Publications
Publication Information
| Title | Real-time artificial intelligence for accelerator control: A study at the Fermilab Booster |
| Authors | J. John, Malachi Schram, Christian Herwig, D. Kafkes, Jovan Mitrevski, William Pellico, G. Perdue, Andres Quintero-Parra, Brian Schupbach, Kiyomi Seiya, Nhan Tran, Javier Duarte, Y. Huang, R. Keller |
| JLAB number | JLAB-CST-21-3606 |
| LANL number | (None) |
| Other number | DOE/OR/23177-5611 |
| Document Type(s) | (Journal Article) |
| Associated with EIC: | No |
| Supported by Jefferson Lab LDRD Funding: | No |
| Funding Source: | High Energy Physics (HEP) |
|
Journal Compiled for Physical Review Accelerators and Beams Volume 24 Page(s) 104601 Refereed | |
| Publication Abstract: | We describe a method for precisely regulating the gradient magnet power supply (GMPS) 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 GMPS, 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 (FPGAs), and show the first machine-learning based control algorithm implemented on an FPGA for controls at the Fermilab accelerator complex. As there are no surprise latencies on an FPGA, this capability is important for operational stability in complicated environments such as an accelerator facility. |
| Experiment Numbers: | other |
| Group: | Data Science |
| Document: | |
| DOI: | https://doi.org/10.1103/PhysRevAccelBeams.24.104601 |
| Accepted Manuscript: | PhysRevAccelBeams.24.104601.pdf |
| Supporting Documents: | |
| Supporting Datasets: | |