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发展项目。该计划旨在开发创新的解决方案，以解决射频空腔控制中的三大挑战：微音抑制是提高束流质量的关键，也是许多工作在连续波（CW）模式下的加速器面临的主要挑战。通过对安装的快速压电元件的相位误差反馈，加速器实现了2-8倍的抑制。为了进一步改善抑制，必须了解不同的力学模式如何影响梁。评估不同模式的射频分析将提供有关需要主动抑制的模式的信息。对系统机械结构的分析进一步提供了关于每种模式的可观测性和可控性的必要信息。后两点提供了用压电元件进行主动传声器控制的局限性，这是加速器界所不知道的。对基于传感器/执行器配置的可能控制系统进行评估可以提高精度。2.TRIUMF的e-LINAC将以CW模式运行。加速腔以矢量和模式操作。虽然连续波运算和矢量和控制是在加速器中可以找到的操作概念，但两者的结合是相当新的，并且伴随着新的挑战。洛伦兹力引起的参量振荡是射频腔工作梯度的限制因素。对耦合腔中洛伦兹力引起的振荡进行分析是理解这类系统稳定性的必要条件。为了提高系统的稳定性，将开发一种依赖于传感器/执行器对的主动振荡抵消系统，该系统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是定期发生的。它们必须被检测和反应，以最大限度地保护机器和最小的停机时间。机器学习在粒子加速器领域正在获得动力，因为这些系统的不同操作领域及其好处正在被评估和学习。故障诊断对于大型和复杂的机器尤其重要。我们建议研究一种基于机器学习的火花检测系统。前两点解决了射频超导腔的场稳定性，从而提高了光束稳定性，第三点解决了加速器LLRF控制的人工智能。随着经验的积累，可以将其应用于上述项目部分，实现更智能的数据采集和控制。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。该提案要求资助研究生的工资和资助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.