Abstract
High fidelity kinetic equilibria are crucial for tokamak modeling and analysis. Manual workflows for constructing kinetic equilibria are time consuming and subject to user error, motivating development of several automated equilibrium reconstruction tools to provide accurate and consistent reconstructions for downstream physics analysis. These automated tools also provide access to kinetic equilibria at large database scales, which enables the quantification of general uncertainties with sufficient statistics arising from equilibrium reconstruction techniques. In this paper, we compare a large database of DIII-D kinetic equilibria generated manually by physics experts to equilibria from the CAKE and JAKE automated kinetic reconstruction tools, assessing the impact of reconstruction method on equilibrium parameters and resulting magnetohydrodynamic (MHD) stability calculations. We find good agreement among scalar parameters, whereas profile quantities, such as the bootstrap current, show substantial disagreement. We analyze ideal kink and classical tearing stability with DCON and STRIDE respectively, finding that the $\delta W$ calculation is generally more robust than $\Delta^\prime$. We find that in $90\%$ of cases, both $\delta W$ stability classifications are unchanged between the manual expert and CAKE equilibria.
Introduction
To perform advanced physics simulations of tokamak plasmas, researchers require a "kinetic equilibrium"—a 2D cross-sectional map of magnetic fields and pressure profiles. Traditionally, these reconstructions are hand-tuned by experts to match diagnostic data, a process that is both time-consuming and subject to user bias. Automated tools like CAKE (Consistent Automatic Kinetic Equilibrium) and JAKE have been developed to standardize this process and enable large-scale database analysis. However, a critical question remains: how do the numerical differences between manual and automated reconstructions impact the predictive physics models that rely on them?
Methodology
This study utilizes a large-scale database comparison from the DIII-D tokamak. The authors compared a historical control dataset of 1,336 manually tuned equilibria against automated reconstructions generated by CAKE and JAKE. The analysis focused on two main areas:
- Parameter Comparison: Assessing scalar parameters (e.g., plasma current, major radius) and profile quantities (e.g., pressure, safety factor $q$, and current density).
- MHD Stability Analysis: Feeding these equilibria into the DCON code (for ideal kink stability, $\delta W$) and the STRIDE code (for classical tearing mode stability, $\Delta^\prime$).
Key Insights: Profile Sensitivity
A primary finding is that while global plasma shape parameters are robust across methods, profile-sensitive metrics vary significantly. Parameters relying on local gradients or edge data, such as the bootstrap current and core safety factor ($q_0$), are highly sensitive to the chosen reconstruction method and internal algorithmic settings like spline knot placement.
Results: Parameter Agreement and Stability Robustness
Statistical analysis shows that CAKE provides a much closer match to manual expert reconstructions than JAKE. Scalar parameters like major radius ($R_0$) and plasma current ($I_p$) show excellent agreement across the database, whereas edge pressure ($P_{edge}$) and core pressure ($P_{core}$) exhibit more scatter.
In terms of downstream physics, ideal kink stability ($\delta W$) proved relatively robust, with a 91% agreement in stability classification between manual and CAKE equilibria. In contrast, classical tearing stability ($\Delta^\prime$) is extremely brittle; the STRIDE code agreed on classification only ~75% of the time, with stability parameter magnitudes varying by orders of magnitude for the same plasma shot due to slight profile differences.
Conclusion
The sensitivity of tearing stability models to reconstruction nuances underscores the need for Uncertainty Quantification (UQ) in tokamak modeling. Because profile quantities are heavily influenced by reconstruction choices, the authors advocate for running downstream physics models on ensembles of equilibria to ensure that scientific conclusions are robust against reconstruction variance.
Summary updated