Berlin 2024 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
AKBP: Arbeitskreis Beschleunigerphysik
AKBP 12: AKBP Poster Session
AKBP 12.16: Poster
Donnerstag, 21. März 2024, 11:00–14:00, Poster A
Algorithmic Quantification of Laser-Plasma Accelerated Electron Bunches for Campaign Bayesian Steering — Tobias Hänel1, Franziska Herrmann1, Susanne Schöbel1, Anna Willmann1, Richard Pausch1, Amin Ghaith2,1, Maxwell Laberge1,3, Ye-Yu Chang1, Patrick Ufer1,4, Paula Twellenkamp1, Teresa D'Orsi Barreto1, Michael Bussmann5,1, Ulrich Schramm1,4, Arie Irman1, and •Jeffrey Kelling1,6 — 1Helmholtz-Zentrum Dresden - Rossendorf, Dresden, Germany — 2Synchrotron SOLEIL, Saint-Aubin, France — 3The University of Texas at Austin, Austin, TX, USA — 4Technische Universität Dresden, Dresden, Germany — 5CASUS, Görlitz, Germany — 6Institut für Physik, TU Chemnitz, Chemnitz, Germany
Laser-plasma accelerators (LPA) are much smaller than conventional systems and can generate electron bunches with uniquely high current and small length, making them ideally suited to seed free electron lasers (FEL). UV lasing in such a setup has recently been demonstrated at the COXINEL experiment. Characteristics of the emitted radiation are sensitive to the shape of the seeding electron bunches, which in turn is determined by the pulse-properties of the driving laser. In order to enable the use of multi-objective Bayesian optimization to efficiently find pareto-optimal laser parameters given electron-bunch energies and spreads, we require a consistent and automatic way of extracting these properties measured energy spectra. Here, we present our approach, based on classical computer-vision methodology and evaluate the efficacy for Bayesian optimization runs.
Keywords: laser plasma acceleration; multi-objective optimization; Bayesian optimization; free-electron laser