Advanced control strategies for fusion energy devices

Fusion energy can provide clean and endless electricity, but stabilizing a fused energy device is extremely complicated. Upgrading the UK’s Fusion Machine Mast and Scientists working at the Dutch Institute have made significant advances in two important areas

How can I spread strong heat evenly within the reactor? How can you control the amount of fuel in the core? These are some of the questions that arise when digging deeper into the control strategies needed to realize fusion as a future energy source.

Experiments by British and Dutch researchers show that even small shifts in the magnetic field in a reactor can instantly change where heat and particles go. We also developed smarter ways to manage fuel levels, such as using frozen pellets and advanced computer models to make the reactor run smoothly. Important developments, like the UK Steps and international ITER projects, paving the way for future fusion power plants.

Power Sharing Dynamics

Additional divers can be installed on top of the machine to distribute the enormous amount of power in the fusion reactor to the vessel wall. This so-called double-null configuration is foreseen in the UK’s Stepping Fusion Power Plant. This is because most of the power is directed towards the outer divers, thereby mitigating relatively vulnerable internal targets.

Such a configuration is formed by overlapping the top and bottom magnetic shapes. When they are closely aligned, balanced power sharing between divers is achieved. However, even slight inconsistencies can result in upper or lower bias. Maintaining a balanced power sharing is expected to be particularly difficult for future nuclear reactors with more demanding requirements than current devices.

A dedicated experiment on Mast Upgrade, conducted in collaboration with Differ, was the first to investigate the dynamics of double-null configurations. The results were noticeable. We found that power sharing variations have a direct and immediate effect on diverter target conditions. Even when power sharing was modulated up to 200 times per second, fluctuations appeared in the target without delay. This means that there is virtually no time to counter such imbalances in the sledgear. Instead, you need to maintain a balanced configuration continuously. This is a major challenge for nuclear reactors employing such a configuration, highlighting the need for dedicated research over the next few years.

Core density control

The most promising fusion reaction for energy production on Earth lies between two flavors of hydrogen: deuterium and tritium (DT). When it collides with sufficient energy, it forms by fusing helium and neutrons. The probability of a collision between deuterium and tritium depends on the number of particles around it. More fuel to the same volume results in more fusion reactions with more denser mixtures. In this way, the fuel density of the core reactor volume determines the fusion power production. Therefore, to control the production of fusion capacity, the core fuel density of the reactor must be controlled.

The experimental mast upgrade reactor began operation in 2020 and then operated with a core density controller that did not meet your needs. This slowed the investigation of physics and exploitation of devices. Uncontrolled, core density often rises to levels where global instability leads to sudden loss of plasma (end of the experiment). At low density, the plasma legs that enter the diverter of the mast upgrade split are split into two, hit the sensor window, trigger the protection probe, and the control system ends the discharge.

To support the development of core density control, we applied the sine potential of the bolt to the gas valve injecting fuel. Next, a laser (interferometer) is used to measure the response of the core density. Only these measurements (data in the form of numeric rows) can be used to measure dynamic behavior. This data shows how long it takes for the core density to rise and how much it will rise when the gas valve is opened at a certain percentage. Use this information to determine the settings for the cruise control offline without any new experiments. Next, in the third experiment, we demonstrated control by following the desired density for cruise control. The control loop and experimental results are shown in Figure 3.².

Therefore, the development of density control in Mast-U was a necessary step in the development of exhaust control, as reported in a previous article.

Injection of frozen pellets

A gas valve was used to control the core density on the mast upgrade. However, in a full-fledged reactor, the edges of the core are very hot so the gas injected into the edges has little effect on the core density. Instead, the fuel from frozen pellets can be injected, penetrating through the edges into the core, increasing density where fusion capabilities are generated.

However, frozen pellets introduce many complications with regard to density control. The main thing among these is the individual nature of the pellet injection. In Itekamac, France’s largest tokamac, currently under construction, each frozen DT pellet constitutes at least 6% of the total number of plasma particles. These plasma density jumps easily exceed the limits of plasma edge density, leading to unacceptable plasma destruction and serious damage to the reactor in ITER. From a control point of view, the individual nature of the pellet injection also makes it difficult to control in real time.

We have developed a predictive controller (also known as Model Predictive Control, or MPC) that can handle the computational complexity of pellet decisions, while ensuring that plasma edge density does not exceed acceptable limits. To test this controller, we combined it with a high fidelity simulation of ITER plasma operation. The results are shown in Figure 4. This controller demo opened up several measures to further improve the robustness of core density control and calculation time.

Contribute to the UK programme on spherical tokamacs for energy production (steps)

All of this is put together in our work with UKAEA, the UK, which began in February 2022 with further development of exhaust management for the UK public-private project step. This was followed in 2024 by particle control. It mainly focuses on a step exhaust controller that is independent but simultaneously capable of controlling the peeling of two diverters in two null configurations.

This is not always guaranteed, so we also investigate the outcomes of disturbances that cause all heat to rise and fall. We also analyzed what would be the best set of actuators and sensors to achieve peeling. An important aspect of assessing whether an exhaust controller works is determining what obstacles arise from the burning core plasma. Therefore, the size and timescale of the disturbance have been studied to determine which controls are suppressed and which ones need to be buffered. Pellets were of particular interest as they are predictable sources of interference that can be suppressed using modern model predictive control techniques.

Both core control and exhaust control are important aspects of the realization of nuclear fusion as an energy source, which has been greatly increased by the development of high-field magnetic tokamaks.

B. Kool, TM van den Doel et al. , ‘Double Null Power Sharing Dynamics (Prelint), Nuclear Fusion (2025) in Mast-u. GL Derks et al. , Fusion Engineering and Design 202 (2024), 114387 J. Harrison, A. Aboutaleb, S. Ahmed, M. Aljunid, S. Allan, H. Anand, Y. Andrew, L. Appel, A. Ash, J. Ashton, 64, no. 11, p. 112017, 2024 Orrico, Christopher Anthony, et al., “Predicted density profile control with discrete pellets applied to integrated simulations of Iter,” Fusion (2025).

This article will also be featured in the 24th edition of Quarterly Publishing.