Description
The physics underlying the spontaneous formation of edge transport barriers above a certain power threshold and the associated transition to improved confinement regimes remains partially understood, despite its critical importance for future fusion reactors. Access to such regimes is known to depend on the complex interplay of a multitude of factors, making the accurate global simulation of the plasma turbulent dynamics during the transition a demanding challenge.
In this contribution, 3D turbulent simulations of the COMPASS tokamak boundary plasma are carried out using the first-principle drift-reduced fluid code SOLEDGE3X. The magnetic equilibrium reproduces the shot #16515 (ohmic L-mode D-shaped discharge), and neutrals are self-consistently included through the coupling with the kinetic code EIRENE.
To address this challenge, a power scan starting from sub-threshold L-mode conditions is conducted. During the power ramp-up (from 50 kW to 200 kW), strong radially localized axisymmetric poloidal flows are observed to develop at the low field side, oscillating close to the theoretical geodesic acoustic mode frequency. Remarkably, once the pressure gradient reaches a certain threshold, a sudden deepening of the radial electric field well and a consequent steepening of the upstream thermodynamic profiles near the separatrix are observed. This occurs only after the introduction of core particle fueling, enabling the pressure to reach values otherwise unattainable by neutral puffing and recycling alone. During the well formation, decomposing the radial force balance reveals that the poloidal flow contribution $-v_\theta B_\phi$ is substantial, comparable in magnitude to the diamagnetic term $\nabla p_i/(eZn_i)$ in the region where the well forms. Concomitantly, the stabilization of strong poloidal flows close to the separatrix is found to correlate with an increase in the peak of the electric Reynolds stress $\langle \tilde{u}_{E,\psi} \tilde{u}_{E,\theta} \rangle$.
We discuss the impact of the core particle source on the barrier onset and the discrepancy observed in the case without the source. The consistency of these results with experimental expectations is critically examined, finally focusing on the adequacy and limitations of the electrostatic fluid model in reproducing these dynamics.
These results aim to contribute to a better understanding of the physics behind L-H transition.