Discovering the Energy of the Future

The essence of our projects is adding new functionality to existing software. However, as software is only a means to an end, a question arises right away: what is it going to be used for? The answer is an exciting one: for simulations of a fusion reactor! Fusion is a fascinating and currently a very “hot” subject, so after an introduction by our mentor and some additional internet research, we feel ready to share our understanding and why HPC is important in this context.

The fusion

First off, let us briefly describe how fusion works. The goal is to release energy. This happens by the following process: first, two atoms (deuterium and tritium, which are both hydrogen isotopes) join (or “fuse”) together to form a new, unstable, molecule. The new molecule releases a neutron, thus becoming a stable Helium particle. As the mass of the resultant molecule is smaller than that of the two original particles, the missing “energy” is released.

Easy as it sounds, fusion can only take place at temperatures around 100 million degrees. At such high temperatures, matter takes form of plasma – a cloud of charged particles. This is the most common state in the universe, and here on earth plasma can be found in lightning or neon signs.

Heating matter to such high degrees causes particles to lose or gain electrons (to become ionised). Problems with plasma arise when one tries to confine it. As it is the hottest state of matter, naturally the particles in it are the most energetic, and hence are hard to keep in a confined space, away from material surfaces. This is one of the most difficult problems in using fusion reactors.

Currently, there are three known ways to confine plasma: gravity, inertia and using magnetic fields. The latter encapsulates the plasma in a current, such as done in a tokamak fusion reactor.

The reactor

The reactor simulated in the project is tokamak: a word derived from Russian for “toroidal chamber with magnetic coils”, which is essentially a very large doughnut-shaped vacuum chamber. A gas of ionised deuterium and tritium particles is pumped into the vacuum, until a high density is reached. Then, a beam of high velocity neutral deuterium particles is fired into the chamber to collide with the gas particles, increasing the temperature of the system greatly. A pulse of electric current is then induced, heating the system even further, and turning ionised gas into plasma. The high temperatures of around 100 million degrees Celsius cause the deuterium and tritium particles to fuse together, thus releasing energy, just as happens inside stars.

Joint European Torus (hereafter affectionately called JET) is the world’s largest tokamak fusion reactor located in Oxfordshire, UK. From the official JET website, it “investigates the potential of fusion power as a safe, clean and virtually limitless energy source for future generations”. To see what JET looks like in action, click here. However, not all the research is done experimentally – simulations also play an important role in understanding the science. While simulating plasma behaviour is relatively easy, problems arise when trying to model what happens near the walls of the tokamak: the high voltage drop causes a so-called “singularity” – something that is virtually impossible to model. That’s the sort of cutting-edge science we will be working with – assisting in visualising JET reactions using HPC.

The motivation

There are numerous benefits to using nuclear fusion, as it is one of the most promising options for generating large amounts of carbon-free energy in the future. According to current estimates, the costs of using fusion energy can be comparable with that of fission, renewables and fossils.

As described earlier, the fuel used in fusion is tritium and deuterium, and their supply is essentially inexhaustible. While deuterium can be extracted from water, tritium can be produced from lithium, which is found in the earth’s crust.

In addition to that, there are numerous other advantages of fusion, namely:
– No emission of harmful toxins
– Only helium is produced in the process, which is already abundant in the atmosphere and will not contribute to global warming
– Efficiency: 1kg of fusion can provide the same amount of energy as 10 million kg of fossil fuels
– Safety: as only small amounts of fuel are used at a time, large-scale nuclear accidents are virtually impossible

The problem and the solution

Unfortunately, currently fusion requires much more energy input than what is produced. Many countries are phasing out larger-scale independent fusion research due to funding restrictions, and instead are concentrating their efforts on joined endeavours such as JET. While the domestic tokamaks are still used for educational purposes and research within the machines’ limits, the break-throughs are expected to come from the larger reactors. And that’s where we fit in too – simulating the process to contrast it with experiment. If the two match, there is a chance we understand the science, allowing us to extend the given model to simulate larger reactors which are not yet physically possible.

 

High Performance Computing

So, why HPC is important for nuclear fusion? Nuclear fusion simulations generate huge amounts of data.This data needs to be processed and visualised to gain an understanding of the experiments, and for the researchers to be able to compare experimental and simulated results. Such processing and visualisation of data require large amounts of processing power that only HPC systems can provide. In these systems the hardware and the software are just as important: the hardware must provide high processing capabilities and the software must be efficient to run fast and make the most of those capabilities. Our task is to extend this software, while keeping in mind high performance, efficiency and the functionality which provides the desired insights into the science of fusion.

To end, here are some final words of motivation: “With adequate funding, the first fusion power plants should be operating by 2040.” [2]

This post was written by Luna and Evguenia.

———

[1] This picture is borrowed from http://en.wikipedia.org/wiki/File:JointEuropeanTorus_internal.jpg with accordance to the copyright

[2] Culham Centre for Fusion Energy, http://www.ccfe.ac.uk/Fusion_power.aspx