Introduction: unlocking stable fusion through smarter control
Fusion energy promises clean, abundant power, but turning that promise into reliable reactors requires mastering how heat and fuel behave inside the machine. Researchers working with the UK’s MAST Upgrade (MAST-U) and the Dutch research institute DIFFER have made significant strides in two critical areas: how to spread intense heat evenly across the reactor and how to regulate the core fuel density. Their work informs future devices such as the STEP project in the UK and the international ITER project, bringing us closer to practical fusion power plants.
Power sharing and the double-null configuration
One major hurdle is distributing the enormous power from the fusion core to the vessel wall so that no single component bears an excessive load. A promising approach is the double-null configuration, where an extra divertor is added at the top of the machine. In this setup, the magnetic geometry is harmonised to steer heat and particles toward both the upper and lower divertors, ideally achieving balanced power sharing.
Researchers at MAST-U, in collaboration with DIFFER, studied the dynamics of this configuration. They found that even tiny shifts in power-sharing can instantly alter divertor conditions. Disturbances can propagate to the targets at high speed, and fluctuations can occur up to 200 times per second with no delay in the divertor response. This means future reactors will require continuously maintained balance rather than reactive adjustments, presenting a formidable control challenge for devices that aim to operate in a double-null configuration over long periods.
Core density control: from gas valves to predictive modeling
The dominant fusion reaction in terrestrial devices involves deuterium and tritium (DT). The rate of fusion scales with how densely the fuel particles populate the plasma core. Controlling core density is thus essential to maintaining stable operation and predictable power output.
In MAST-U, a basic core-density control using a gas valve faced limitations, often allowing the core density to drift into regimes that trigger global instabilities or protective shutdowns. A novel approach replaced ad hoc adjustments with a closed-loop control system that uses sinusoidal voltage inputs to the fuel-gas valve and real-time interferometric measurements of core density. The data feed a model that characterises how density responds to valve changes, enabling a “cruise control” strategy to reach and sustain target density levels offline before running experiments. The result was a substantial improvement in experimental reliability—about 20% more successful runs without unexpected terminations—and a foundation for more advanced exhaust-control schemes.
Looking further ahead, gas fueling at the edge becomes less effective in full-scale reactors where edge temperatures are extreme. The team explored injecting frozen DT pellets, which penetrate the edge and increase core density more directly. However, pellets introduce discretization: each pellet causes a sizeable density jump, risking edge limits and possible disruptions. To manage these challenges, researchers developed predictive control methods, including model predictive control (MPC), to decide pellet timing and size while safeguarding edge density. They validated MPC with high-fidelity ITER-style simulations, demonstrating its potential to stabilize core density with discrete fueling actions and reduced computational cost—an essential step toward robust control in future devices.
Connecting core control to exhaust and detachment in STEP
The control strategies developed for MAST-U feed into the UK’s STEP initiative, led by UKAEA, which seeks to demonstrate a pilot fusion power plant. The research focuses on exhaust control within a double-null geometry to independently and simultaneously control detachment (the cooling of the exhaust plasma) in the upper and lower divertors. A key concern is how disturbances originating from the burning core affect diversion and detachment. By analyzing disturbance size and timescales, scientists are identifying which perturbations can be suppressed by modern control algorithms and which must be buffered by reactor design. Pellets, as a predictable disturbance, provide a useful test case for developing robust MPC-based exhaust controllers.
Why these advances matter for ITER and beyond
Although ITER operates at a different scale and with different design choices, the underlying control challenges are shared: preventing instabilities, distributing power safely, and maintaining steady fusion reactions. The integration of predictive density control, advanced pellet-fueling strategies, and continuous double-null power sharing lays the groundwork for reliable operation of future reactors. The close collaboration between MAST-U and DIFFER demonstrates how cross-institutional experiments can reveal dynamic plasma behaviours that must be countered with real-time, high-precision control systems.
Looking ahead: a path to scalable, safe fusion power
As fusion research progresses, the fusion community will increasingly rely on sophisticated control architectures that couple physics models with fast diagnostics and robust actuators. The lessons from MAST-U and DIFFER show that precise magnetic-field control, continuous power-sharing, and predictive core-density management are not optional enhancements—they are essential foundations for tomorrow’s fusion power plants, including STEP and ITER’s broader mission to deliver practical fusion energy.
Key takeaways
- Tiny magnetic-field changes can instantly affect divertor conditions in double-null configurations, necessitating continuous, precise control.
- Core-density control using real-time measurements and predictive modeling improves experimental reliability and enables advanced exhaust control planning.
- Pellet injection introduces discrete density steps; model predictive control helps manage these jumps, improving safety for large-scale devices like ITER.
- Integrating core and exhaust control strategies supports the feasibility of future fusion power plants such as STEP and multi-institution ITER operations.
