Dresden 2020 – wissenschaftliches Programm

Die DPG-Frühjahrstagung in Dresden musste abgesagt werden! Lesen Sie mehr ...

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

AKBP: Arbeitskreis Beschleunigerphysik

AKBP 14: Focus session: Machine Learning

AKBP 14.2: Hauptvortrag

Donnerstag, 19. März 2020, 10:00–10:30, MOL 213

Towards Micro-Bunching Control at Storage Rings with Reinforcement Learning — •Tobias Boltz — KIT

The operation of ring-based synchrotron light sources with short electron bunches can provide intense coherent synchrotron radiation (CSR) up to the THz frequency range. Yet, the continuous reduction in bunch length and stable emission of CSR are limited by the self-interaction of the bunch with its own radiation field. Above a machine-specific threshold current, the emitted CSR power starts fluctuating rapidly and continuously due to the formation of dynamically evolving micro-structures in the longitudinal charge distribution. As these small spatial structures lead to an increased emission of CSR at higher frequencies, this effect might also be desirable dependent on the application at hand. In this contribution, we discuss complementary approaches to both excitation and mitigation of the micro-bunching dynamics in order to optimize the emitted CSR for each application individually. Therefore, we motivate the usage of an RF modulation scheme to exert control over the longitudinal beam dynamics and illustrate how reinforcement learning methods can be applied to optimize towards different objective functions.