Real-time Artificial Intelligence for Accelerator Control: A Study at the Fermilab Booster

We describe a method for precisely regulating the gradient magnet power supply at the Fermilab Booster accelerator complex using a neural network trained via reinforcement learning. We demonstrate preliminary results by training a surrogate machine-learning model on real accelerator data to emulate...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2021-10
Main Authors: Jason St John, Herwig, Christian, Kafkes, Diana, Mitrevski, Jovan, Pellico, William A, Perdue, Gabriel N, Quintero-Parra, Andres, Schupbach, Brian A, Seiya, Kiyomi, Tran, Nhan, Schram, Malachi, Duarte, Javier M, Huang, Yunzhi, Keller, Rachael
Format: Article
Language:English
Subjects:
Artificial intelligence
Field programmable gate arrays
Machine learning
Neural networks
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We describe a method for precisely regulating the gradient magnet power supply at the Fermilab Booster accelerator complex using a neural network trained via reinforcement learning. We demonstrate preliminary results by training a surrogate machine-learning model on real accelerator data to emulate the Booster environment, and using this surrogate model in turn to train the neural network for its regulation task. We additionally show how the neural networks to be deployed for control purposes may be compiled to execute on field-programmable gate arrays. This capability is important for operational stability in complicated environments such as an accelerator facility.
ISSN:2331-8422
DOI:10.48550/arxiv.2011.07371