Author: Scheinker, A.
Paper Title Page
TUOXGD3 6D Phase Space Diagnostics Based on Adaptively Tuned Physics-Informed Generative Convolutional Neural Networks 776
 
  • A. Scheinker
    LANL, Los Alamos, New Mexico, USA
  • F.W. Cropp V
    UCLA, Los Angeles, USA
  • D. Filippetto
    LBNL, Berkeley, California, USA
 
  Funding: US Department of Energy, DOE Office of Science Graduate Student Research (SCGSR) contract numbers 89233218CNA000001 and DE-AC02-05CH11231 and by the NSF under Grant No. PHY-1549132.
A physics-informed generative convolutional neural network (CNN)-based 6D phase space diagnostic is presented which generates all 15 unique 2D projections (x,y), (x,y’),…, (z,E) of a charged particle beam’s 6D phase space (x,y,z,x’,y’,E)*. The CNN is trained by supervised learning over a wide range of input beam distributions, accelerator parameters, and the associated 6D beam phase spaces at multiple accelerator locations. The CNN is applied in an un-supervised adaptive manner without knowledge of the input beam distribution or accelerator parameters and is robust to their unknown time variation. Adaptive feedback automatically tunes the low-dimensional latent space of the encoder-decoder CNN to predict the 6D phase space based only on 2D (z,E) longitudinal phase space measurements from a device such as a transverse deflecting RF cavity (TCAV). This method has the potential to provide diagnostics beyond the existing state of the art at many accelerator facilities. Studies are presented for two very different accelerators: the 5-meter-long ultra-fast electron diffraction (UED) HiRES compact accelerator at LBNL and the kilometer long plasma wakefield accelerator FACET-II at SLAC.
*A. Scheinker. "Adaptive machine learning for time-varying systems: low dimensional latent space tuning." Journal of Instrumentation 16.10, 2021: P10008. https://doi.org/10.1088/1748-0221/16/10/P10008
 
slides icon Slides TUOXGD3 [3.112 MB]  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2022-TUOXGD3  
About • Received ※ 21 May 2022 — Revised ※ 13 June 2022 — Accepted ※ 15 June 2022 — Issue date ※ 16 June 2022
Cite • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
 
TUPOST055 Toward Machine Learning-Based Adaptive Control and Global Feedback for Compact Accelerators 991
 
  • F.W. Cropp V, P. Musumeci
    UCLA, Los Angeles, USA
  • D. Filippetto, A. Gilardi, S. Paiagua, D. Wang
    LBNL, Berkeley, California, USA
  • A. Scheinker
    LANL, Los Alamos, New Mexico, USA
 
  Funding: This work was supported by the DOE Office of Science Graduate Student Research (SCGSR) program, by the DOE Office of Basic Energy Sciences under Contract No. DE-AC02-05CH11231, … continued
The HiRES beamline at Lawrence Berkeley National Laboratory (USA) is a state-of-the-art compact accelerator providing ultrafast relativistic electron pulses at MHz repetition rates, for applications in ultrafast science and for particle accelerator science and technology R&D. Using HiRES as testbed, we seek to apply recent developments in machine learning and computational techniques for machine-learning-based adaptive control, and eventually, a full control system based on global feedback. The ultimate goal is to demonstrate the benefits of such a suite of controls to UED, including increased temporal and spatial resolution. Concrete steps toward these goals are presented, including automatic, model-independent tuning for accelerators, and energy virtual diagnostics with direct application to improving UED temporal resolution.
… [continued from below] by the DOE Office of Science, Office of High Energy Physics under contract number 89233218CNA000001 and DE-AC02-05CH11231 and by the NSF under Grant No. PHY-1549132.
 
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2022-TUPOST055  
About • Received ※ 08 June 2022 — Revised ※ 12 June 2022 — Accepted ※ 16 June 2022 — Issue date ※ 20 June 2022
Cite • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)