IPAC2015 - List of Authors (Stabile, P.)

Paper	Title	Page
MOPWI028	Initial Experimental Results of a Machine Learning-Based Temperature Control System for an RF Gun	1217
	A.L. Edelen, S. Biedron, S.V. Milton CSU, Fort Collins, Colorado, USA B.E. Chase, D.J. Crawford, N. Eddy, D.R. Edstrom, E.R. Harms, J. Ruan, J.K. Santucci, P. Stabile Fermilab, Batavia, Illinois, USA
	Colorado State University (CSU) and Fermi National Accelerator Laboratory (Fermilab) have been developing a control system to regulate the resonant frequency of an RF electron gun. As part of this effort, we present experimental results for a benchmark temperature controller that combines a machine learning-based model and a predictive control algorithm for improved settling time, overshoot, and disturbance rejection relative to conventional techniques. Such improvements have implications for machine up-time and management of reflected power. This work is part of an on-going effort to develop adaptive, machine learning-based tools specifically to address control challenges found in particle accelerator systems.
DOI •	reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2015-MOPWI028
Export •	reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)

Paper

Title

Page

Initial Experimental Results of a Machine Learning-Based Temperature Control System for an RF Gun

1217

A.L. Edelen, S. Biedron, S.V. Milton
CSU, Fort Collins, Colorado, USA
B.E. Chase, D.J. Crawford, N. Eddy, D.R. Edstrom, E.R. Harms, J. Ruan, J.K. Santucci, P. Stabile
Fermilab, Batavia, Illinois, USA

Colorado State University (CSU) and Fermi National Accelerator Laboratory (Fermilab) have been developing a control system to regulate the resonant frequency of an RF electron gun. As part of this effort, we present experimental results for a benchmark temperature controller that combines a machine learning-based model and a predictive control algorithm for improved settling time, overshoot, and disturbance rejection relative to conventional techniques. Such improvements have implications for machine up-time and management of reflected power. This work is part of an on-going effort to develop adaptive, machine learning-based tools specifically to address control challenges found in particle accelerator systems.

DOI •

reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2015-MOPWI028

Export •

reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)