Systematic investigation and improvement of LLRF controls for LINACs and cyclotrons

该提案要求资金支持加速器科学的研究和开发计划。该计划旨在开发创新的解决方案，以解决射频腔控制中的三个主要挑战，具体如下：1.微音抑制对于提高光束质量至关重要，这被许多以连续波（CW）模式运行的加速器视为主要挑战。加速器已经通过向安装的快速作用压电元件反馈相位误差实现了2-8倍的抑制。为了进一步改善抑制，必须了解不同的机械模式如何影响光束。评估不同模式的RF分析将提供关于需要主动抑制的模式的信息。系统的机械结构的分析进一步提供了关于每个模式的可观测性和可控性的必要信息。后两点提供了使用压电元件的主动颤噪控制的限制，这是加速器社区所不知道的。基于并置传感器/致动器的可能的控制系统的评估可以提高精度。 2.TRIUMF的e-LINAC将在CW模式下运行。加速腔以矢量和模式操作。虽然CW操作和矢量和控制是可以在加速器中找到的操作概念，但两者的结合是相当新的，并伴随着新的挑战。洛伦兹力引起的参数振荡可能是RF腔的操作梯度的限制因素。分析耦合腔中洛仑兹力引起的振荡对于理解这种系统的稳定性是必要的。将开发一种依赖于同位传感器/致动器对的主动振荡消除系统，以增加系统的稳定性，这可能适用于全球其他加速器实验室。3.回旋加速器中的火花经常发生。它们必须被检测到并做出反应，以最大限度地保护机器并最大限度地减少停机时间。机器学习在粒子加速器领域的发展势头越来越大，因为这些系统的不同操作领域以及好处正在评估和学习。故障诊断对于大型和复杂的机器尤其重要。我们建议研究基于机器学习的火花检测系统。前两点解决了RF超导腔的场稳定性，从而提高了射束稳定性，第三点解决了加速器LLRF控制的人工智能。随着经验的积累，它可以应用于上述项目部分，以更智能的数据收集和控制。该项目为科学和工程领域的年轻研究人员提供了一个完美的培训环境。该提案要求提供资金支持研究生的工资，并提供资金支持年轻研究人员参加会议的旅费。

This proposal requests funds to support a program in research and development in accelerator science. The program is aimed at developing innovative solutions to address three major challenges in the control of RF cavities as follows: 1.Microphonics suppression is crucial to improve the beam quality, which is considered as a major challenge by many accelerators operating in continuous wave (CW) mode. Accelerators have achieved a suppression of a factor of 2-8 by phase error feedback to an installed fast acting piezo element. To improve the suppression further, how the different mechanical modes affect the beam must be understood. An RF analysis evaluating the different modes will provide the information regarding the modes that need to be actively suppressed. An analysis of the mechanical structure of the system further provides the necessary information regarding observability and controllability of each mode. The latter two points provide the limitations of active microphonics control with a piezo element, which are unknown to the accelerator community. An evaluation of a possible control system based on a collocated sensor/actuator can improve the accuracy. 2.TRIUMF's e-LINAC will be operated in the CW mode. The acceleration cavities are operated in vector sum mode. While CW operation and vector sum control are operational concepts which can be found in accelerators, the combination of the two is rather new and accompanied by new challenges. Lorentz force induced parametric oscillations can be a limiting factor for the operational gradient of RF cavities. An analysis of Lorentz force induced oscillations in coupled cavities is necessary to understand the stability of such systems. An active oscillation cancellation system that relies on a collocated sensor/actuator pair will be developed to increase the system stability, which may be applicable to other accelerator laboratories worldwide. 3.Sparks in cyclotrons occur on a regular basis. They have to be detected and reacted upon for maximum machine protection and minimal down time. Machine learning is gaining momentum in the field of particle accelerators as the different operational areas for such systems and the benefits are evaluated and learned. Fault diagnostics, can be especially important for large and complex machines. We are proposing to investigate a spark detection system based on machine learning. The first two points address the field stability of an RF superconducting cavity and hence improving the beam stability, the third point addresses artificial intelligence towards accelerator LLRF control. With experience gained it can be applied to the above project parts in terms of more intelligent data collection and control. The proposed project is a perfect training environment for young researchers in the fields of science and engineering. The proposal requests funds to support graduate students' salaries and funds to support travel to conferences for young researchers.