IPAC2018 - List of Authors (Pfeiffer, S.)

Paper	Title	Page
WEPAF051	LLRF Operation and Performance at the European XFEL	1934
	M. Omet, V. Ayvazyan, J. Branlard, Ł. Butkowski, M. Hierholzer, M. Killenberg, D. Kostin, L. Lilje, S. Pfeiffer, H. Schlarb, Ch. Schmidt, V. Vogel, N. Walker DESY, Hamburg, Germany
	The European X-ray Free-Electron Laser (XFEL) at Deutsches Elektronen-Synchrotron (DESY), Hamburg, Germany is a user facility providing ultrashort hard and soft X-ray flashes with a high brilliance. All LLRF stations of the injector, covering the normal conducting RF gun, A1 (8 1.3 GHz superconducting cavities (SCs)) and AH1 (8 3.9 GHz SCs), were successfully commissioned by the end of 2015. The commissioning of LLRF stations A2 to A23 (32 1.3 GHz SCs each) in the XFEL accelerator tunnel (XTL) was concluded in June 2017. SASE light was produced in SASE undulator section SA1 and delivered to the first users in September 2017, marking the beginning of regular user operation. The current state of the LLRF systems, the experience gained during operation and the performance achieved in terms of stability and energy reach are presented.
DOI •	reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2018-WEPAF051
Export •	reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)

WEPMF058	Anomaly Detection for Cavity Signals - Results from the European XFEL	2502
	A.S. Nawaz, S. Pfeiffer DESY, Hamburg, Germany G. Lichtenberg HAW, Hamburg, Germany P. Rostalski Institute for Electrical Engineering in Medicine, Lübeck, Germany
	The data throughput of the European XFEL DAQ is about 1.5 Gb/s. Data depicting the cavity signal behavior is currently only saved manually. This either happens, when cavity tests are being performed, or an operator detects a fault in the cavity system, that has to be further investigated. Those instances of interest are neither systematically nor automatically stored. It can therefore be assumed that unwanted or degraded cavity behavior is detected late or not at all. It is proposed to change the focus from detecting known faults (such as quenches) to additionally detect anomalies in the cavity system behavior. In order to detect anomalies in the cavity signals, an algorithm is proposed using a cavity model. It aims on finding those data sets, which diverge from the nominal cavity behavior, saving those instances for later analysis. The nominal behavior is defined by the cavity electromagnetic resonance model with beam loading as well as the model for the mechanical oscillations due to the Lorentz Forces. By using such an approach, the detection of anomalies, as well as faults could be automated. This contribution aims to summarize the influence of beam loading on the detection and gives examples for anomalies that were found in several cavities.
DOI •	reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2018-WEPMF058
Export •	reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)

Paper

Title

Page

LLRF Operation and Performance at the European XFEL

1934

M. Omet, V. Ayvazyan, J. Branlard, Ł. Butkowski, M. Hierholzer, M. Killenberg, D. Kostin, L. Lilje, S. Pfeiffer, H. Schlarb, Ch. Schmidt, V. Vogel, N. Walker
DESY, Hamburg, Germany

The European X-ray Free-Electron Laser (XFEL) at Deutsches Elektronen-Synchrotron (DESY), Hamburg, Germany is a user facility providing ultrashort hard and soft X-ray flashes with a high brilliance. All LLRF stations of the injector, covering the normal conducting RF gun, A1 (8 1.3 GHz superconducting cavities (SCs)) and AH1 (8 3.9 GHz SCs), were successfully commissioned by the end of 2015. The commissioning of LLRF stations A2 to A23 (32 1.3 GHz SCs each) in the XFEL accelerator tunnel (XTL) was concluded in June 2017. SASE light was produced in SASE undulator section SA1 and delivered to the first users in September 2017, marking the beginning of regular user operation. The current state of the LLRF systems, the experience gained during operation and the performance achieved in terms of stability and energy reach are presented.

DOI •

reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2018-WEPAF051

Export •

reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)

WEPMF058

Anomaly Detection for Cavity Signals - Results from the European XFEL

2502

A.S. Nawaz, S. Pfeiffer
DESY, Hamburg, Germany
G. Lichtenberg
HAW, Hamburg, Germany
P. Rostalski
Institute for Electrical Engineering in Medicine, Lübeck, Germany

The data throughput of the European XFEL DAQ is about 1.5 Gb/s. Data depicting the cavity signal behavior is currently only saved manually. This either happens, when cavity tests are being performed, or an operator detects a fault in the cavity system, that has to be further investigated. Those instances of interest are neither systematically nor automatically stored. It can therefore be assumed that unwanted or degraded cavity behavior is detected late or not at all. It is proposed to change the focus from detecting known faults (such as quenches) to additionally detect anomalies in the cavity system behavior. In order to detect anomalies in the cavity signals, an algorithm is proposed using a cavity model. It aims on finding those data sets, which diverge from the nominal cavity behavior, saving those instances for later analysis. The nominal behavior is defined by the cavity electromagnetic resonance model with beam loading as well as the model for the mechanical oscillations due to the Lorentz Forces. By using such an approach, the detection of anomalies, as well as faults could be automated. This contribution aims to summarize the influence of beam loading on the detection and gives examples for anomalies that were found in several cavities.

DOI •

reference for this paper ※ https://doi.org/10.18429/JACoW-IPAC2018-WEPMF058

Export •

reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)