5th ICFA Beam Dynamics Mini-Workshop on Machine Learning for Particle Accelerators

Bayesian optimization for the SIS18 injection

Not scheduled

20m

80/1-001 - Globe of Science and Innovation - 1st Floor (CERN)

Poster Optimisation and Control Poster session

Speaker

Sabrina Maria Appel

Description

The complexity of the GSI/FAIR accelerator facility demands a high level of automation to maximize the time for physics experiments. Accelerator laboratories across the globe are investigating numerous techniques to achieve this goal, including classical optimization, Bayesian optimization (BO), and reinforcement learning. This presentation will provide an overview of recent activities in these domains at GSI.
Beginning with conventional optimization, the beam loss during the multi-turn injection into the SIS18 synchrotron was reduced from 40% to 15% in approximately 15 minutes, whereas manual adjustments may takeup to 2 hours.
The implementation of Multi-Objective BO has resulted in the first physical measurement of a Pareto front on the SIS18 injection. Physics-informed BO for automated injection optimization is currently under investigation and has been utilized for a potential two-plane multi-turn injection via simulation.