The idea for ITER was initiated by Gorbachev and Reagan in the 1980s, at the height of the Cold War, when the two leaders agreed to cooperate on developing nuclear fusion, a clean energy technology that could one day power the world. The project began a year after the European Union, Japan, the Soviet Union and the United States began designing a massive international fusion facility.
In the early 2000s, China, South Korea, and India joined four other countries in collaborating to build and operate an experimental reactor, still under construction in Cadarache, outside Aix-en-Provence, France, to enable scientists and engineers around the world to learn enough about nuclear fusion to design prototype commercial reactors within the next few decades.
The seven members are making both financial and in-kind contributions in the form of parts and services that help each country develop local know-how in specific areas. The in-kind contributions help achieve one of the project’s sub-goals: to foster a global industry so that when fusion becomes commercially viable, many parts of the world will know enough about it to run their own power plants.
Recently, the ITER Council agreed to start allocating resources to support the growing private fusion industry, and held a major workshop to kick off this new work. The first private sector workshop took place from 27 to 29 May at the ITER Organization headquarters in Cadarache, France.
ITER Public Affairs Manager Laban Koblenz said people from all over the world were in attendance, representing a range of roles in the growing ecosystem, including investors, policymakers, fusion companies and various supply chain partners.
The event brought together about 300 people from a variety of organizations, many of whom compete at some level and who may not agree on the best approach to convergence, but provided an opportunity to share ideas and find common ground in their vision for the future.
The various approaches to nuclear fusion may be summed up into two broad categories: the first is magnetic confinement fusion (MCF), which uses a magnetic field to confine plasma, requiring large machines such as the ITER tokamak, still under construction, or the Wendelstein 7-X stellarator, which began operation in 2015. The second broad category is inertial confinement fusion (ICF), which uses a powerful impulse to compress the plasma and initiate the fusion reaction.
According to Andrew Holland, CEO of the Fusion Industries Association (FIA), the number of private startups aiming to build nuclear reactors using unique approaches has surged since around 2021, with about $4 billion invested in the same period. Of the $6 billion invested in fusion worldwide, about 80% has gone to U.S. companies, but geographic diversity is increasing.
“We are seeing more ambitious startups emerge and attract more investment, particularly here in Europe and Japan,” Holland said in his opening presentation at the workshop.
Computer technology supports growth of the nuclear fusion industry
Regardless of the approach, nuclear fusion requires computer technology in at least five key areas for every project. First, to run simulations to verify the underlying physics before the machine is designed. Second, to run one or more control systems that coordinate the various components in the machine. Third, to support diagnostic systems that observe different aspects of the operation so that immediate action can be taken or the machine’s behavior can be investigated. Fourth, to power the robots needed to perform repairs and maintenance in the harsh environment of a reactor. Fifth, to collect and store data from the diagnostic systems to improve the models so that better machines can be designed in the future.
Many different types of computer technology are required, including high-performance computing to run complex simulations and analyze diagnostic data, radiation-hardened integrated circuits to support diagnostic equipment and control robots used for repair and maintenance, high-speed storage networks to collect and store data during operation, control software to synchronize components, and specialized software for simulation.
Ignition Computing, a Dutch company represented at the event, represents a very specific niche in the growing ecosystem involving the creation and integration of simulation models. [modelling software] “Typically we focus on a specific physical phenomenon, such as magnetic equilibrium,” Daan van Vugt, founder and CEO of Ignition Computing, explains to Computer Weekly . “There may also be other codes that model transport, that is, the movement of the particles themselves.”
Whatever the approach to fusion, the first step is to “discretize” the data: the world is analog, but computers need digital information, so both the measurements and the time of each measurement must be digitized before they can be used in a simulation.
“You have to break the data down into discrete points so that it can be processed by a computer,” says Van Vugt, “and through this discretization, you turn the physics equations into the simultaneous equations that you’re trying to solve.”
In most cases, you’ll need to study more than one phenomenon at a time. This is called multiphysics simulation, and van Vugt says there are two different approaches to dealing with it. One is to solve each phenomenon one at a time — for example, solve for density, then solve for temperature, and iterate in between. The other approach is to solve them together, putting them into one coupled equation and solving them all at once.
Code developed to model an aspect of physics is sometimes released as open source. Depending on the approach to fusion, different existing codes can be reused from one project to another, saving significant time and effort. We hope that lessons learned at the first Private Sector Workshop will lead to significantly increased sharing for the benefit of humanity.
