Phenomenological model of a free-electron laser using machine learning

A.M. Kalitenko

The simulation of free-lectron lasers (FELs) requires complex programs and qualified personal to adjust the numerical parameters. A numerical modeling is not always close to experiments, and many programs are unavailable or inaccessible for public use. To design new facilities, it is necessary to estimate the size, cost of the FEL, as well as the radiation characteristics. Therefore, it is necessary to consider a lot of FEL configurations before starting construction or upgrading existing systems. This paper presents a technique for building a neural network for analyzing FEL parameters. We collected numerical simulation data of about 2000 configurations, built and trained a neural network that can analyze several FEL configurations in a short time depending on the undulator and the electron beam parameters. This technique can easily be extended to more complex systems and applied to real facilities to take into account the individual characteristics of facilities.

Cite the code:

Article: https://iopscience.iop.org/article/10.1088/1402-4896/acf814