Advanced methods for storage ring nonlinear beam dynamics design and implementation Song, Minghao Physics Autoresonance Design optimization Gaussian process Nonlinear beam dynamics Online optimization Storage ring To meet the increasing demands of scientific researchers for brighter photonbeams, storage ring beam emittance is continually pushed down to a new ultra-low level. It, therefore, becomes correspondingly more challenging to ensure such storage rings have good nonlinear beam dynamics performance. This thesis work is focused on developing advanced methods for low emittance storage ring nonlinear beam dynamics design and implementation.Nonlinear beam dynamics optimization is essential to low emittance storagering design. A highly efficient multi-objective optimization algorithm is needed to simultaneously achieve a large dynamic aperture and a large local momentum aperture. Work was done to improve and test a machine learning-based algorithm called multi-generation Gaussian process optimizer (MG-GPO). This advanced method uses constructed GP models to pre-select solutions, and benchmarking of results on toy problems shows that MG-GPO converges significantly faster than traditional algorithms. The MG-GPO algorithm was successfully applied to nonlinear lattice design optimization, for example, to the SPEAR3 upgrade 7-nm lattice, and it was demonstrated to converge faster than NSGA-II and MOPSO. This was due to its capability of selecting candidates that tend to have better performance. This algorithm will help accelerate nonlinear lattice studies.Correction of nonlinear beam dynamics is also important for low emittancestorage ring commissioning and operation. In order to measure and correct features relevant to the nonlinear beam dynamics, an effective method is needed to excite sustained beam oscillations to large amplitude. A method based on the concept of autoresonance was proposed. This advanced technique excites nonlinear transverse beam motion in storage rings by sweeping the drive frequency. The theory for the autoresonance threshold was derived for the nonlinear optics systems in storage rings, both with and without damping effects, using Hamiltonian dynamics. The theoretical predictions for the drive amplitude threshold were found to agree well with simulations for a simple storage ring model, as well as for simulations with the actual SPEAR3 and APS lattices. The theory was also compared favorably to historical data from experiments on SPEAR3. Simulations verified that an oscillation driven by autoresonant excitation matches the character of a free oscillation, so that beam oscillation data taken during the ramping process can confidently be used to characterize the nonlinear beam dynamics performance. The precision of measurements can be improved by using autoresonant excitation since large amplitude beam oscillations are sustained significantly longer. Simulations of autoresonant excitation demonstrated the measurements of the detuning coefficients and resonance driving terms. The use of autoresonant excitation for the detection of faulty magnets and correction of resonance driving terms was also demonstrated.Online optimization is an alternative way to effectively improve nonlinear beamdynamics performance in a real storage ring. The greater efficiency of an advanced optimization algorithm is also needed to find globally optimal solutions in the limited experimental time that is typically available. The MG-GPO algorithm was implemented for SPEAR3 vertical emittance minimization and injection efficiency optimization. Again, the optimized solutions demonstrate that MG-GPO is more efficient than the commonly used PSO algorithm. SPEAR3 performance was successfully improved during the online optimization runs with MG-GPO. Spentzouris, Linda 2022 Dissertation application/pdf islandora:1024949 http://hdl.handle.net/10560/islandora:1024949 Illinois Institute of Technology PHYS / Physics en In Copyright http://rightsstatements.org/page/InC/1.0/ Restricted Access