Speaker
Description
Designing superconducting magnets presents a challenge due to their inherently multi-faceted nature. It requires a deep understanding of diverse physics domains and the integration of specialized tools. To streamline the design process, CEA-IRFU started developing ALESIA, a novel optimization and data management platform.
Our platform leverages advanced algorithms, including nonlinear programming techniques, evolutionary algorithms, active learning strategies, and surrogate modelling, to navigate the intricate parameter space efficiently and converge rapidly towards optimal magnet designs while minimizing computational cost. Its flexible architecture seamlessly integrates with any existing physics simulation software, currently encompassing OPERA for magnetic field calculations and CAST3M for mechanical analysis.
Crucially, its automated optimization loop aims at considering, at the same time, all the critical stages: electromagnetism, conductor properties, mechanics, and quench behavior, ensuring holistic and robust design solutions. Currently, ALESIA is already employed in the development of superconducting magnets for two projects: the SPIN ROTATORS project for the Electron-Ion Collider (EIC) and the ASTERICS magnets for the NEW GAIN ion injector at GANIL. This paper details the optimization algorithms used during all the stages and its applications.
