Anomaly Detection for Beam Loss Maps in the Large Hadron Collider

8th International Particle Accelerator Conference

Valentino, Gianluca^1,2 ; Bruce, Roderik^1 ; Redaelli, Stefano^1 ; Rossi, Roberto^1 ; Theodoropoulos, Panagiotis^1 ; Jaster-Merz, Sonja^3

CERN, Geneva 23

CH1211, Switzerland^1

University of Malta, Msida

MSD2080, Malta^2

University of Hamburg, Hamburg

20146, Germany^3

关键词: Anomaly detection;    Beam loss;    Large Hadron Collider;    Local Outlier Factor;    Loss distribution;    Setting errors;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/874/1/012002/pdf
DOI  :  10.1088/1742-6596/874/1/012002

来源: IOP

【 摘 要 】

In the LHC, beam loss maps are used to validate collimator settings for cleaning and machine protection. This is done by monitoring the loss distribution in the ring during infrequent controlled loss map campaigns, as well as in standard operation. Due to the complexity of the system, consisting of more than 50 collimators per beam, it is difficult to identify small changes in the collimation hierarchy, which may be due to setting errors or beam orbit drifts with such methods. A technique based on Principal Component Analysis and Local Outlier Factor is presented to detect anomalies in the loss maps and therefore provide an automatic check of the collimation hierarchy.