Engineers at the Lawrence Livermore National Laboratory made history in 2022 by successfully demonstrating fusion ignition, marking a crucial milestone by generating more fusion energy than what was originally delivered through traditional laser means. This landmark achievement has reignited optimism regarding the future of fusion energy, a field often criticized for its protracted timeline toward commercial application.
In a recent survey conducted by the US Clean Air Task Force, the integration of artificial intelligence (AI) with fusion energy development was prominently featured, revealing a collaborative synergy that could significantly alter the trajectory of energy production in the coming years. According to Sehila Gonzalez de Vicente, global director of Fusion Energy at the Task Force, AI is progressing at an unprecedented pace and meets the growing demand for clean energy, making fusion an appealing solution. The report posits that AI could accelerate fusion technology development while fusion itself offers a much-needed emission-free energy source to support advancing AI capabilities.
The report highlights five pivotal technologies where AI and high-performance computing (HPC) can propel progress: materials discovery, tritium breeding technologies, superconducting magnets, inertial fusion energy technology, and advanced diagnostics and control systems. It includes a collection of case studies showcasing various organisations that are leveraging AI to hasten the pace of fusion development.
A notable example comes from Commonwealth Fusion Systems, an enterprise spun out from the Massachusetts Institute of Technology in 2018. The firm is employing HPC technology to model its SPARC fusion demonstration plant, enhancing the efficiency of particle transport and fluid mechanics simulations. Principal engineer Jaydeep Deshpande indicated a marked shift from several weeks required for a neutronics model down to several days, which marks a significant improvement in design insight and iteration. The company's chief operating officer, Rick Needham, asserted that they aim to achieve a net energy gain by 2027, effectively reducing the time frame from 30 years to a mere 30 months.
In the UK, nTtau Digital is providing generative AI tools for fusion engineers to simulate interdependencies within fusion systems. Their software optimises tritium breeding while minimising coil design costs, paving the way for innovative designs that were previously unexplored. This technology is expected to integrate with Nvidia's Omniverse platform, enabling collaborative design in a virtual space, which organisations like the UK Atomic Energy Agency (UKAEA) are exploring to streamline the design process without physical prototyping.
Another emerging player in the field, Focused Energy, is concentrating on laser-driven fusion plants and utilising generative AI for cognitive simulation. By developing sophisticated models that consider long-term impacts of design decisions, the company hopes to expedite the transition towards viable power plants.
The role of AI extends to real-time control within fusion environments, as validated by Google’s AI division, DeepMind. In 2022, they utilised reinforcement learning algorithms to create the first AI-run magnetic control system in a nuclear fusion reactor, demonstrating AI's effectiveness in responding to the dynamic and complex nature of plasma conditions in a tokamak.
The increasing energy demands of AI and cloud computing are prompting tech giants to explore fusion as a potential power source. Lindsay McQuade, director of energy at Amazon Web Services, highlighted the urgency of identifying new energy solutions, citing predictions of a doubling in electricity consumption from data centres and AI by 2026—enough to power an entire nation like Japan.
Market interest in fusion technology is noticeably heating up, with the number of companies involved in fusion quadrupling from 10 to 40 since 2017. However, the uncertain cost of developing commercial fusion plants poses risks; a recent MIT study suggested a broad potential future contribution from fusion to global electricity generation could range between 10% to 50% by the year 2100.
Investment plays a critical role in this evolving landscape. Total equity investments in the fusion sector have exceeded US$8 billion, with US$2 billion directed towards Commonwealth Fusion Systems, which is striving to bring its technology to commercialisation by the early 2030s. Recent collaborations, such as a US$3 billion deal between Microsoft and Constellation Energy regarding a nuclear reactor in Three-Mile Island, illustrate the intersection of technology and energy that is being actively pursued.
As the fusion industry embarks on this innovative journey, the interplay between AI and fusion technology underscores a transformative opportunity that can impact both energy production and the broader tech landscape.
Source: Noah Wire Services