Uncertainty hints at a surer path to clean fusion energy

Fusion has been described as the key to producing carbon-free energy at scale but the degree of uncertainty involved has been a deterrent to research because of the likely time and cost.

At issue is the turbulent behaviour of plasma at over 150 million °C, which bleeds energy and impedes stable reactions. Predicting turbulence reliably is essential, but traditional computational models have struggled to predict turbulence with sufficient certainty and have required demand millions of CPU hours to undertake the challenge.

However, AI firm digiLab’s Uncertainty Engine has enabled the construction of models that quantify what the company dubs the “known unknowns” in plasma behaviour even in the absence of complete data.

Operating in conjunction with UKAEA’s Spherical Tokamak for Energy Production (STEP) programme, the technology has helped save hundreds of thousands of CPU hours, said the firm.

Head of business development at digiLab Amanda Niedfeldt remarked: “With digiLab’s Uncertainty Engine, we’ve helped UKAEA create fast, explainable models that don’t just make predictions, but also quantify uncertainty making data actionable, even when it's sparse.”

“Some examples of the powerful impacts this unlocks are drastically faster simulations, fewer redundant runs, and smarter diagnostic decisions, which accelerates design programmes and prevents costly late-stage redesigns.” 

Adam Stephen, UKAEA head of advanced control unit described the success as more than a technical breakthrough.

“It is a proof point that uncertainty-aware AI can unlock progress where the stakes are highest. For fusion, that means bringing humanity closer to a future of limitless, clean energy,” he explained.

“For other industries, it signals a new way to manage complexity, reduce risk and accelerate innovation.”

Pic: MAST-U fusion energy machine at UKAEA Culham Campus (SMD Photography)