Modeling Temperature Profiles in the Pedestal of NSTX with Reduced Models

This paper presents a predictive modeling framework for NSTX H-mode pedestal profiles by coupling the ASTRA transport solver with neoclassical and reduced gyrokinetic models, revealing that while neoclassical and ETG turbulence dominate specific channels, KBM/MHD-like modes are essential for accurately capturing transport in both ion and electron thermal channels.

The Gist

Imagine you are trying to keep a pot of soup boiling perfectly on a stove. You want the soup to be hot enough to cook the ingredients, but not so hot that it boils over and spills everywhere. In the world of nuclear fusion, scientists are trying to do the exact same thing, but instead of a pot of soup, they are trying to contain a super-hot ball of gas (plasma) inside a magnetic bottle.

The "pedestal" mentioned in this paper is like the rim of the pot. It's the critical edge layer where the super-hot plasma meets the cooler, empty space outside. If this rim is too weak, the heat escapes, and the fusion reaction dies. If it's too strong, the pressure builds up until the whole thing explodes (a "disruption").

The scientists at the University of Texas and Princeton wanted to build a better recipe book to predict exactly how hot this rim should be in their specific fusion machine, called NSTX.

Here is how they did it, broken down into simple concepts:

1. The Problem: The "Recipe" Was Missing Ingredients

For a long time, scientists had a basic recipe for the rim, but it kept failing. They could predict the shape of the heat, but the numbers were always wrong. It was like trying to bake a cake with a recipe that only listed flour and sugar, but forgot the eggs and baking powder. The cake would look like a cake, but it wouldn't rise properly.

They realized the "rim" of the plasma is controlled by three different types of "chaos" (turbulence) happening at the same time:

• The Electron Chaos (ETG): Tiny, fast-moving electrons creating ripples.
• The Ion Chaos (KBM/MHD): Bigger, slower-moving heavy particles (ions) creating waves.
• The "Old School" Physics (Neoclassical): The standard, predictable way particles bump into each other.

2. The Experiment: Building a Better Simulator

The team built a computer simulation (a "digital twin" of the machine) to test how these three types of chaos interact. They used a tool called ASTRA, which is like a sophisticated video game engine for physics.

They started by testing just one type of chaos (the Electron Chaos).

• The Result: It worked okay for the electrons, but the heavy ions (the "meat" of the soup) got way too hot. The simulation predicted the ions would be scorching hot, but in the real machine, they were cooler.
• The Analogy: It was like turning on the stove burner for the soup, but forgetting to open the lid to let steam escape. The heat built up too fast.

3. The Breakthrough: Adding the "Heavy" Chaos

They realized they were missing the Ion Chaos (specifically something called Kinetic Ballooning Modes, or KBM).

• The Analogy: Imagine the plasma rim is a balloon. The electron chaos is like tiny pinpricks, but the ion chaos is like someone squeezing the balloon from the outside. If you don't account for the squeezing, the balloon (the temperature) gets too big.
• They created a new "surrogate model" (a smart shortcut) based on complex supercomputer simulations to predict exactly how much this "squeezing" happens.

4. The Final Recipe: A Perfect Balance

When they combined all three ingredients—the electron chaos, the ion chaos, and the standard physics—their simulation finally matched the real machine perfectly.

• The "Secret Sauce": They found that the Ion Chaos is the main thing holding back the temperature from getting too high. It acts like a safety valve.
• The "Fine Tuning": They also found that the electron chaos is very sensitive to how dense the gas is at the edge. If the density is low, the electron chaos goes wild.
• The "One Magic Number": Amazingly, they only had to tweak one single number in their entire model to make it work for two completely different types of fusion shots (one with a wide rim, one with a narrow rim). This suggests they finally found the fundamental rule that governs the plasma's edge.

Why Does This Matter?

Think of this paper as the moment a chef finally figures out the exact temperature and timing to bake the perfect soufflé.

Before this, scientists were guessing. Now, they have a reliable tool that can predict how the "rim" of the plasma will behave. This is crucial for the future of fusion energy (like the upcoming NSTX-U machine). If we can predict the rim perfectly, we can design machines that run hotter, longer, and safer, bringing us one step closer to unlimited clean energy.

In short: They stopped guessing and started measuring the "chaos" at the edge of the fusion fire. By understanding how the tiny electrons and the heavy ions dance together, they finally wrote a recipe that works.

Technical Summary

1. Problem Statement

Accurately predicting H-mode pedestal temperature profiles is critical for forecasting the performance of magnetically confined fusion plasmas. While macroscopic MHD instabilities (like peeling-ballooning modes) define global pedestal limits, turbulence driven by micro-instabilities (such as Electron Temperature Gradient (ETG) modes, Kinetic Ballooning Modes (KBM), and Microtearing Modes) regulates the internal transport.

• The Challenge: Existing predictive capabilities for spherical tokamaks (like NSTX) are limited. Previous approaches often relied solely on KBM constraints or adapted the EPED model, failing to fully capture the complex, multi-scale interactions between electron and ion transport channels.
• Specific Gap: There is a lack of a fully validated, integrated modeling framework that couples reduced gyrokinetic models for both electron-scale (ETG) and ion-scale (KBM/MHD) turbulence with neoclassical transport to predict time-evolving temperature profiles in spherical tokamaks.

