Author: Johnston, S.C.
Paper Title Page
WEPAB244 Optimization and Machine Learning Applied to the RF Manipulations of Proton Beams in the CERN PS 3201
 
  • A. Lasheen, H. Damerau, S.C. Johnston
    CERN, Meyrin, Switzerland
  
  The 25 ns bunch spac­ing in the LHC is de­fined by a se­quence of RF ma­nip­u­la­tions in the Pro­ton Syn­chro­tron (PS). Mul­ti­ple RF sys­tems cov­er­ing a large range of rev­o­lu­tion har­mon­ics (7 to 21, 42, 84, 168) allow per­form­ing RF ma­nip­u­la­tions such as beam split­ting, and non-adi­a­batic bunch short­en­ing. For the nom­i­nal beam sent to LHC, each bunch is split in 12 in the PS. The rel­a­tive am­pli­tude and phase set­tings of the RF sys­tems need to be pre­cisely ad­justed to min­i­mize the bunch-by-bunch vari­a­tions in in­ten­sity, lon­gi­tu­di­nal emit­tance, and bunch shape. How­ever, due to tran­sient beam-load­ing, the ideal set­tings, as well as the best achiev­able beam qual­ity, vary with beam in­ten­sity. Slow drifts of the hard­ware may also af­fect beam qual­ity. In this paper, au­tom­a­tized op­ti­miza­tion rou­tines based on par­ti­cle sim­u­la­tions with in­ten­sity ef­fects are pre­sented, to­gether with the first con­sid­er­a­tions of ma­chine learn­ing. The op­ti­miza­tion rou­tines are used to as­sess the best achiev­able lon­gi­tu­di­nal beam qual­ity ex­pected with the PS RF sys­tems up­grades, in the frame­work of the LHC In­jec­tor Up­grade pro­ject.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2021-WEPAB244  
About • paper received ※ 19 May 2021       paper accepted ※ 01 July 2021       issue date ※ 24 August 2021  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)