Abstract: Stellarators can be designed in such a way as to provide plasma confinement without any contribution from the plasma itself. They are inherently steady-state, and mostly free from disruptions. However, these advantages as compared to other confinement devices come at the cost of increased technological complexity, and a design parameter space which has a much higher dimension.
The high dimensionality of the design parameter space can be viewed as an opportunity to rely predominantly on numerical optimization to find particularly attractive reactor designs. But it can only be a practical strategy for the rapid development of commercial fusion if one possesses high performance numerical tools to efficiently explore the parameter space, and optimize over it.
In this talk, I will present new solvers developed with that goal in mind, as part of the ongoing Simons collaboration on Hidden Symmetries and Fusion Energy. Specifically, I will introduce new methods for computing stellarator equilibria more efficiently and accurately, and also present the first results of a new approach for stellarator optimization based on direct coil-plasma optimization with analytically computed derivatives. Although the presentation focuses on new numerical solvers, it is intended for a physics audience, and will highlight the central physics ideas in our work rather than algorithmic and mathematical details.