WAO2023 International Workshop on Accelerator Operations

Session

How Does the Machine Learning Integrate with Operation?

12 Sept 2023, 09:00

Tsukuba International Congress Center (EPOCHAL)

21. OR22 - Machine Learning as a tool for accelerator operation: status, pros, cons and futur deployments.

Bertrand Lefort (CERN)

12/09/2023, 09:00

"Machine Learning (ML) is now widely used in the Industry. Applications in robotic and autonomous vehicles have proven that control theories can benefit greatly from these technologies. The same is true in science, Deep Learning (DL) offers great expectations for science and has demonstrated its ability to extract unexpected relationships between inputs and outputs in complex data sets. One of...

22. OR21 - Online reinforcement learning control of beam collision at IP for BEPCII

Jiaqi Fan (IHEP)

12/09/2023, 09:25

For BEPCII, operators need to tune the transverse offset in displacement and angular deviation (x, x’, y, y’) of the two beams at the interaction point (IP) to maintain a high luminosity as the beam current decays in a normal run. Since the optimal offset change with current nonlinearly in a run and varies in each run. It is laborious to keep the best beam offset at IP for high luminosity at...

23. OR24 - Application examples of the machine learning in KEK

Miho Shimada (KEK)

12/09/2023, 09:45

Machine learning (ML) is widely used and several tools are accessible for free so we easily tried them in several accelerator operations in KEK. One of the applications of compact ERL is beam tuning for infrared SASE FEL. It is used to maximize the intensity of the FEL light, which requires bunch compression, transverse beam matching at the undulator, small energy spread, and transverse...

24. OR23 - Transitioning the Use of Machine Learning from Physics Study to Production Operation*

Charles Peters (SNS)

12/09/2023, 10:05

"The Spallation Neutron Source (SNS) at the Oak Ridge National Laboratory (ORNL) is a megawatt class neutron production proton accelerator. The facility is undergoing a beam power upgrade from 1.4 MW to 2 MW and increasing the yearly scheduled neutron production operating time from 4500 to 5000 hours. One area of focus is to utilize machine learning to try to help achieve these lofty...