Göttingen 2025 – wissenschaftliches Programm
Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe
AKBP: Arbeitskreis Beschleunigerphysik
AKBP 7: Novel Accelerator Concepts II and FELs
AKBP 7.4: Vortrag
Mittwoch, 2. April 2025, 17:00–17:15, ZHG004
A virtual spectral diagnostic for plasma accelerated bunches at FLASHForward — •Philipp Burghart1,2, Lewis Boulton1, and Jonathan Wood1 for the FLASHForward collaboration — 1Deutsches Elektronen-Synchrotron DESY, Hamburg, Germany — 2University of Hamburg, Germany
Plasma-wakefield acceleration (PWFA) promises to reduce the size of future machines significantly by providing multi-GeV/m acceleration gradients, orders of magnitude higher than conventional RF accelerators. However, PWFA is a process with many non-linear dependencies, making it difficult to understand the influence of input parameters. Moreover, measurements of e.g. energy spectra are destructive, preventing the output beam from being used for applications whilst only allowing for the diagnosis of one bunch in a bunch train simultaneously. Neural networks trained on non-destructive measurements can be used to predict the properties of accelerated bunches, which would provide more insight into sources of variability and potential shot-to-shot, non-destructive measurements for whole bunch trains. Using experimental data collected at FLASHForward - a beam-driven plasma acceleration experiment at DESY, Hamburg - a neural network-based virtual diagnostic predicting the spectral properties of plasma accelerated bunches is being investigated. In this contribution, we present first results from this project.
Keywords: Plasma wakefield acceleration; Neural Networks; Machine Learning