Abstract
A machine learning-based optimization algorithm, the multigeneration Gaussian process optimizer, is used to optimize the nonlinear beam dynamics of the Advanced Photon Source storage ring. The dynamic aperture (DA) and the local momentum aperture (LMA) are first optimized separately with sextupole knobs. Solutions found with these optimizations are used to seed the initial population of seeds in two-objective optimizations that simultaneously optimize the DA and LMA, which lead to a distribution of solutions with different DA and LMA performances, from which a setting suitable for operation can be selected.
Highlights
As storage rings push for lower emittances, sextupole magnets used in the lattices for chromaticity correction become stronger
The approach we demonstrate here would be important for the commissioning of the generation of storage rings as their lattices have many more magnets and more error sources
We conducted a series of experiments to optimize the dynamic aperture (DA) and local momentum aperture (LMA) for the Advanced Photon Source (APS) storage ring with sextupole knobs, using primarily the MG-GPO method, a stochastic, MLbased optimization algorithm
Summary
As storage rings push for lower emittances, sextupole magnets used in the lattices for chromaticity correction become stronger. Online optimization of nonlinear beam dynamics for storage rings has been proposed and demonstrated [4] in experiments, in which sextupole knobs were used to improve the DA by optimizing the injection efficiency. At the Advanced Photon Source (APS), offline sextupole optimization was done previously with global sextupole families and local sextupole knobs around the limiting physical aperture [9] The optimization improved both the DA and LMA simultaneously as the optimized solution effectively enlarged the physical acceptance by changing the nonlinear deformation of the phase space at the minimum aperture location. This is the first time such a two-objective nonlinear storage ring beam dynamics optimization is done in experiments.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
