Speaker
Description
The advent of nuclear fusion energy has been an highly anticipated one for over a century. In this work, we explore a promising concept to achieve it through magnetic confinement, the stellarator, which breaks the tokamak symmetry, allowing steady-state operation. Optimization of this design requires the iterative calculation of the magnetic equilibrium, tending to be computationally expensive. To expedite optimization, an approximation around the magnetic axis was implemented. For this work, we survey the differences between orbits of energetic particles for both the conventional construction and the approximate one. We then obtain optimized configurations with enhanced particle confinement by employing both local and global optimization methods. Finally, we show how machine learning algorithms may help identify novel stellarator designs.
