Pre L-H Transition Radial Electric Field and Transport Validations of Edge and Scrape-off Layer Gyrokinetic Simulations at ASDEX Upgrade

This paper presents a stepwise validation of full-f gyrokinetic simulations using the GENE-X code for the ASDEX Upgrade tokamak, demonstrating excellent agreement with experimental radial electric field and transport profiles during the pre L-H transition phase and highlighting the critical roles of turbulence-driven flows and neutral gas ionization sources in reproducing edge plasma behavior.

The Gist

Imagine a fusion reactor as a giant, super-hot pot of soup (plasma) that we are trying to keep boiling without spilling over the sides. To get the most energy out of this soup, we want it to enter a special "high-confinement mode" (H-mode), where the heat stays trapped inside much better. But getting there is tricky; the soup has to cross a threshold, like a door that only opens if you push hard enough.

This paper is about building a super-accurate computer simulation to understand exactly what happens in the "kitchen" of the pot (the edge of the plasma) just before that door opens. The researchers used a powerful tool called GENE-X to simulate the ASDEX Upgrade tokamak, a real fusion experiment in Germany.

Here is the breakdown of their findings using simple analogies:

1. The "Step-by-Step" Cooking Method

Instead of trying to simulate the entire slow process of heating the soup from cold to hot in one go (which is very hard to get right), the researchers took a "step-by-step" approach. They looked at four specific moments in time as the heating power increased, stopping at each step to check if their simulation matched reality.

• The Analogy: Imagine taking a photo of a cake rising in an oven every few minutes. Instead of trying to predict the whole rise at once, they checked the cake at 2:30, 3:30, 4:30, and just before it was done. At each stop, they adjusted their simulation inputs to match what the real oven was doing.

2. The Invisible "Electric Wall" (The Radial Electric Field)

The most important thing they studied is something called the Radial Electric Field ($E_r$). Think of this as an invisible electric "wall" or "fence" that forms at the edge of the plasma.

• The Goal: For the plasma to switch into the high-performance mode, this electric fence needs to get very deep and strong (like a deep moat).
• The Discovery: The simulation showed that this "moat" gets deeper and deeper as the heating power increases, matching real-world measurements perfectly.
• The Secret Sauce: They found out why the moat gets deep. It's not just the pressure of the plasma pushing against the wall. It's mostly caused by turbulence-driven winds (poloidal flows) swirling around the edge. Imagine a whirlpool in a bathtub; the swirling water creates a depression in the center. The simulation showed that these turbulent swirls are the main reason the electric "moat" forms.

3. The Missing Ingredient: The "Gas Source"

In their first attempts, the simulation was a bit off. It predicted the density of the plasma (how crowded the particles are) was too low near the edge, and the heat escaping was too high.

• The Fix: They realized they were missing a crucial ingredient: neutral gas ionization. In the real world, cold gas from the walls gets hit by the hot plasma and turns into new particles (ionization).
• The Analogy: It's like baking a cake but forgetting to add the rising agent (yeast or baking powder). The cake wouldn't rise properly. By adding a "density source" to their code to mimic this gas turning into plasma, the simulation suddenly matched the real experiment. The plasma density profile looked right, and the heat escaping was no longer too high.

4. Turbulence: The "Storm" in the Soup

The edge of the plasma is a stormy place with tiny whirlwinds (turbulence) that try to carry heat away.

• The Battle: The researchers found two types of "storms" fighting for dominance: electron drift waves and trapped-electron modes.
• The Result: The "electron drift waves" were the main drivers of the chaos. However, when they added the "gas source" (the missing ingredient mentioned above), it smoothed out the density gradients (the steepness of the slope), which acted like a calm wind, stabilizing the storm and reducing the heat loss.

5. The Final Verdict: A Better Recipe

The paper concludes that their new, more complete simulation (which includes the whole edge and the "scrape-off layer" where particles escape) is a major success.

• Why it matters: Previous simulations were like looking at a small slice of the cake and guessing the rest. This new method looks at the whole edge self-consistently.
• The Achievement: They successfully predicted the depth of the electric "moat" and the amount of heat flowing out, matching the real machine's data very closely. This proves that their computer model is mature enough to help predict the "power threshold" needed to switch a future fusion reactor into its high-performance mode.

In summary: The researchers built a high-fidelity computer model of a fusion plasma edge. By adding a realistic "gas source" and tracking the swirling turbulent winds, they successfully recreated the formation of the critical electric field barrier that allows fusion reactors to operate efficiently. They didn't just guess; they validated their recipe against real experimental data at every step.

Technical Summary

Technical Summary: Pre L-H Transition Radial Electric Field and Transport Validations of Edge and Scrape-off Layer Gyrokinetic Simulations at ASDEX Upgrade

Problem Statement
Accurate prediction of turbulence and transport in the plasma edge and scrape-off layer (SOL) is a critical challenge for future fusion reactors, as these regions determine energy confinement and fusion power. While the transition from low-confinement (L-mode) to high-confinement (H-mode) is governed by edge turbulence and transport processes, the strong coupling between the confined plasma edge and the SOL, combined with the presence of magnetic X-points, challenges the applicability of local transport models. Previous validation studies using local, $\delta f$ gyrokinetic (GK) simulations (e.g., using the GENE code) successfully reproduced turbulent transport levels but relied on experimentally prescribed equilibrium profiles and radial electric fields ($E_r$) as inputs. Consequently, these local approaches cannot provide a predictive description of edge dynamics or the evolution of the $E_r$ well, which is central to regulating edge turbulence and the L-H transition power threshold.

