Machine learning-based longitudinal phase space prediction of particle accelerators

Self Cite

124

The National User Facility for Advanced Accelerator Experimental Tests II (FACET-II) at SLAC National Accelerator Laboratory expands upon the experiments conducted at FACET. Its purpose is to build upon the decades-long experience developed conducting accelerator R&D at SLAC in the areas of advanced acceleration and coherent radiation techniques with high-energy electron and positron beams. This paper summarizes the motivations for the design and resulting capabilities of the FACET-II facility.

Section: Introductionmentioning

confidence: 99%

FACET-II facility for advanced accelerator experimental tests

Yakimenko

Alsberg

Bong

et al. 2019

Self Cite

124

“…Machine operators could use these models to check the potential impact of setting changes before trying them out on the actual machine, or to assess a new course of action as goals change during an operating shift. These models could also be used as a diagnostic tool to provide predictions about unmeasured beam parameters (i.e., as a virtual diagnostic [21,26,46]) or to flag when the system has changed substantially (i.e., model-based anomaly detection). They could also be exploited in model-based control and model-guided optimization routines (i.e., using the model to help guide the search for optimal settings, as is done in model predictive control and Bayesian optimization [47]).…”

Section: A Incorporation Into On-line Modeling and Model-based Controlmentioning

confidence: 99%

Machine learning for orders of magnitude speedup in multiobjective optimization of particle accelerator systems

Edelen¹,

Neveu²,

Frey³

et al. 2020

High-fidelity physics simulations are powerful tools in the design and optimization of charged particle accelerators. However, the computational burden of these simulations often limits their use in practice for design optimization and experiment planning. It also precludes their use as on-line models tied directly to accelerator operation. We introduce an approach based on machine learning to create nonlinear, fastexecuting surrogate models that are informed by a sparse sampling of the physics simulation. The models are Oð10 6 Þ-Oð10 7 Þ times more computationally efficient to execute. We also demonstrate that these models can be reliably used with multiobjective optimization to obtain orders-of-magnitude speedup in initial design studies and experiment planning. For example, we required 132 times fewer simulation evaluations to obtain an equivalent solution for our main test case, and initial studies suggest that between 330-550 times fewer simulation evaluations are needed when using an iterative retraining process. Our approach enables new ways for high-fidelity particle accelerator simulations to be used, at comparatively little computational cost.

“…Powerful NN tools have also been developed for ML-based longitudinal phase space prediction of transverse deflecting cavity readings in particle accelerators, which are some of the most important diagnostics that exist for measuring a beam's longitudinal phase space [52,53]. A novel Bayesian optimization framework that uses sparse online Gaussian processes has been applied for quadrupole magnet tuning in an FEL [54].…”

Section: B Machine Learningmentioning

confidence: 99%

Model-independent tuning for maximizing free electron laser pulse energy

Scheinker

Bohler

Tomin

et al. 2019

The output power of a free electron laser (FEL) has extremely high variance even when all FEL parameter set points are held constant because of the stochastic nature of the self-amplified spontaneous emission (SASE) FEL process, drift of thousands of coupled parameters, such as thermal drifts, and uncertainty and time variation of the electron distribution coming off of the photo cathode and entering the accelerator. In this work, we demonstrate the application of automatic, model-independent feedback for the maximization of average pulse energy of the light produced by free electron lasers. We present experimental results from both the European x-ray free electron laser at DESY and from the Linac Coherent Light Source at SLAC. We demonstrate application of the technique on rf systems for automatically adjusting the longitudinal phase space of the beam, for adjusting the phase shifter gaps between the undulators, and for adjusting steering magnets between undulator sections to maximize the FEL output power. We show that we can tune up to 105 components simultaneously based only on noisy average bunch energy measurements.