Machine learning based closures for high-collisionality deuterium-carbon plasmas

Various plasmas in laboratory and space environments consist of multiple ion species. For instance, in tokamak edge plasmas, the presence of ionized impurities emitted from the material surface has a significant impact on plasma transport phenomena. To describe multi-ion plasmas using fluid equations, precise closure relations are required to close the system of fluid equations. In this work, we introduce the development of fitting formulas for parallel closures using a machine learning algorithm, incorporating recent advancements in closure theory for multi-ion plasmas [Plasma Phys. Control. Fusion 65, 075014 (2023)]. The effectiveness of this approach is demonstrated by applying it to a high-collisionality deuterium-carbon plasma. The machine learning-based method proves to be a practical and accurate approach for developing closure relations and can be extended to a broader range of plasma systems.