PACMAN – Particle Accelerators and Machine Learning

Co-PIs:

  • Jochem Snuverink, PSI
  • Tatiana Pieloni, EPFL
  • Andreas Adelmann, PSI
  • Markus Janousch, PSI
  • Vittoria Rezzonico, EPFL

Project presentation
May 13th 2019

Objectives:

Particle accelerator facilities have a wide range of operational needs when it comes to tuning, optimisation, and control. For instance, at the Large Hadron Collider (LHC) at CERN and the High Intensity Proton Accelerator (HIPA) at Paul Scherrer Institute (PSI), users require long-term machine stability, or maximisation of a key beam parameter, for example collision rate. However, for other machines, like the new SwissFEL accelerator at PSI, there is a strong incentive to reduce time spent switching between user-requested operating conditions. In order to meet these sorts of demands, particle accelerators rely on interactions with control systems, on fine-tuning of machine settings by operators, online optimisation routines, and on databases of previous settings that were known to be optimal for some desired operating condition. With more than 35,000 particle accelerators in operation worldwide, spanning over a wide range of applications: from cancer therapy to industrial application to fundamental physics, any increase in performance has an immediate large, potential societal impact. We propose to bring Machine Learning (ML) to particle accelerator operation, in order to increase the performance. Each of the mentioned operational needs have corresponding ML-based approaches that could be used to supplement the existing workflows. ML, and more generally, model-independent tuning, is proposed as a novel method of operation. The project is likely to have a game-changing impact in how we model and operate charged particle accelerators in the near future.