Speaker
Chong Shik Park
(Korea University Sejong Campus)
Description
We present a machine-learning surrogate model for the RAON LEBT that enables fast prediction of beam centroids at multiple diagnostics. A dataset of TRACK simulations spanning relevant steering-magnet and electrostatic-quadrupole settings is used to train fully connected neural networks. The surrogate model reproduces the underlying beam dynamics with high accuracy while providing orders-of-magnitude faster evaluation. This approach supports rapid orbit studies, optimization, and data-driven beam control in the RAON front-end transport system.
| Paper status | Resubmitted proceeding files received and assigned to an editor. Accepted. |
|---|
Author
Chong Shik Park
(Korea University Sejong Campus)