Master theses
Simulation of high-power laser spectra using generative machine learning
Author
Anna Kapitánová
Year
2025
Type
Master thesis
Supervisor
Ing. Matěj Mráček
Reviewers
Mgr. Alexander Kovalenko, Ph.D.
Department
Summary
High-power laser systems, such as the L1-Allegra laser at ELI Beamlines, are essential in laser wakefield acceleration, offering compact alternatives to conventional particle accelerators. However, their operation is complex, expensive, and highly sensitive to system parameters. This thesis explores the use of diffusion-based generative models to simulate the spectral output of the L1-Allegra system, providing insights that are not accessible during real-time operation. In addition to reconstructing unmeasured spectra, the model supports the optimization of laser settings, helping to improve system performance and reduce the need for extensive experimental tuning.