Methodology
This work employs a stepwise validation strategy using first-principles, electromagnetic, full-$f$, and collisional GK simulations of edge and SOL turbulence, including X-point geometry. The simulations utilize the GENE-X code to model a dedicated hydrogen discharge (#38176) at the ASDEX Upgrade (AUG) tokamak.

• Simulation Strategy: Instead of simulating the slow temporal evolution of plasma profiles under prescribed power, the authors perform independent simulations for four successive experimental L-mode time slices ($t = 2.7, 3.8, 4.5, 4.8$ s) approaching the L-H transition. The inner edge boundary is fixed to experimental profiles, while edge and SOL profiles evolve self-consistently in response to turbulence, transport, and density sources.
• Density Source Modeling: To address the lack of edge density sources in standard models, an ad-hoc kinetic density source ($S_n$) is introduced near the separatrix to mimic neutral gas ionization. This source is tuned to match experimental density profiles near the separatrix at the final time slice ($t=4.8$ s).
• Boundary Conditions: Dirichlet boundary conditions are applied at the inner edge and SOL boundaries. The simulations are run to a quasi-steady state (QSS), where particle and energy fluxes balance the sources and sinks.
• Analysis: The study validates Outboard Midplane (OMP) profiles (density, temperature, $E_r$) against experimental measurements. It further decomposes the radial electric field using force balance analysis, characterizes edge turbulence via dispersion relations, and analyzes particle and heat transport fluxes.

Key Contributions and Results

1. Validation of $E_r$ Profiles: The simulations show excellent qualitative and quantitative agreement with experimental $E_r$ profiles, particularly the depth and width of the negative $E_r$ well near the separatrix. The inclusion of the density source ($S_n$) at $t=4.8$ s is shown to be essential for reproducing the experimentally observed deepening of the $E_r$ well, bringing the simulated minimum values ($\approx -15$ kV/m) close to H-mode measurements.
2. Force Balance and Poloidal Flows: A force balance decomposition reveals that while the edge region ($\rho_{pol} \lesssim 0.98$) is dominated by the ion pressure gradient and toroidal rotation, the $E_r$ well structure is governed by turbulence-driven poloidal flows. These flows deviate significantly from neoclassical (NC) predictions. The deepening of the $E_r$ well is directly linked to the strengthening of these nonlinear, turbulence-driven poloidal flows.
3. Role of Density Source: The introduction of the edge density source is critical for reproducing experimentally relevant density profiles and reducing the overestimation of electron and ion heat fluxes. The source acts as a localized temperature sink, flattening profiles and stabilizing turbulence, which leads to a reduction in turbulent transport.
4. Turbulence Characterization: Edge turbulence is identified as a competition between electron drift waves (eDWs) and trapped-electron modes (TEMs), with eDWs being dominant. The inclusion of the density source narrows the turbulence spectrum and shifts the peak to lower wavenumbers, consistent with collisional eDW dominance.
5. Transport Predictions:
   • Without Density Source: Simulations systematically overestimate heat fluxes, particularly electron heat flux ($P_e$), and fail to reproduce the correct density profiles near the separatrix.
   • With Density Source: The inclusion of $S_n$ reduces both ion ($P_i$) and electron ($P_e$) heat fluxes, yielding quantitative agreement with experimental ASTRA estimates (within 20% for $P_i$).
   • Diamagnetic Contribution: Close to the L-H transition, the ion heat flux ($Q_i$) arises not only from turbulent transport but also from a significant diamagnetic contribution (approx. 30% of total ion power). This highlights the necessity of a self-consistent treatment of turbulent and diamagnetic fluxes, which local flux-tube approaches cannot capture.
   • Scaling: In the absence of the density source, ion heat and particle fluxes follow a Gyro-Bohm scaling dominated by nonlinear $E \times B$ advection.

Significance and Claims
The paper claims that this work constitutes an important milestone toward predictive, first-principles gyrokinetic simulations of the L-H transition power threshold. By moving beyond local models to global, full-$f$ simulations, the study demonstrates that:

• Global edge-SOL turbulence simulations can reproduce critical L-mode physics ingredients, specifically the $E_r$ structure and ion heat fluxes ($Q_i$), which are believed essential for triggering the L-H transition.
• The evolution of the $E_r$ well is governed by nonlinear turbulence-mean flow interactions, a mechanism absent in simulations where $E_r$ is externally imposed.
• Accurate prediction of edge transport requires a consistent treatment of both turbulent and diamagnetic fluxes, as well as an accurate model for edge particle sources (neutral ionization).
• The results provide a valuable dataset for the development of reduced transport models (e.g., TGLF) by offering turbulent flux spectra from full-$f$ GK simulations.

The authors note that while the model captures key physics, future work must incorporate self-consistent neutral gas dynamics and finite Larmor radius (FLR) effects to further enhance physics fidelity. The study does not claim to simulate the actual dynamic L-H transition event but validates the pre-transition conditions required for it.