Speaker
Description
Beam transport line of uniformization is an essential step before beam shooting and plays a critical role in applications such as neutron imaging and isotope production. In practical experiments, particle beams generated by cyclotrons typically exhibit an approximately Gaussian intensity distribution. Local bright spots on the targets for high-power Gaussian beams create difficult cooling problems and shorten the lifetime of the target. To address this issue, a beam transport line with multiple beam elements is designed. Considering the complexity of multi-parameter optimization process, an automatic optimization method based on deep reinforcement learning is investigated. In the simulation environment, the Soft Actor-Critic ( SAC ) algorithm is employed to adjust multiple beam elements for optimizing the uniformity of the beam. Results indicate that the SAC algorithm based on the maximum entropy principle achieves over 85% uniformity on the target. Compared with manually optimizing parameters, the proposed method demonstrates higher efficiency and superior uniformization performance.
