Accelerator Systems and Components
Operation, Artificial Intelligence, Machine Learning
Paper Title Page
TU2I2 Development and Commissioning of the K500 Superconducting Heavy Ion Cyclotron 46
  • S. Som, A. Bandyopadhyay, S. Bandyopadhyay, S. Bhattacharya, P. Bhattacharyya, T. Bhattacharyya, U. Bhunia, N. Chaddha, J. Debnath, M.K. Dey, A. Dutta Gupta, S. Ghosh, A. Mandal, P.Y. Nabhiraj, C. Nandi, Z.A. Naser, S. Pal, S. Pal, U. Panda, J. Pradhan, A. Roy, S. Saha, S. Seth, S.K. Thakur
    VECC, Kolkata, India
  Funding: Work supported by the DAE, Government of India.
The K500 Superconducting Cyclotron (SCC) has been developed indigenously and commissioned at VECC. The three-phase Radio-Frequency (RF) system of SCC, consists of three half-wave cavities placed vertically 120 deg. apart. Each half-wave cavity has two quarter-wave cylindrical cavities tied together at the centre and symmetrically placed about median plane of the cyclotron. Each quarter-wave cavity is made up of a short circuited non-uniform coaxial transmission line (called "dee-stem") terminated by accelerating electrode (called "Dee"). The SCC, operating in the range 9 to 27 MHz, has amplitude and phase stability within 100 ppm and 0.1 deg. respectively. The overview of all the subsystems of the cyclotron along with low-level RF (LLRF), high and low power RF amplifiers, cavity analysis, absolute Dee voltage measurement using X-ray method, amplitude and phase control loops will be presented in the talk. The commissioning of the cyclotron with first harmonic Nitrogen4+ beam extracted at 252 MeV, while operating at 14 MHz RF frequency, along with the correction of first harmonic magnetic field error by repositioning the cryostat within 120 micron accuracy, will be discussed briefly.
slides icon Slides TU2I2 [18.037 MB]  
DOI • reference for this paper ※  
About • Received ※ 15 June 2022 — Revised ※ 27 June 2022 — Accepted ※ 10 August 2022 — Issue date ※ 05 September 2022
Cite • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)  
TH1I2 Reinforcement Learning and Bayesian Optimization for Ion Linac Operations 130
  • J.L. Martinez Marin, B.R. Blomberg, K.J. Bunnell, G.M. Dunn, E. Letcher, B. Mustapha, D. Stanton
    ANL, Lemont, Illinois, USA
  Funding: This work was supported by the U.S. Department of Energy, under Contract No. DE-AC02-06CH11357. This research used the ATLAS facility, which is a DOE Office of Nuclear Physics User Facility.
The use of artificial intelligence can significantly reduce the time needed to tune an accelerator system such as the Argonne Tandem Linear Accelerator System (ATLAS) where a new beam is tuned once or twice a week. After establishing automatic data collection procedures and having analysed the data, machine learning models were developed and tested to tune subsections of the linac. Models based on Reinforcement Learning (RL) and Bayesian Optimization (BO) were developed, their respec-tive results are discussed and compared. RL and BO are well known AI techniques, often used for control systems. The results were obtained for a subsection of ATLAS that contains complex elements such as the radio-frequency quadrupole (RFQ). The models will be later generalized to the whole ATLAS linac, and similar models can be devel-oped for any accelerator with a modern control system.
slides icon Slides TH1I2 [4.617 MB]  
DOI • reference for this paper ※  
About • Received ※ 09 July 2022 — Revised ※ 10 August 2022 — Accepted ※ 19 September 2022 — Issue date ※ 19 September 2022
Cite • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)