Search

The research carried out in particle accelerator facilities does not concern only particle and condensed matter physics, although these are the main topics covered in the field. Indeed, since a particle accelerator is composed of many different sub-systems, its proper functioning depends both on each of these parts and their interconnection. It follows that the study, implementation, and improvement of the various sub-systems are fundamental points of investigation too. In particular, an interesting aspect for the automation engineering community is the control of such systems that usually are complex, large, noise-affected, and non-linear. The doctoral project fits into this scope, investigating the introduction of new methods to automatically improve the performance of a specific type of particle accelerators: seeded free-electron lasers. The optimization of such systems is a challenging task, already faced in years by many different approaches in order to find and attain an optimal working point, keeping it optimally tuned despite drift or disturbances. Despite the good results achieved, better ones are always sought for. For this reason, several methods belonging to reinforcement learning, an area of machine learning that is attracting more and more attention in the scientific field, have been applied on FERMI, the free-electron laser facility at Elettra Sincrotrone Trieste. The research activity has been carried out by applying both model-free and model-based techniques belonging to reinforcement learning. Satisfactory preliminary results have been obtained, that present the first step toward a new fully automatic procedure for the alignment of the seed laser to the electron beam. In the meantime, at the Conseil Européen pour la Recherche Nucléaire, CERN, a similar investigation was ongoing. In the last year of the doctoral course, a collaboration to share the knowledge on the topic took place. Some of the results collected on the largest particle physics laboratory in the world are presented in the doctoral dissertation.

The CERN H− linear accelerator, LINAC4, served as a test bed for advanced algorithms during the CERN Long Shutdown 2 in the years 2019/20. One of the main goals was to show that reinforcement learning with all its benefits can be used as a replacement for numerical optimization and as a complement to classical control in the accelerator control context. Many of the algorithms used were prepared before- hand at the electron line of the AWAKE facility to make the best use of the limited time available at LINAC4. An overview of the algorithms and concepts tested at LINAC4 and AWAKE will be given and the results discussed., peer-reviewed

Reinforcement Learning (RL) is one of the most promising techniques in Machine Learning because of its modest computational requirements with respect to other algorithms. RL uses an agent that takes actions within its environment to maximize a reward related to the goal it is designed to achieve. We have recently used RL as a model-free approach to improve the performance of the FERMI Free Electron Laser. A number of machine parameters are adjusted to find the optimum FEL output in terms of intensity and spectral quality. In particular we focus on the problem of the alignment of the seed laser with the electron beam, initially using a simplified model and then applying the developed algorithm on the real machine. This paper reports the results obtained and discusses pros and cons of this approach with plans for future applications., Proceedings of the 17th International Conference on Accelerator and Large Experimental Physics Control Systems, ICALEPCS2019, New York, NY, USA

We report on the successful operation of a Free Electron Laser (FEL) in the Echo Enabled Harmonic Generation (EEHG) scheme at the FERMI facility at Sincrotrone Trieste. The experiment required a modification of the FEL-2 undulator line which, in normal operation, uses two stages of high-gain harmonic generation separated by a delay line. In addition to a new seed laser, the dispersion in the delay-line was increased, the second stage modulator changed and a new manipulator installed in the delay-line chicane hosting additional diagnostic components. With this modified setup we have demonstrated the first evidence of strong exponential gain in a free electron laser operated in EEHG mode at wavelengths as short as 5 nm., Proceedings of the 39th Free Electron Laser Conference, FEL2019, Hamburg, Germany

Despite the large number of feedback loops running simultaneously at the FERMI Free Electron Laser (FEL), they are not sufficient to keep the optimal machine working point in the long term, in particular when the machine is tuned in such a way to be more sensitive to drifts of the critical parameters. In order to guarantee the best machine performance, a novel software application which minimizes the shot to shot correlation between these critical parameters and the FEL radiation has been implemented. This application, which keeps spatially and temporally aligned the seed laser and the electron beam, contrary to many algorithms that inject noise in the system to be optimized, run transparently during the experiment beam times. In this paper we will also present a newly developed method to calculate a beam 'quality factor' starting from the images provided by a photon spectrometer, which tries to mimic the evaluation of machine physicists, as well as the results obtained using two model-less algorithms to optimize the FEL performance through maximization of the quality factor., Proceedings of the 16th Int. Conf. on Accelerator and Large Experimental Control Systems, ICALEPCS2017, Barcelona, Spain

Nature photonics 13(8), 555 - 561 (2019). doi:10.1038/s41566-019-0427-1, X-ray free-electron lasers (FELs), which amplify light emitted by a relativistic electron beam, are extending nonlinear optical techniquesto shorter wavelengths, adding element specificity by exciting and probing electronic transitions from core levels. Thesetechniques would benefit tremendously from having a stable FEL source, generating spectrally pure and wavelength-tunable pulses.We show that such requirements can be met by operating the FEL in the so-called echo-enabled harmonic generation (EEHG) configuration.Here, two external conventional lasers are used to precisely tailor the longitudinal phase space of the electron beambefore emission of X-rays. We demonstrate high-gain EEHG lasing producing stable, intense, nearly fully coherent pulses at wavelengthsas short as 5.9 nm (~211 eV) at the FERMI FEL user facility. Low sensitivity to electron-beam imperfections and observationof stable, narrow-band, coherent emission down to 2.6 nm (~474 eV) make the technique a prime candidate for generatinglaser-like pulses in the X-ray spectral region, opening the door to multidimensional coherent spectroscopies at short wavelengths., Published by Nature Publ. Group, London [u.a.]

The radiation generated by a seeded free-electron laser (FEL) is characterized by a high temporal coherence, which is close to the Fourier limit in the ideal case. The setup and optimization of a FEL is a non-trivial and challenging operation. This is due to the plethora of highly sensitive machine parameters and to the complex correlations between them. The fine tuning of the FEL process is normally supervised by physicists and is carried out by scanning various parameters with the aim of optimizing the spectrum of the emitted pulses in terms of intensity and line-width. In this article we introduce a novel quantitative method for the evaluation of the FEL spectrum via a quality index. Moreover, we investigate the possibility of optimization of the FEL parameters using this index as the objective function of an automatic procedure. We also present the results of the preliminary tests performed in the FERMI FEL focused on the effectiveness and ability of the automatic procedure to assist in the task of machine tuning and optimization.