Update on the design of the Columbia Stellarator eXperiment

TL;DR

The paper presents the final CSX configuration, balancing neoclassical physics near quasi-axisymmetry with engineering constraints through an improved Boozer-surface optimization that now accounts for finite-build coil effects. A multi-filament finite-build model is developed to capture magnetic-field errors and HTS-tape strain, followed by a finite-build refinement that preserves QS quality while keeping strain within tolerances. Neoclassical predictions from SFINCS indicate reduced flow damping in CSX relative to CNT and favorable heat-flux characteristics, supporting experimental feasibility. Sensitivity analyses quantify tolerances and demonstrate experimental flexibility via IL-coil rotation, enabling controlled exploration of QA deviations, islands, and error-field effects. The work establishes a practical pathway toward constructing CSX, including robust optimization, detailed engineering constraints, and a clear plan for diagnostics and future studies of neoclassical transport and island physics.

Abstract

We present the final configuration chosen to be build for the Columbia Stellarator eXperiment (CSX), a new stellartor experiment at Columbia University. In a recent publication, Baillod et al. (NF, 2025) discussed in detail the different objectives, constraints, and optimization algorithms used to find an optimal configuration for CSX. In this paper, we build upon this first publication and find a configuration that satisfies all the constraints. We describe this final configuration including discussion of the coil finite build effects, sensitivity analyses, and the plasma neoclassical physics properties using the SFINCS code. These post-processing calculations provide a confirmation that the experimental goals of CSX can be achieved with the presented configuration.

Figures (15)

• Figure 1: CSX design scheme. The first section, highlighted in blue, is performed with a single-filament representation of the coils. The second section, in red, uses a multi-filament representation of the coils to better model the magnetic field generated by the coils.
• Figure 2: Scaling of normalized magnetic field error with increasing number of normal filaments.
• Figure 3: Comparison of HTS tape stack cross-sections with and without enforcing that $\mathbf{\hat{n}}_{\text{fb}} \perp \mathbf{\hat{k}}_{\text{HTS}}$.
• Figure 4: Visualization of filament placement for strain for MFC's. The filaments are placed in the center of the HTS tape stack in the $\mathbf{\hat{k}}_{\text{HTS}}$ direction, as required in paz-soldan_non-planar_2020.
• Figure 5: Strain for a CSX coil compared between a SFC and its MFC counterpart, before re-optimizing the MFC to reduce strain. The shaded area represents the variation between the minimum and maximum strain across the HTS tape stack. The maximum strain in certain filaments exceeds the strain threshold of the HTS tape.