Design and construction of the MUSE permanent magnet stellarator

Abstract

This paper documents the design and construction of MUSE, the world's first permanent magnet (PM) stellarator and the first quasi-axisymmetric experiment. The purpose of MUSE is to develop and assess a new way of building optimised stellarators that uses simple planar coils PMs. Our PM optimisation algorithm consists of initialising a geometry to pack dipoles densely, running the FAMUS code to minimise surface field error subject to PM constraints and applying discrete jumps to reach a physically realisable solution. FAMUS treats the PM system as a set of ideal point dipoles. From there we construct finite-volume magnet towers to be housed in 3D-printed PM holders. We describe the design of the PM holders, which were validated by laser metrology. We analyse the effects of finite permeability, sensitivity to perturbations and magnetostatic forces. An exact analytic formula for the magnetic field from a finite-volume PM tower is presented to compute PM–PM forces and stress on the PM holder. Stellarator construction is complete and experiments are underway.

Quasi-single-stage optimization for permanent magnet stellarators

“…The MUSE team has made a tremendous effort to improve the final design with many refinement techniques. Details are documented in [26]. Here, we are not going to adopt the same procedure, as it is time-consuming and requires a lot of manual refinements.…”

“…In addition, the MUSE magnets were designed to avoid specific regions to allow diagnostic access to the plasma, including vertical views. To distinguish from the original MUSE PM design, we shall call the new PM design 'MUSE-0' since it is similar to one of the generation-0 solutions in figure 7 of [26]. The comparisons of PM among MUSE++, MUSE-0, and MUSE are listed in table 2.…”

“…All the methods assume a target plasma equilibrium which is fixed during the optimization of permanent magnets. In other words, PM stellarators, like PM4Stell [23,24] and MUSE [25,26], are based on the two-stage optimization approach and therefore face similar difficulties. Some equilibria might require numerous PMs or cannot be realized by available PMs.…”

“…We apply our quasi-single-stage optimization approach to obtain a new quasi-axisymmetric PM stellarator. It starts from MUSE [25,26]. MUSE was originally designed with the two-stage optimization approach [25] and is the first completely constructed PM stellarator in the world [26].…”

“…It starts from MUSE [25,26]. MUSE was originally designed with the two-stage optimization approach [25] and is the first completely constructed PM stellarator in the world [26].…”