2023年
12/01
開催
開催日時
2023/12/01 10:30 – 12:00
開催場所
J-PARC CCR会議室 + Zoom
講演者
Dr. Sharankova, Ralitsa V. (FNAL)
言語
English
詳細ページ
お問い合わせ
劉 勇 LIU Yong (PHS 4735)
概要
The Linac delivers 400 MeV H- beam to the rest of the accelerator chain at Fermilab. Beam quality and efficiency of the Linac directly affect beam losses in downstream machines and ultimately beam quality to users. Therefore, delivering stable beam with maximal throughput from the Linac is crucial for the complex. To operate a high current beam, accelerators must minimize uncontrolled particle loss; this can be accomplished by minimizing beam longitudinal emittance via RF parameter optimization. In practice, RF tuning is required daily since the resonance frequencies of the accelerating cavities are affected by ambient temperature and humidity variations and thus drift with time. Moreover, the energy and phase space distribution of particles emerging from the ion source are subject to fluctuations. Time drift is not unique to Fermilab; it affects all accelerator complexes around the world. As such, time-drift aware RF tuning has been an area of active research with increasingly sophisticated methods. Over the last few years applications of machine learning (ML) to accelerator optimization have grown exponentially, however this work is mainly still at the proof-of-concept stage. At the Fermilab Linac, we have developed and successfully tested several such algorithms on real data, which are expected to significantly improve accelerator performance once integrated into daily operations.