Adaptive RMP-ELM Control in DIII-D towards complete ELM suppression from the H-mode transition while maintaining good confinement

This work presents access to the high-confinement ELM-suppressed state [1] from the very beginning of the H-mode phase in the DIII-D tokamak using integrated feedback control of resonant magnetic field perturbations (RMPs). This achievement is made possible by combining an early RMP-ramp scheme [2] alongside the adaptive RMP [3,4] with n=3 RMP in ITER-similar shape plasmas (q₉₅~3.4). The confinement enhancement observed with such n=3 adaptive RMP exhibits distinct characteristics from the previous n=1 adaptive RMP [3] in that the rotation pedestal plays a crucial role in boosting confinement from standard pedestal recovery. The recovered rotation pedestal by adaptive RMP results in a favorable change of ExB flow, lowering the RMP threshold and reducing the turbulent core transport. Their synergy allows the maintenance of ELM suppression with a very low RMP amplitude (1.5 kA) and promotes strong confinement recovery up to H₉₈~1.5. This low-RMP ELM-suppressed state is advantageous in an ITER, as it minimizes the loss of fast particles and fusion gain. Lastly, the ongoing progress in advanced 3D control in DIII-D using an ML/AI surrogate model is also presented. [1] T. E. Evans et al., PRL 53, 1453, [2] T. E. Evans et al., PFR 12, 2402046, [3] S. K. Kim et al., NF 62, 026043, [4] R. Shousha et al., POP 29, 032514