JACoW is a publisher in Geneva, Switzerland that publishes the proceedings of accelerator conferences held around the world by an international collaboration of editors.
TY - CONF AU - Morozov, I.A. AU - Piminov, P.A. ED - Kuzin, Maksim V. ED - Schaa, Volker RW TI - Detection of Anomalies in BPM Signals at the VEPP-4M J2 - Proc. of RuPAC2021, Alushta, Crimea, 27 September-01 October 2021 CY - Alushta, Crimea T2 - Russian Particle Accelerator Conference T3 - 27 LA - english AB - Beam position monitors (BPMs) are widely used for beam diagnostics in particle accelerators. Turn-by-turn (TbT) beam centroid data provide a means to estimate performance-critical accelerator parameters, like betatron frequency and optical functions. Parameter estimation accuracy is heavily related to TbT data quality. BPM faults might lead to erroneous estimation of accelerator parameters and should be accounted for achieving accurate and reliable results. Several anomaly detection methods for TbT data cleaning are considered. Derived features of BPM signals along with their robust dispersion estimation are used to flag faulty BPM signals. Estimated contamination factor is used with unsupervised learning methods (Local Outlier Factor and Isolation Forest). Application of anomaly detection methods for the VEPP-4M experimental TbT data is reported. PB - JACoW Publishing CP - Geneva, Switzerland SP - 420 EP - 422 KW - ISOL KW - detector KW - experiment KW - optics KW - simulation DA - 2021/10 PY - 2021 SN - 2673-5539 SN - 978-3-95450-240-0 DO - doi:10.18429/JACoW-RuPAC2021-WEPSC40 UR - https://jacow.org/rupac2021/papers/wepsc40.pdf ER -