Metal Casting of a Stellarator Shell for Experimental Evaluation of the Thickness Variation Method

Abstract

Current stellarator designs rely on complex coil systems that carry high costs, hindering experimental stellarator research. To reduce costs, a novel stellarator design concept known as the thickness variation method aims to produce a coil-free stellarator by constructing a continuous toroidal shell of varying thickness. In practice, the approach demands significant geometric precision and material uniformity, posing a key manufacturing challenge. This work addresses this by developing a manufacturing process designed to produce a thickness-varying toroidal shell stellarator. An example stellarator geometry was first established, and a sectional shell prototype was designed as the target for fabrication. Direct cavity mold casting using PET-CF and plaster molds was identified as the most promising manufacturing method, offering geometric accuracy alongside advantages in accessibility, rapid production, and low cost. To improve material compositional uniformity, diffusion during casting was mitigated by drying the plaster molds and casting in a nitrogen atmosphere. A sample casting was halved and found to contain no internal voids. Laser inspection of half-shells produced using the same manufacturing process revealed thickness errors of 3.36% and 1.98% on the PET-CF-molded and plaster-molded surfaces, respectively. The experimental feasibility of the shell prototype was then evaluated through simulation, confirming that the shell can generate a sufficiently strong magnetic field and withstand the thermal demands of relevant experiments. These results demonstrate the ability of the proposed manufacturing approach to produce geometrically accurate thickness-varying toroidal shells, supporting its application to future research on the thickness variation method.