Speaker
Description
Modern FEL operation increasingly relies on advanced software tools, automation, and data-driven optimization methods. At the European XFEL, a strategy is being developed to improve operational efficiency, overall FEL performance, and to reduce the dependence on experts during non-standard operation modes.
Recent developments include automated scanning and tuning procedures, the use of large language models for improved documentation and access to various knowledge bases, as well as operator support via chat-based interfaces, and agentic AI systems for performance optimization.
In addition, efforts are ongoing to combine individual optimization tools and software services in order to automate larger and more complex operational workflows.
This paper summarizes the overall concept, the current implementation status, and first experiences with AI- and machine-learning-related tools in daily accelerator operation.
| I have read and accept the Privacy Policy Statement | Yes |
|---|