Accelerator Diagnosis and Control Based on Low Energy Empirical Model

基于低能经验模型的加速器诊断与控制

• 批准号: SAPPJ-2014-00028
• 负责人: Chao, YuChiu
• 金额: $ 1.46万
• 依托单位: TRIUMF
• 依托单位国家: 加拿大
• 项目类别: Subatomic Physics Envelope - Project
• 财政年份: 2014
• 资助国家: 加拿大
• 起止时间: 2014-01-01 至 2015-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=568026
• 关键词: Accelerator; Diagnosis; Control; Based; Low

This proposal aims to develop and validate a new approach to creating online accelerator models, and to further establish it as the basis for accelerator diagnostic and control applications in realistic operation and commissioning environments, especially at the Advanced Rare Isotope Laboratory (ARIEL) project at TRIUMF. The ability to correctly and efficiently model accelerator transport and beam dynamics is critical to the success of any real-time online diagnosis and control task. This is even more so at low energy where key performance parameters are often “frozen in” and too late to change at higher energy. Unfortunately, efficiency and correctness are not always compatible objectives for low energy (low-beta) modeling. In many cases faster analytic algorithms are used online to model effects at low energy, producing incorrect physics. We propose to bridge this gap with an online empirical model to be described later. Failure of analytical models for low-beta beam transport arises from the following: • Velocity changes within cavities • Non-relativistic dynamics, naïve field-rigidity scaling, and arrival-time dependence • RF waveform detail defying sinusoidal approximation but important at low energy • Significant longitudinal-transverse momentum transfer • Higher order aberrations in solenoids and in cavities • Strong impact of chromaticity due to large momentum spread • Elements inside fringe fields of other elements • Increased need to account for space charge at low energy, possibly with subdivided elements • Nontrivial interplay between all phase space dimensions preventing reduction/decoupling of model • Algorithm deficiency to deal with time-varying elements at off-design phase • Inability of S-coordinate based formulation to capture time-of-flight at element boundaries • Any unconventional element defying accurate analytical treatment (now limited to low-beta) While in the highly relativistic regime most analytical accelerator and beam modeling tools can be free of these problems, toward the non-relativistic end, which can encompass range of beta from 0 up to 0.999 or beyond, this dilemma between speed and accuracy is very real, often with detrimental effects. The solution put forth in the current proposal, henceforth referred to as the low energy (low beta) empirical model, attempts to mitigate this dilemma by adopting an efficient empirical approach, independent of the nature and detail of the physical process being modeled as long as a dedicated modeling tool exists that correctly models the process of interest as functions of a limited number of control and beam parameters. The modeling tool can account for all pertinent physical details without need of closed-form formulation or approximation to physical processes. The outcome is rendered into efficient numerical models that can quickly and accurately predict detailed machine and beam behavior. This process is conceptually simple and scalable, and can be applied to any modeling tool for any component. Possible directions in which to extend this approach are many. Basically, complexity in underlying physics does not present additional levels of conceptual difficulty for this approach. Experience gained and infrastructure developed within the current proposed scope of low energy empirical model can serve as the basis for more ambitious projects such as empirical model based on real data from controlled beam based measurements or statistically accumulated machine data. Such development can lead to the capability of deterministically and efficiently controlling the traditionally most volatile sections of an accelerator.

该提案旨在开发和验证一种新的方法来创建在线加速器模型，并进一步将其作为加速器诊断和控制应用的基础，特别是在TRIUMF的高级稀有同位素实验室（ARIEL）项目中。正确有效地模拟加速器传输和束流动力学的能力对于任何实时在线诊断和控制任务的成功至关重要。这在低能量下更是如此，其中关键性能参数通常被“冻结”，并且在较高能量下改变太晚。不幸的是，效率和正确性并不总是低能量（低β）建模的兼容目标。在许多情况下，在线使用更快的分析算法来模拟低能量下的效应，产生不正确的物理学。我们建议用稍后描述的在线经验模型来弥合这一差距。低β射束输运的分析模型的失败源于以下原因：·非相对论动力学，朴素场刚性标度，和到达时间依赖性· RF波形细节不服从正弦近似，但在低能量下很重要·显著的纵向-横向动量传递·在谐振腔和腔中的高阶像差·由于大的动量扩散导致的色度的强烈影响·其他元素的边缘场内的元素·增加了在低能量下考虑空间电荷的需要，·所有相空间维度之间的非平凡相互作用防止模型的简化/解耦·在非设计阶段处理时变元素的算法缺陷·基于S坐标的公式无法捕获元素边界处的飞行时间·任何非常规元素无法进行精确的分析处理（现在限于低β）虽然在高度相对论性的制度，大多数分析加速器和束建模工具可以没有这些问题，朝向非相对论性的端部（其可以包括从0到0.999或更高的β的范围），速度和准确性之间的这种困境是非常真实的，通常具有有害的影响。在当前提议中提出的解决方案（此后称为低能量（低β）经验模型）试图通过采用有效的经验方法来缓解这种困境，该方法独立于被建模的物理过程的性质和细节，只要存在将感兴趣的过程正确地建模为有限数量的控制和射束参数的函数的专用建模工具。建模工具可以考虑所有相关的物理细节，而不需要封闭形式的公式或近似的物理过程。结果被渲染成高效的数值模型，可以快速准确地预测详细的机器和梁的行为。这个过程在概念上是简单的和可伸缩的，并且可以应用于任何组件的任何建模工具。推广这一方法的可能方向有很多。基本上，基础物理学的复杂性不会为这种方法带来额外的概念难度。在当前提出的低能量经验模型范围内获得的经验和开发的基础设施可以作为更雄心勃勃的项目的基础，例如基于受控射束测量的真实的数据或统计积累的机器数据的经验模型。这样的发展可以导致确定性和有效地控制加速器的传统上最不稳定的部分的能力。