involve thousands of sensitive factors that affect outcomes.
“The highly nonlinear and temporarily varying interaction between the plasma, its environment and external controls presents a considerable complexity in these experiments. A further difficulty arises from the fact that there is no single objective metric that fully captures both plasma quality and equipment constraints,” the researchers wrote in Scientific Reports.
The Optometrist Algorithm combines human and machine intelligence; investigators are asked to choose between pairs of outcomes in order to narrow down to an optimal set of parameters. This human input is necessary to prevent experiments being devised which could be dangerous or useless.
“To increase the speed of learning and optimisation of plasma, we developed the Optometrist Algorithm. Just as in a visit to the optometrist, the algorithm offers a pair of choices to a human, and asks which one is preferable.”
“Given the choice, the algorithm proceeds to offer another choice,” the researchers wrote in Scientific Reports.
The algorithm allowed the researchers to hold extremely hot nuclear plasma stable for longer periods during work on Tri Alpha Energy’s C-2U plasma generator. The start-up hopes to be able to use the Optimetrist Algorithm on its new reactor, Norman, and suggests that it could be applied to other problems unrelated to energy.
Standard nuclear fusion reactors, such as the Joint European Torus in Oxfordshire, tend to take a ‘doughnut’ ring shape (the tokamak design). California-based Tri Alpha Energy, however, has developed an alternative nuclear fusion design which fires beams of plasma into a vessel pinned in place, spinning, by a magnetic field.
Michl Binderbauer, president and CTO of Tri Alpha Energy, described the research as having “huge promise”, and having the potential to “speed up” research in this difficult area.