Systematic investigation and improvement of LLRF controls for LINACs and cyclotrons

这项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，以应对以下三个主要挑战：1.微波抑制是提高光束质量的关键，这被许多运行在连续波(CW)模式下的加速器视为主要挑战。加速器通过将相位误差反馈到已安装的快速作用压电元件，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对梁的影响。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。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。后两点提供了压电式元件的有源微音学控制的局限性，这对于加速器社区来说是未知的。对基于传感器/执行器配置的可能的控制系统进行评估可以提高精度。2.TRIUMF的e-直线加速器将以连续波模式运行。加速腔以矢量和模式运行。虽然CW运算和矢量和控制是在加速器中可以找到的运算概念，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射频腔工作梯度的一个限制因素。为了了解这类系统的稳定性，有必要对耦合腔中洛伦兹力引起的振荡进行分析。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消除系统，这可能适用于世界各地的其他加速器实验室。3.回旋加速器中的火花定期发生。必须对它们进行检测和反应，以实现最大限度的机器保护和最短的停机时间。机器学习正在粒子加速器领域获得动力，因为这种系统的不同操作领域及其好处得到了评估和学习。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。前两点涉及射频超导腔的场稳定性，从而提高束流稳定性，第三点涉及人工智能对加速器LLRF的控制。随着经验的积累，它可以在更智能的数据采集和控制方面应用于上述项目部分。建议的项目为科学和工程领域的青年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来支持研究生的工资，并为年轻研究人员参加会议提供资金。

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.