Towards a Data-driven Digital Twin of a Free Electron Laser

Sources of soft X-rays are highly appealing in research as they allow to image atomic- and
molecule- scaled structures, however high requirements to technical equipment complicate
application of such systems. Free electron laser is one of famous sources of ultra-intense
coherent X-ray beams. Convenient kilometer-scale electron accelerators make these
facilities expensive and difficult to maintain, while laser-driven electron accelerators might
significantly reduce size of free-electron lasers. In order to control such a source of X-rays
there are required time consuming numerical and experimental research. A rising demand
on statistical and mathematical methods for inversion of the system state, comprehension of
measurement data and quantification of data stability can only be met by a comprehensive
machine learning based digital twin for Free Electron Laser. The digital twin potentially
accelerates theoretical comprehension of the system, novel means for design space
exploration and promises reliable in-situ analysis of experimental diagnostics and
parameters which leads to democratization of laser-driven FELs accelerating fundamental
science in research field MATTER by collaborative efforts in Matter and Technologies. Digital
twin is comprising of multiple surrogate models for electron acceleration processes by virtue
of that one could unveil beam dynamics on the scope of collected diagnostic. This
formulation allows us to derive observables within the beamline promising physics-informed
inversion of the beamline meaning that we are able to explain observations guided by our
theoretical understanding.