Beam-based Identification of Magnetic Field Errors in a Synchrotron using Deep Lie Map Networks

Abstract

We present the first experimental validation of the Deep Lie Map Network (DLMN) approach for recovering both linear and non-linear optics in a synchrotron. The DLMN facilitates the construction of a detailed accelerator model by integrating charged particle dynamics with machine learning methodology in a data-driven framework. The primary observable is the centroid motion over a limited number of turns, captured by beam position monitors. The DLMN produces an updated description of the accelerator in terms of magnetic multipole components, which can be directly utilized in established accelerator physics tools and tracking codes for further analysis. In this study, we apply the DLMN to the SIS18 hadron synchrotron at GSI for the first time. We discuss the validity of the recovered linear and non-linear optics, including quadrupole and sextupole errors, and compare our results with alternative methods, such as the LOCO fit of a measured orbit response matrix and the evaluation of resonance driving terms. The small number of required trajectory measurements, one for linear and three for non-linear optics reconstruction, demonstrates the method's time efficiency. Our findings indicate that the DLMN is well-suited for identifying linear optics, and the recovery of non-linear optics is achievable within the capabilities of the current beam position monitor system. We demonstrate the application of DLMN results through simulated resonance diagrams in tune space and their comparison with measurements. The DLMN provides a novel tool for analyzing the causal origins of resonances and exploring potential compensation schemes.