Mitigation of ground motion effects via feedback systems in the Compact Linear Collider

The Compact Linear Collider (CLIC) is a future multi-TeV electron positron collider, which is currently being designed at CERN. To achieve its ambitious goals, CLIC has to produce particle beams of the highest quality, which makes the accelerator very sensitive to ground motion. Four mitigation methods have been foreseen by the CLIC design group to cope with the feasibility issue of ground motion. This thesis is concerned with the design of one of these mitigation methods, named linac feedback (L-FB), but also with the simultaneous simulation and validation of all mitigation methods. Additionally, a technique to improve the quality of the indispensable system knowledge has been developed. The L-FB suppresses beam oscillations along the accelerator. Its design is based on the decoupling of the overall accelerator system into independent channels. For each channel an individual compensator is found with the help of a semi- automatic control synthesis procedure. This technique allows the designer to incorporate expert knowledge, which is used by an optimisation algorithm to minimise the luminosity loss due to ground motion. This approach speeds up the design process significantly, while at the same time improving the orbit feedback performance compared to standard methods. Beside the L-FB, simple but effective designs for the interaction point feedback and cost reduction options for the quadrupole stabilisation are presented. For the design of all these feedback systems models of the ground motion influence on different beam parameters such as beam offset, beam size and luminosity have been derived by adapting and extending existent models. To design, improve and validate the ground motion mitigation methods, a simulation framework was set up, which includes a ground motion generator, beam tracking, beam-beam interaction and all mitigation methods. The simulations show that the ground motion mitigation methods can efficiently preserve the CLIC luminosity. Due to our design of the L-FB, the specifications of the beam position monitor resolution could be relaxed significantly. The robustness of the L-FB was also verified with respect to many other imperfections. Only a certain sensitivity to beam energy variations was observed, which could be resolved by filtering dispersive orbits from the measurements. Further simulation results were an essential input for the redesign of the quadrupole stabilisation system leading to a significant performance improvement of the system. Due to the high importance of the system knowledge for many applications, a system identification scheme was developed. It is capable of adapting the parameters of the system model to changes of the main linac behaviour (orbit response matrix) on-line, during the regular operation of the linac. By focusing only on the most significant system changes, the identification speed could be improved strongly compared to standard algorithms. The identified parameters of the orbit response matrix can be used to improve the performance of beam-based alignment algorithms and orbit feedbacks and are an important input for diagnosis and error detection tools.