LINAC2022 - List of Authors (Walker, N.)

Paper	Title	Page
MOPOPA03	Beam-Transient-Based LLRF Voltage Signal Calibration for the European XFEL	80
	N. Walker, V. Ayvazyan, J. Branlard, S. Pfeiffer, Ch. Schmidt DESY, Hamburg, Germany
	The European XFEL linac consists of 25 superconducting RF (SRF) stations. With the exception of the first station which is part of the injector, each station comprises 32 1.3-GHz SRF TESLA cavities, driven by a single 10-MW klystron. A sophisticated state-of-the-art low-level RF (LLRF) system maintains the complex vector sum of each RF station. Monitoring and maintaining the calibration of the cavity electric field (gradient) probe signals has proven critical in achieving the maximum energy performance and availability of the SRF linac. Since there are no dedicated diagnostics for cross-checking calibration of the LLRF system, a procedure has been implemented based on simultaneously measuring the beam transient in open-loop operation of all cavities. Based on methods originally developed at FLASH, the European XFEL procedure makes use of automation and the XFEL LLRF DAQ system to provide a robust and relatively fast (minutes) way of extracting the transient data, and is now routinely scheduled once per week. In this paper, we will report on the background, implementation, analysis methods, typical results, and their subsequent application for machine operation.
DOI •	reference for this paper ※ https://doi.org/10.18429/JACoW-LINAC2022-MOPOPA03
About •	Received ※ 13 August 2022 — Revised ※ 23 August 2022 — Accepted ※ 14 September 2022 — Issue date ※ 27 September 2022
Cite •	reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)

THPOPA26	Machine Learning Assisted Cavity Quench Identification at the European XFEL	798
THOPA09	use link to see paper's listing under its alternate paper code
	J. Branlard, A. Eichler, J.H.K. Timm, N. Walker DESY, Hamburg, Germany
	A server-based quench detection system is used since the beginning of operation at the European XFEL (2017) to stop driving superconducting cavities if they experience a quench. While this approach effectively detects quenches, it also generates false positives, tripping the accelerating stations when failures other than quenches occur. Using the post-mortem data snapshots generated for every trip, an additional signal (referred to as residual) is systematically computed based on the standard cavity model. Following an initial training on a set of such residuals derived from quench as well as non-quench events, two independent machine learning engines analyze routinely the trip snapshots and their residuals to identify if a trip was indeed triggered by a quench or has another root cause. The outcome of the analysis is automatically appended to the data snapshots and distributed to a team of experts. This constitutes a fully deployed example of machine-learning-assisted failure classification to identify quenches, supporting experts in their daily routine of monitoring and documenting the accelerator uptime and availability.
	Slides THPOPA26 [0.695 MB]
	Poster THPOPA26 [0.975 MB]
DOI •	reference for this paper ※ https://doi.org/10.18429/JACoW-LINAC2022-THPOPA26
About •	Received ※ 19 August 2022 — Revised ※ 24 August 2022 — Accepted ※ 01 September 2022 — Issue date ※ 01 September 2022
Cite •	reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)

Paper

Title

Page

Beam-Transient-Based LLRF Voltage Signal Calibration for the European XFEL

80

N. Walker, V. Ayvazyan, J. Branlard, S. Pfeiffer, Ch. Schmidt
DESY, Hamburg, Germany

The European XFEL linac consists of 25 superconducting RF (SRF) stations. With the exception of the first station which is part of the injector, each station comprises 32 1.3-GHz SRF TESLA cavities, driven by a single 10-MW klystron. A sophisticated state-of-the-art low-level RF (LLRF) system maintains the complex vector sum of each RF station. Monitoring and maintaining the calibration of the cavity electric field (gradient) probe signals has proven critical in achieving the maximum energy performance and availability of the SRF linac. Since there are no dedicated diagnostics for cross-checking calibration of the LLRF system, a procedure has been implemented based on simultaneously measuring the beam transient in open-loop operation of all cavities. Based on methods originally developed at FLASH, the European XFEL procedure makes use of automation and the XFEL LLRF DAQ system to provide a robust and relatively fast (minutes) way of extracting the transient data, and is now routinely scheduled once per week. In this paper, we will report on the background, implementation, analysis methods, typical results, and their subsequent application for machine operation.

DOI •

reference for this paper ※ https://doi.org/10.18429/JACoW-LINAC2022-MOPOPA03

About •

Received ※ 13 August 2022 — Revised ※ 23 August 2022 — Accepted ※ 14 September 2022 — Issue date ※ 27 September 2022

Cite •

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

THPOPA26

Machine Learning Assisted Cavity Quench Identification at the European XFEL

798

THOPA09

use link to see paper's listing under its alternate paper code

J. Branlard, A. Eichler, J.H.K. Timm, N. Walker
DESY, Hamburg, Germany

A server-based quench detection system is used since the beginning of operation at the European XFEL (2017) to stop driving superconducting cavities if they experience a quench. While this approach effectively detects quenches, it also generates false positives, tripping the accelerating stations when failures other than quenches occur. Using the post-mortem data snapshots generated for every trip, an additional signal (referred to as residual) is systematically computed based on the standard cavity model. Following an initial training on a set of such residuals derived from quench as well as non-quench events, two independent machine learning engines analyze routinely the trip snapshots and their residuals to identify if a trip was indeed triggered by a quench or has another root cause. The outcome of the analysis is automatically appended to the data snapshots and distributed to a team of experts. This constitutes a fully deployed example of machine-learning-assisted failure classification to identify quenches, supporting experts in their daily routine of monitoring and documenting the accelerator uptime and availability.

Slides THPOPA26 [0.695 MB]

Poster THPOPA26 [0.975 MB]

DOI •

reference for this paper ※ https://doi.org/10.18429/JACoW-LINAC2022-THPOPA26

About •

Received ※ 19 August 2022 — Revised ※ 24 August 2022 — Accepted ※ 01 September 2022 — Issue date ※ 01 September 2022

Cite •

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