Machine Learning Integration of CSR and Space-Charge Calculations

Advanced accelerator concepts and their applications require increasingly sophisticated control over the beam distribution. Plasma and structure wakefield accelerators benefit from precise longitudinal beam shaping, while applications such as free-electron lasers require significant compression. Bunch compression through magnetic chicanes, while highly effective, often leads to deleterious effects such as the generation of coherent synchrotron radiation (CSR). Moreover, at low energies, the interplay between CSR and space-charge becomes important when trying to preserve the beam brightness for use in HEP experiments. Our efforts are aimed at advancing the state of the art both in how CSR and space-charge affect the beam dynamics for ultra-short, high-brightness bunches, and in speeding up the computation of these effects for more efficient optimization. This talk will focus on recent results aimed at utilizing machine learning to speed up the evaluation of CSR effects. We will also detail our plans to integrate CSR and space-charge calculations into a unitary ML based solver that can replace traditional Lienard-Wiechert solvers in particle tracking codes.