17–22 May 2026
C.I.D
Europe/Zurich timezone

Improving Injection at Diamond using Dimensionality Reduction Techniques and Bayesian Optimisation

WEP5057
20 May 2026, 16:00
2h
C.I.D

C.I.D

Deauville, France
Poster Presentation MC5.D01: Beam Optics Lattices, Correction Schemes, Transport Poster session

Speaker

Shaun Preston (University of Oxford, John Adams Institute for Accelerator Science, Diamond Light Source)

Description

Injection efficiency into the Diamond Light Source Storage Ring (SR) is currently optimised by operators performing a grid scan over the final pair of corrector magnets inside the Booster-to-Storage Ring (BTS) transfer line. The phase advance between the pair is sufficient to provide good control over the position and angle of the beam as it enters the SR. However, the method is slow and the strengths of the corrector pair can approach power supply limits over time as machine conditions drift. We propose a Bayesian optimisation algorithm to optimise leading right-singular vector coefficients obtained from a Singular Value Decomposition (SVD) of the BTS response matrix over all corrector magnets, improving sample efficiency. We further transform the proposed coefficients through a non-linear map to restrict the set of solutions to a desirable range as determined by the user. We compare our results to the grid scan technique and suggest further refinements to the algorithm.

In which format do you inted to submit your paper? LaTeX

Author

Shaun Preston (University of Oxford, John Adams Institute for Accelerator Science, Diamond Light Source)

Co-authors

Ian Martin (Diamond Light Source) Philip Burrows (John Adams Institute for Accelerator Science) Neven Blaskovic Kraljevic (Diamond Light Source) Richard Fielder (Diamond Light Source)

Presentation materials

There are no materials yet.