A new preprint details a computational optimization of an ARC-like compact tokamak, achieving a simulated fusion power increase of approximately 30% to 65% compared to a baseline design. The study employed Bayesian optimization to navigate a complex, multidimensional operating space defined by plasma composition, shape, and pedestal density. By systematically varying these parameters, the model identified a new operating point with significantly higher fusion power and power density. The findings, presented in a paper submitted to arXiv, highlight the potential for significant performance gains in high-field tokamaks through integrated core and pedestal optimization, without altering the physical device footprint. Source: arXiv

The optimization focused on several key levers that influence plasma performance. Researchers manipulated the effective ionic charge (Zeff), pedestal density, and plasma shaping parameters including elongation, triangularity, and squareness. The model captured the complex interplay between these variables, such as the stabilization of Ion Temperature Gradient (ITG) turbulent transport by increased impurity content. While higher Zeff can improve confinement, it also leads to fuel dilution, creating a trade-off that the optimization algorithm was designed to navigate. The study underscores the importance of squareness as a shaping parameter, which the authors note is a powerful but less commonly considered tool for enhancing performance. Source: arXiv

The optimization focused on several key levers that influence plasma performance.

A critical aspect of the analysis was the treatment of the pedestal pressure, which is constrained by magnetohydrodynamic instabilities. The model accounts for the separate impacts of plasma parameters on peeling and ballooning modes, which together define the pedestal stability boundary. The complicated dependence of these modes on the operational parameters makes simultaneous optimization challenging. This work is particularly relevant for compact, high-field devices like the ARC concept developed by Commonwealth Fusion Systems, where maximizing the pressure within a given magnetic field geometry is essential for achieving high fusion gain and economic viability. Source: arXiv

The initial optimization, holding elongation constant, found an operating point yielding ~30% more fusion power than the ARC V3A reference case. This was achieved through increased shaping and a higher Zeff. When elongation was also included as a variable in the optimization, the model identified a scenario with a ~65% increase in total fusion power. This more aggressive shaping pushes the plasma closer to the limits of stability but demonstrates a significant performance ceiling available through geometric controls. These results, derived from computational modeling, provide a roadmap for experimental campaigns on future devices aiming to maximize fusion output in steady-state operations. Source: arXiv