What is fusion? Harnessing the energy of the stars.

Meet Theo Brown, a PhD researcher at the UK Atomic Energy Authority (UKAEA) working as part of our ongoing collaboration to design fusion energy reactors.

We are working with UKAEA to help make fusion energy a reality. As a part of our ongoing collaboration to accelerate fusion energy research our teams are using complex modelling and simulation techniques to further our understanding of fusion. Theo Brown researches how to apply state-of-the-art machine learning techniques to design future fusion reactors. He tells us more about fusion energy and shares some of the challenges our teams face when looking to build a fusion energy reactor for the first time.

Harnessing the energy of the stars

As we look to the future of climate change, we need to tackle the environmental challenges around energy production. In many ways, fusion would be the ideal energy source. Fusion is the process of combining two light nuclei to produce a heavy nucleus, a reaction that produces an enormous amount of energy – in fact, it’s the reaction that occurs naturally in the core of a star.  It is estimated that if we could harness fusion for energy generation, we could meet the energy needs of the world for more than a million years. The fuel for the reaction is hydrogen, which can be extracted from seawater. Fusion generates no greenhouse gas emissions or long-lasting radioactive waste. There is also no possibility of runaway chain reactions, making it relatively safe.

Harnessing the energy of the stars

In the words of Nobel physics laureate Pierre-Gilles de Gennes:

“We say that we will put the sun into a box. The idea is pretty. The problem is, we don’t know how to make the box.”

To achieve fusion hydrogen fuel must be heated to millions of degrees Celsius and compressed to very high densities, replicating the conditions at the centre of a star. This is a huge engineering challenge.

The fusion reactor

The most promising fusion reactor which can create the necessary reaction environment for commercial fusion is the tokamak. A tokamak uses magnetic fields to levitate fuel plasma in a doughnut-shaped chamber. The UK is currently designing a flagship next-generation reactor, the Spherical Tokamak for Energy Production (STEP), which aims to deliver electricity generated from fusion to the national grid by 2040. 

What is the safety factor?

One of the biggest challenges we face in building a fusion reactor is understanding how the plasma fuel moves. We use high-performance computing (HPC) to model the different properties of fusion plasma. Particles in a tokamak plasma follow a magnetic field and to keep the particles contained the field is twisted into a spiral. The angle of the magnetic field spiral, known as the safety factor, q, is one of our key priorities to optimise as the safety factor affects the efficiency and stability of the plasma. Designing a tokamak with good safety factor properties is vital in achieving commercial fusion energy production.

Example tori with different safety factor, q, credit: Theo Brown.

Designing tokamaks

Even on high-performance supercomputers, simulating tokamak designs is an incredibly time-consuming process as the physics of plasmas is highly complicated. Previously, the optimisation of the safety factor profile typically required hundreds of simulations. To help speed up the design process, we used a machine learning technique called Bayesian optimisation, where a model is trained to predict the performance of potential designs before selecting which designs to simulate in more detail. The model predictions help the optimiser find better solutions in less time. With supercomputers, these models are enabling us to design tokamaks in-silico, reducing production time and cost, and helping to make commercial fusion energy a reality. 

Looking forwards

Climate change is a global issue, and fusion energy has the potential to be a key part of the solution to resolve it. To make commercial fusion energy a reality we need to optimise our designs of fusion reactors. HPC is helping us to improve our modelling and tackle a lot of the difficulties we face around speeding the design process. As a part of our ongoing collaboration, we are looking to build a full digital twin of a fusion reactor bringing us closer to our goals of fusion energy production.