Speaker
Description
In computational plasma physics kinetic models are used to simulate plasma phenomena where small scale physics is expected to be of importance. These models contain the full information of the particle velocity distribution function but are computationally expensive. Therefore, computationally cheaper models are utilized, which can then be deployed to larger scales, e. g. 10-moment fluid models or magnetohydrodynamics (MHD). However, the large scale behavior can be influenced by small scale processes (Verscharen et al., 2022). Thus, models are required that can include kinetic processes, in reduced form, into large scale simulations. At the moment, analytical closures are used to close the hierarchy of fluid equations, but these closures are strictly valid only in certain regimes. Finding suitable closure equations is an ongoing research topic that gets increasingly more difficult in complex systems. In this study, we try to improve fluid models by learning a suitable symbolic closure applying machine learning methods to data from kinetic simulations.\par
At first, less complex physical settings are chosen to be able to compare results with theory, such as Landau damping and the Kelvin-Helmholtz instability. For these test cases the robustness of the method to the choice of parameters and data selection is investigated. Hereby, we aim to reproduce results from previous works, e. g. Long et al. (2019) and Cheng et al. (2023). In the long term the method will be applied to more complex physical settings and could improve existing closure equations and provide insights into the underlying physics.
