Search

This paper presents a new approach to space charge dominated beamline design using an artificial intelligence (AI)-based optimization code named giotto. The code incorporates advanced algorithms for multiobjective genetic optimizations in particle accelerators, allowing efficient exploration of the parameter space and improved beam quality. The study demonstrates the application of giotto in the design of a high-brightness injector for an Energy Recovery Linac (ERL) called BriXSinO. The optimized injector features a low energy (4.5 MeV) and relatively high bunch charge (100 pC) operation. The results show promising beam parameters comparable to other ERL projects. Furthermore, the paper introduces innovative techniques, including bunches back-rotation the use of Lorentzian distributions in the fitness function. The approach successfully achieves dispersion closure in a space charge dominated dogleg. Overall, this work contributes to the advancement of accelerator science, offering a powerful methodology for beamline design and optimization. The new techniques and methodologies introduced have the potential to enhance the performance and stability of particle accelerators in various applications.