Multi-device investigation and validation of density limit and plasma energy balance models for tokamak disruption prediction in DECAF*

Increasing plasma density in tokamaks improves fusion power output, but can lead to a disruption if a density limit is exceeded. The commonly used empirical threshold known as the Greenwald Limit is insufficiently stringent to accurately predict and avoid disruptions, especially for highly shaped plasmas employing neutral beam injection heating. Research continues to find a density limit that would allow tokamaks to reliably operate closer to the threshold, maximizing fusion power without risk of disruption. Several theoretical underpinnings of the density limit have been proposed, such as models based on radiative power balance, radiative islands [1], and microinstabilities [2]. In this study, we use the DECAF [3] approach to evaluate the efficacy of these models against databases with large numbers of plasmas from multiple machines, including MAST-U, NSTX, and KSTAR. The predictive capacity of these models is evaluated by assessing whether the crossing of the density limit results in a disruption exhibiting physical manifestations characteristic of density limits, such as multifaceted asymmetric radiation from the edge (MARFE) events, high radiative power balance fractions, and collapse of key plasma profiles. Analysis of chains of DECAF Events leading to disruption shows that chains initiated by impurity radiative collapses (IRC) are long and relatively slow compared to the constituent Events. This suggests that IRC Events are often not the direct cause of disruptions, but rather a trigger for more complex disruption dynamics.

[1] D. Gates and L. Delgado-Aparicio, PRL 108 (2012) 165004; doi: 10.1103/PhysRevLett.108.165004

[2] M. Giacomin et al., PRL 128 (2022) 185003; doi: 10.1103/PhysRevLett.128.185003

[3] S. A. Sabbagh et al., Phys. Plasmas 30 (2023) 032506; doi: 10.1063/5.0133825