Fast retuning of superconducting radiofrequency accelerating cavities in the European Spallation Source linear accelerator.

Abstract: The aim of the degree project is to continue the work of PhD student YuanShuai Qin on creating a fast retuning of the ESS accelerator lattice in the event of an accelerating cavity being offline. The main goal of the project was evaluating and improving the retuning routine created by YuanShuai Qin. The major concern is to maintain beam stability by reducing emittance growth and keeping the beam phase advance smooth. Firstly, the original Matlab code for retuning authored by YuanShuai Qin was translated to python. This work included creating a genetic algorithm used as the optimizer on the retuning process. To make sure that this had been done correctly, the transfer matrix of a small lattice containing a few cavities were compared between TraceWin, the original Matlab code and the new python code. The resulting transfer matrix of the Pyhton code was closer to TraceWin than the transfer matrix calculated from the original Matlab script. One of the major approximations used in the retuning routine was the exclusion of space charge effects. To determine if this was a too crude approximation a new lattice was created, which is the ESS lattice but matched without the effects of space charge. The Twiss parameter was accurately calculated, with TraceWin as reference, when space charge was not considered in the simulations but not when the effects were included. However, the matching routine was tested on both lattices for failed cavities at different positions in the accelerator. The main conclusion was that the beam stability was maintained only slightly better when the effects of space charge was excluded compared to the effects were included. One of the major issues, including or excluding space charge effects, was an increasing beam phase advance for the retuned lattices. This was attempted to be rectified using beam phase advance optimization in the fitness function fed to the genetic algorithm used for the retuning process. This new fitness function made the retuning extremely slow and did not improve the resulting lattice much. Therefore, the conclusion is that if the retuning routine is to be improved, space charge should be included in the optimization, perhaps by creating the transfer matrices with respect to the effects of space charge.