2. Methodology

The authors developed a new modeling capability by integrating reduced gyrokinetic transport models into the ASTRA (Automated System for TRansport Analysis) 1.5D transport solver. The study focused on two representative NSTX discharges with contrasting characteristics:

• Shot 132543: Non-lithiated, ELMy, narrow pedestal (analyzed at 700 ms, post-ELM recovery).
• Shot 132588: Lithiated, ELM-free, wide pedestal (analyzed at 650 ms, steady state).

Key Modeling Steps:

1. Fixed Density: To isolate thermal transport mechanisms, electron and ion density profiles were held fixed based on experimental data, while temperature profiles ($T_e$ and $T_i$) were evolved dynamically.
2. Reduced ETG Model: A simple algebraic reduced model for ETG turbulence (from Ref. [20]) was used. It was validated against nonlinear GENE and CGYRO simulations. The study found the baseline model underestimated heat flux by a factor of ~2 for NSTX but retained the unscaled version initially to test sensitivity.
3. Quasi-linear Surrogate for KBM/MHD: To address missing ion transport, a quasi-linear surrogate model was developed.
   • Database: Linear gyrokinetic simulations were performed using GENE across a database of equilibria reconstructed with SPIDER and CHEASE.
   • Stability Scan: Growth rates ($\gamma$) were mapped on $s-\alpha$ diagrams (magnetic shear vs. normalized pressure gradient) for various toroidal mode numbers.
   • Surrogate Construction: A Radial Basis Function (RBF) interpolator was trained on this database to predict transport diffusivity ($\chi_{mix}$) as a function of local $T_e$, $T_i$, and magnetic shear.
4. Coupled Evolution: The model coupled electron and ion channels via electron-ion collisional power transfer ($P_{ei}$).
5. Refinements: The study tested the impact of:
   • Scaling the ETG transport by a factor of 2 ($2\times Q_{ETG}$).
   • Explicit $E \times B$ shear suppression factors.

3. Key Contributions

• Integrated Transport Framework: Successfully coupled reduced ETG, neoclassical, and ion-scale KBM/MHD models within a time-dependent transport solver for spherical tokamaks.
• Surrogate Modeling: Developed a robust, physics-based quasi-linear surrogate model for KBM/MHD transport derived from a comprehensive linear gyrokinetic database, requiring only one free parameter ($c_0 = 0.0008$) calibrated to a single discharge.
• Mechanism Identification: Systematically quantified the relative roles of different transport mechanisms, revealing that neoclassical transport dominates the ion channel across the entire pedestal, while ETG and KBM/MHD modes provide critical anomalous transport in specific regions.
• Robustness Analysis: Demonstrated that the model is resilient to variations in density profiles and that the transport is "stiff," meaning profiles are tightly constrained by marginal stability boundaries.

4. Key Results

• Neoclassical Dominance (Ion Channel): Neoclassical transport is the dominant mechanism across the entire pedestal for ions. However, neoclassical transport alone is insufficient to explain experimental ion temperatures, leading to massive overpredictions when ETG is the only other mechanism included.
• ETG Role (Electron Channel): ETG turbulence is highly localized to the plasma edge and low-density gradient regions ($\eta_e \gg 1$). It contributes substantially to electron transport but cannot account for the full profile without ion-scale transport.
• Necessity of KBM/MHD: The inclusion of the KBM/MHD surrogate model was essential. It provides the missing anomalous ion transport and contributes significantly to the electron channel as well.
• Profile Prediction Accuracy:
  • Baseline (ETG + Neo): Failed to reproduce experimental profiles, overpredicting both $T_e$ and $T_i$ significantly.
  • Full Model (ETG + Neo + KBM): With the KBM surrogate and a $2\times$ scaling of the ETG model, the simulations achieved good quantitative agreement with experimental $T_e$ and $T_i$ profiles for both discharges from the pedestal top to the separatrix.
• Sensitivity:
  • Scaling: Doubling the ETG transport or adding $E \times B$ shear suppression resulted in only minor, quantitative adjustments to the profile shape, not qualitative changes. The system is stiff; small changes in transport coefficients are compensated by slight adjustments in the temperature gradient.
  • Density: The model showed high resilience to $\pm 20\%$ variations in density gradients due to a cancellation effect between turbulent drive changes and collisional energy exchange ($P_{ei}$).
• Discharge 132543 Discrepancy: The model slightly overpredicted temperatures for Shot 132543. This was attributed to the fact that the experimental data was taken during a transient post-ELM recovery phase (non-steady state), whereas the solver assumes a steady state.

5. Significance and Future Outlook

• Predictive Capability: This work lays the foundation for predictive modeling of future devices (e.g., NSTX-U, STEP). The framework captures dominant physics with minimal empirical tuning (only one free parameter).
• Physics Insight: The study confirms that accurate pedestal modeling requires a coupled approach where ion-scale turbulence (KBM) and electron-scale turbulence (ETG) act complementarily, constrained by the density profile shape.
• Future Work:
  • Extending the model to predict density profiles (currently fixed).
  • Replacing heuristic surrogate components with direct physics-based closures.
  • Validating against other spherical tokamaks (e.g., MAST) and investigating low-recycling scenarios in NSTX-U.
  • Evaluating Microtearing Mode (MTM) contributions, though these are expected to be secondary to ETG and KBM.

In summary, the paper demonstrates that a reduced-model approach, grounded in first-principles gyrokinetic simulations and coupled with a transport solver, can successfully predict complex pedestal temperature profiles in spherical tokamaks by balancing neoclassical, ETG, and KBM transport mechanisms.