Author: Xu, X.Y.
Paper Title Page
WEPP021 Machine Learning Image Processing Technology Application in Bunch Longitudinal Phase Data Information Extraction 561
 
  • X.Y. Xu, Y.M. Zhou
    SINAP, Shanghai, People’s Republic of China
  • Y.B. Leng, Y.M. Zhou
    SSRF, Shanghai, People’s Republic of China
  • X.Y. Xu
    University of Chinese Academy of Sciences, Beijing, People’s Republic of China
 
  To achieve the bunch-by-bunch longitudinal phase measurement, Shanghai Synchrotron Radiation Facility (SSRF) has developed a high resolution measurement system. We used this measurement system to study the injection transient process, and obtained the longitudinal phase of the refilled bunch and the longitudinal phase of the original stored bunch. A large number of parameters of the synchronous damping oscillation are included in this large amount of longitudinal phase data, which are important for the evaluation of machine state and bunch stability. The multi-turn phase data of a multi-bunch is a large two-dimensional array that can be converted into an image. The convolutional neural network (CNN) is a machine learning model with strong capabilities in image processing. We hope to use the convolutional neural network to process the longitudinal phase two-dimensional array data, and extract important parameters such as the oscillation amplitude and the synchrotron damping time.  
DOI • reference for this paper ※ https://doi.org/10.18429/JACoW-IBIC2019-WEPP021  
About • paper received ※ 23 August 2019       paper accepted ※ 10 September 2019       issue date ※ 10 November 2019  
Export • reference for this paper using ※ BibTeX, ※ LaTeX, ※ Text/Word, ※ RIS, ※ EndNote (xml)