Estimating coil features from an equilibrium Preprint: AIP/123-QED
E. Rodríguez, W. Sengupta · 2026
Read Estimating coil features from an equilibrium Preprint: AIP/123-QED by E. Rodríguez, W. Sengupta (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-12339.txt.
Introduction
Rodríguez and Sengupta build coils for vacuum stellarator fields from equilibrium traits alone, through a current potential on flux surfaces. Level sets of the Boozer toroidal angle serve as proxy coils, and key complexity measures, above all coil non-planarity, prove strongly governed by local magnetic field traits. The approach promises a practical proxy for coil design and deeper insight into equilibrium-coil links.
Catalog record
| CREATOR | E. Rodríguez, W. Sengupta |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 16472 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.12339
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Estimating coil features from an equilibrium Preprint: AIP/123-QED E. Rodríguez, W. Sengupta Abstract We present an explicit theoretical framework for constructing artificial modular coils based solely on equilibrium properties, achieved through the formulation of a current potential defined on flux surfaces. We demonstrate that key measures of coil complexity –particularly coil non-planarity– are strongly governed by local magnetic field properties and show promise as predictors for more realistic coil configurations. This approach provides both a pathway towards deeper understanding of equilibrium–coil relationships and a potential practical proxy for coil design. I Introduction The stellarator 21 ; 1 ; 7 is an attractive magnetic confinement concept for achieving controlled thermonuclear fusion. Its flexibility and generality make it unique in its ability to confine plasmas while avoiding common shortcomings of other confinement designs 20 ; 1 . However, achieving good confinement in stellarators requires careful tailoring of the magnetic field geometry 18 . In practice, reproducing three-dimensional (3D) stellarator magnetic fields requires coils, or other current distributions, of considerable geometric complexity. This imposes a significant burden on stellarator design and necessitates careful consideration within the design process itself. A natural question therefore arises: which stellarator fields are particularly amenable to simple coils? A central difficulty in addressing such a question is the fundamental lack of uniqueness of the inverse coil problem: the same magnetic field within a volume can, in principle, be reproduced to arbitrary accuracy by many external current distributions. 14 ; 12 Thus, in practice, the problem is commonly approached as an optimisation task in which the field error is minimised. Despite substantial progress, particularly in optimisation strategies and regularisation techniques, 17 ; 10 ; 2 ; 5 ; 23 an important question remains: to what extent is a given current distribution solution intrinsic to the target magnetic field, and to what extent is it a consequence of the optimisation procedure and the imposed constraints? Recent insight into this question has been gained through the analysis of intrinsic magnetic field length scales, such as L_{\nabla\mathbf{B}} . 13 ; 9 However, such measures remain one step removed from the coil design problem, as they do not directly describe properties of the current distributions. In this Letter, we exploit a theoretical construction by which a coil set can be defined directly and uniquely for a given stellarator equilibrium. The construction is based on a surface current defined on a magnetic flux surface, thereby removing the inherent non-uniqueness of the coil problem. Although the resulting coil set is artificial, it directly encodes intrinsic properties of the magnetic field that can be leveraged to assess the behaviour of more realistic coil sets. We therefore propose it as a practical tool for coil assessment and as a framework for gaining deeper theoretical insight into stellarator coil design. II Constructing the coil set We begin from the standard Merkel formulation 17 of the coil problem. Let \mathbf{B} be a vacuum magnetic field in a toroidal volume \Omega (we leave the finite plasma \beta problem to future work), whose boundary \partial\Omega a flux surface, \mathbf{B}\cdot\hat{\mathbf{n}}=0 , with \hat{\mathbf{n}} the surface normal. Introducing a toroidal winding surface S enclosing \Omega , the Merkel problem consists of finding a scalar potential \Phi(\theta,\phi) as a function of the poloidal and toroidal angles, respectively, such that the surface current \mathbf{K}=\frac{1}{\mu_{0}}\hat{\mathbf{n}}\times\nabla\Phi, (1) reproduces \mathbf{B} in \Omega . In practice, this inverse problem is solved numerically via least-squares minimisation (e.g. NESCOIL 17 , REGCOIL 12 , QUADCOIL 4 ). Different solutions obtained in this way depend on the imposed regularisation 4 and the choice of S . The lack of uniqueness is due to the finite volume between S and \partial\Omega , but it can be eliminated by considering instead the limiting case, S\to\partial\Omega . In this limit, the winding surface coincides with a flux surface, and by direct application of the virtual casing principle 3 ; 6 , we have a closed form for the surface current \mathbf{K}=-\frac{1}{\mu_{0}}\hat{\mathbf{n}}\times\mathbf{B}, (2) evaluated on \partial\Omega . The surface current is simply orthogonal to the magnetic field everywhere and therefore introduces a discontinuity in its tangential component. It can be seen as a special case of [ 16 ]. That the above current reproduces \mathbf{B} follows directly from the vanishing of the exterior field. Applying Biot–Savart and writing \mathbf{B}=-\nabla\Phi_{0} in \Omega , one obtains \displaystyle\mathbf{B}_{\bar{\Omega}} \displaystyle=\frac{\mu_{0}}{4\pi}\int_{S}\mathbf{K}\times\nabla^{\prime}\left(\frac{1}{\rho}\right)\mathrm{d}S^{\prime} (3) \displaystyle=\frac{\mu_{0}}{4\pi}\int_{S}\hat{\mathbf{n}}\cdot\mathbf{B}\,\nabla^{\prime}\left(\frac{1}{\rho}\right)\mathrm{d}S^{\prime}, (4) which vanishes identically since \mathbf{B}\cdot\hat{\mathbf{n}}=0 on a flux surface. The tangential discontinuity thus enforces \mathbf{B} throughout \Omega . It is therefore imperative that S be a flux surface, since the problem in which S and \partial\Omega deviate in a non-trivial way cannot be treated so straightforwardly. We shall later see with a practical example what such deviations look like in practice. In this construction, then, the current potential is precisely the restriction of the vacuum scalar potential to the boundary, \Phi=\Phi_{0}|_{\partial\Omega} (up to a constant), and is unique. In fact, this is just a scaled version of the Boozer toroidal \varphi angle. If we then interpret this current density, as it is customary, as the limit of a set of filamentary coils (see an example in Figure 1 ), then the coils must be modular; i.e. , they close poloidally. Figure 1: Ideal coil set and principal curvature on last close flux surface. 40-coil construction computed using Eq. ( 2 ) for the precise QA stellarator 11 at \rho=1 . The surface colour represents the largest principal curvature of the flux surface, indicating the central role of surface curvature on controlling the coil curvature. III Coil geometry and intrinsic complexity The geometry of the resulting coils provides direct insight into their intrinsic complexity and how the equilibrium properties influence it. For coils defined as contours of constant \Phi on a flux surface, we can express their curvature as \displaystyle\tilde{\kappa}_{n}=2H-\kappa_{n},\qquad\tilde{\kappa}_{g}=\frac{1}{B}\,\hat{\tau}\hat{\tau}:\nabla\mathbf{B}, (5) where \tilde{\kappa}_{g} and \tilde{\kappa}_{n} are the geodesic and normal curvatures of the coils, \kappa_{n} the curvature of the magnetic field lines, \hat{\boldsymbol{\tau}}=\mathbf{B}\times\hat{\boldsymbol{n}}/|\mathbf{B}| and H is the mean curvature of S . 22 The total curvature of the coils is then by definition \tilde{\kappa}=(\tilde{\kappa}_{g}^{2}+\tilde{\kappa}_{n}^{2})^{1/2} . The normal curvature is primarily set by the surface geometry. Because MHD stability generally pushes magnetic field lines to have as negative a normal curvature as possible, but also due to their tendency to minimise bending energy, coils which are orthogonal to them tend to align with the direction of largest principal curvature \kappa_{1} ( i.e. , poloidally). The role of \kappa_{1} is illustrated in Figure 1 , and clearly shows why sharp surface features or extreme elongation are to be avoided. Projecting this behaviour outwards, we expect the normal curvature of the coils to drop as roughly \tilde{\kappa}_{n}\sim 1/\rho , the normalised radial flux coordinate, as the winding surface is pushed away from the plasma. The particulars of the winding surface shape will of course matter. Figure 2: Illustrations of geodesic curvature and non-planar behaviour of coils. a) Diagram showing a flux surface \partial S in the presence of a significant breathing, i.e. , toroidal variation due to change in toroidal magnetic field magnitude. It clearly shows the misalignment between curvature of the coil and the surface, and thus a significant geodesic curvature. b) Sketch of a flux surface patch with a