The European X-Ray Free-Electron Laser Facility (European XFEL) is an X-ray
research laser facility commissioned during 2017. A free-electron laser
generates high-intensity electromagnetic radiation by accelerating electrons
to relativistic speeds and directing them through special magnetic structures.
European XFEL will be an international research facility of superlatives: 27000
X-ray flashes per second and a brilliance that is a billion times higher than
that of the best conventional X-ray sources will open up completely new
opportunities for science. Research groups from around the world will
investigate nanometer-scale structures, fast processes, and extreme states,
taking 3D images of viruses and proteins, and filming chemical reactions.
Creating the world's most brilliant X-ray beam is a technological challenge.
The beam, originally an electron beam that is used to create the photon
beam, needs to be controlled along its path to reach the end of the tunnel,
and needs to be appropriately focused and to be in the right place for the
experiment. The process of so-called 'beam alignment' can take up to an hour
or longer and has to be carried out at the beginning of every experimental
shift by skilled staff. In this project, we aim to automate the process to reduce
time requirements and increase the stability and reliability of this process.
In cooperation with University of Szeged, this project will use the machine learning technology to establish a system to
automatically fulfill the task beam alignment before each experimental shift.
To carry out the project, the alignment will be first investigated and
documented to help establish understanding of the process and gather as
many precedents for the later machine learning as possible. The proper
machine learning algorithms will be chosen and the alignment system will be
designed based on the mechanism. The system will later be developed,
optimized based on the precedents and the machine learning process. Then,
the developed system will be test evaluated and optimized in experiments
before put into use. The data and experience gathered in the project may also
be useful to automate classification of incorrect beam alignment, or detect
arising hardware problems early.
Topics: control, optimisation, machine learning, Automation and Control Systems; Data Analytics & Machine learning
előírt nyelvtudás: English felvehető hallgatók száma: 1
Jelentkezési határidő: 2019-06-30
2024. IV. 17. ODT ülés Az ODT következő ülésére 2024. június 14-én, pénteken 10.00 órakor kerül sor a Semmelweis Egyetem Szenátusi termében (Bp. Üllői út 26. I. emelet).
Bejelentkezés
