强流高功率直线加速器基于空间电荷效应的发射度增长及其控制研究

High-current, high-power linacs are strict in beam loss. Due to the serious space charge effect and the significant influences of beam currents and duty cycles on facilities, the space charge field and the structure field perform nonlinear effect, which would cause both emittance growth and beam loss. Based on the superconducting linac of the Strategic Pilotity Research Program, the emittance growth and control of high-current, high-power linacs will be analyzed around space charge effect, and during this procedure, the engineering error should be taken into account. Firstly, a PIC code integrating design, simulation and optimization will be developed on GPU platform for post-process convenience and calculation time. Secondly, the general factors that result in emttance growth will be analyzed; And the beam halo formation and transmission in the shaper section of RFQ will be studied innovatively. Thirdly, beam dynamics will be optimized to obtain the general emittance control methods including beam halo collimation and emittance redistribution; Explorations of high-order emittance match and "excitation - scraping - deexcitation" will be studied with an open mind. Fourthly, by digging out the cloud mass data and managing the Pareto optimal set by neural network, an off-line Lattice database will be built. The methodology, simulation and optimization of emittance growth and control will be applied to CIADS and LEAF.

强流高功率直线加速器对束流损失要求严格，但其空间电荷效应严重、不同流强和占空比对设备运行影响很大，容易造成内场和外场的非线性，引起发射度增长，造成束流损失。本项目拟依托ADS先导专项直线加速器，以空间电荷效应为中心，考虑工程误差，对强流高功率直线加速器基于空间电荷效应的发射度增长及控制进行研究。首先从便利操作和计算时间考虑，拟开发GPU平台的PIC设计、模拟和优化程序。其次研究发射度增长的一般性因素，并考虑RFQ俘获聚束段的势阱对发射度增长的影响。然后优化束流动力学，获得发射度控制的刮束和发射度重分配方案，同时创新性地研究发射度的高阶匹配和“激励-刮束-反激励”方案。最后挖掘海量云储存数据，通过神经网络管理Pareto解集，搭建离线Lattice数据库。通过项目拟掌握发射度增长及控制的研究方法、模拟技巧和优化策略，为CIADS和LEAF的设计、模拟和优化提供束流物理基础。

强流高功率直线加速器中，发射度的增长会造成束损，影响加速器的运行，因此对发射度的控制非常重要。为了更快更精确地模拟强流束流，使用python语言和机器学习库pytorch开发了基于CPU和GPU双平台的并行强流束流动力学多粒子多束流PIC模拟程序。本项目研究了在ADS超导直线中引起束流发射度增长的主要因素：初始束流分布、束流失配和共振。利用宏粒子的增减，开发了RFQ出口束流分布的重建程序。研究了4阶不相干共振、2阶相干共振、束核主共振和束晕主共振。在发射度控制研究中，首先开发了强流高功率加速器中的全发射度计算程序，然后提出了旋转BPM的数学原理并开发了程序用于验证，最后研究了通过束晕耦合的方法进行刮束。基于真实加速器，使用云存储数据，实现了深度学习Lattic设置，进行了实时轨道校正。束损和稳定运行是强流高功率加速器面临的主要问题。该课题对发射度增长和控制的研究，为ADS超导强流直线加速器解决束损问题提供支持；该课题通过深度机器学习利用云存储数据对Lattice的设定，对ADS超导强流直线加速器未来的稳定运行研究积累经验。