Speaker
Description
In this paper, we show different machine learning algorithms to control and optimize the linear accelerator and the performance of the Free Electron Laser (FEL) at the University of Hawaiʻi at Mānoa (UHM).
We discuss the methods of Bayesian Optimization and Reinforcement learning in optimizing six dimensional beam distribution going through the lattices of the accelerator. The lattices consist of an injector section, the linac and a Double-Bend Achromat (DBA), prior to injection into the FEL undulator. In order to analyze the application of Machine Learning methods, we use a digital twin developed at UHM to simulate the beam dynamics.
The goal of this paper is to optimize the beam parameters at two specific locations in the beamline and to study the correlation between the real beam size measurements and the digital twin.
Funding Agency
The U.S. Department of Energy, Office of the Science, under Contract No. DE-SC0025583
| In which format do you inted to submit your paper? | LaTeX |
|---|