Real-time tracking of the turbulence-minimum state for optimizing plasma confinement in LHD

Abstract Turbulence-driven anomalous transport is one of the primary factors that degrade plasma confinement performance and remains a common challenge in both tokamaks and stellarators/heliotrons. In the Large Helical Device, the dominant turbulence mode—either ion temperature gradient or resistive interchange—varies according to the plasma operational regime, thereby determining the confinement characteristics (Kinoshita et al 2024 Phys. Rev. Lett. 132 235101). In particular, when the turbulent mode switches, both turbulence and anomalous transport are minimized. In this study, by employing an exhaustive search combined with a support vector machine, we successfully identified the condition under which this turbulence transition occurs as a function of externally controllable parameters. The results demonstrate that the transition boundary—representing the state where turbulence is minimized—can be effectively characterized by a linear relationship: n e ( 10 19 m − 3 ) = 4.25 T e ( keV ) − 5.36 . By implementing real-time feedback control to actively track this boundary, a sustained turbulence-suppressed state was achieved. This strategy provides a robust foundation for optimizing plasma performance in future fusion devices, where the synergy between real-time boundary tracking and transport-property exploitation can autonomously guide the plasma toward its highest confinement potential.