| Paper | Title | Page |
|---|---|---|
| MOCPL04 | Software Architecture for Automatic LHC Collimator Alignment Using Machine Learning | 78 |
|
||
|
The Large Hadron Collider at CERN relies on a collimation system to absorb unavoidable beam losses before they reach the superconducting magnets. The collimators are positioned close to the beam in a transverse setting hierarchy achieved by aligning each collimator with a precision of a few tens of micrometers. In previous years, collimator alignments were performed semi-automatically*, requiring collimation experts to be present to oversee and control the entire process. In 2018, manual, expert control of the alignment procedure was replaced by dedicated machine learning algorithms, and this new software was used for collimator alignments throughout the year. This paper gives an overview of the software re-design required to achieve fully automatic collimator alignments, describing in detail the software architecture and controls systems involved. Following this successful deployment, this software will be used in the future as the default alignment software for the LHC.
*G. Valentino et al., "Semi-automatic beam-based LHC collimator alignment", Physical Review Special Topics-Accelerators and Beams vol. 15, no. 5, 2012. |
||
|
Slides MOCPL04 [5.933 MB] | |
| DOI • | reference for this paper ※ https://doi.org/10.18429/JACoW-ICALEPCS2019-MOCPL04 | |
| About • | paper received ※ 28 September 2019 paper accepted ※ 09 October 2019 issue date ※ 30 August 2020 | |
| Export • | reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml) | |
| MOPHA010 | Automatic Beam Loss Threshold Selection for LHC Collimator Alignment | 208 |
|
||
| The collimation system used in the Large Hadron Collider at CERN is positioned around the beam with a hierarchy that protects sensitive equipment from unavoidable beam losses. The collimator settings are determined using a beam-based alignment technique, where collimator jaws are moved towards the beam until the beam losses exceed a predefined threshold. This threshold needs to be updated dynamically, corresponding to the changes in the beam losses. The current method for aligning collimators is semi-automated requiring a collimation expert to monitor the loss signals and continuously select and update the threshold accordingly. The human element in this procedure is a major bottleneck for speeding up the alignment. This paper therefore proposes a method to fully automate this threshold selection. A data set was formed from previous alignment campaigns and analyzed to define an algorithm that produced results consistent with the user selections. In over 90% of the cases the difference between the two was negligible and the algorithm presented in this study was used for collimator alignments throughout 2018. | ||
|
Poster MOPHA010 [1.763 MB] | |
| DOI • | reference for this paper ※ https://doi.org/10.18429/JACoW-ICALEPCS2019-MOPHA010 | |
| About • | paper received ※ 28 September 2019 paper accepted ※ 08 October 2019 issue date ※ 30 August 2020 | |
| Export • | reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml) | |