Speaker
Description
The particle, momentum and energy exchanges between the plasma and the neutral gas from recycling are key aspects of divertor physics. Atoms and molecules are often not in a collisional regime and a kinetic description is required to get transport right. In detached regimes, where the temperature in the divertor drops below a few eVs, the complexity of the reaction channels for neutrals increases. The Monte Carlo approach implemented in EIRENE allows to solve the kinetic problem on a 3D grid, and can accommodate both the geometrical complexity and the relevant species and reactions channels. However, in large machines such as ITER, in which the mean free path of neutrals can be very short compared to the size of the divertor, regions where neutrals become collisional can appear. Such regions make the Monte Carlo approach computationally much more costly, since atoms or molecules may undergo tens of thousands of collisions before being ionized or pumped. It also makes fluid approaches accurate at least in these regions. This presentation will focus on EIRENE simulations in ITER conditions, showing that the computational cost increase is related to physics and that sacrificing the physics is not an option if the goal is to assess e.g. peak heat fluxes in the divertor. Advanced fluid models, fully consistent with the underlying kinetic model used in EIRENE, have been developed for atoms and shown promising results. However, they are generally not valid everywhere in the simulation domain and hybrid kinetic-fluid models are shown to provide a way to combine accuracy and efficiency.