Neutral particles play a crucial role in the plasma edge due to their strong interaction with the plasma. To capture all physics details most often a kinetic equation is solved using a Monte Carlo (MC) approach. Unfortunately, the MC noise hampers the convergence assessment of the solution of the coupled plasma-neutral equations. Moreover, the MC calculation time increases for high-collisional detached cases. In these high-collisional cases, however, the kinetic model may be well-approximated by a fluid neutral model, which can be solved deterministically. In this paper, we assess different fluid neutral models by comparing the resulting plasma sources to the sources from an MC simulation of the kinetic equation. The fluid models take into account the microscopic cross-sections and rate coefficients from the AMJUEL-HYDHEL databases, as well as the microscopic TRIM reflection model. We accomplish the latter without the introduction of any user-defined fitting parameters. We show that a pure pressure-diffusion equation provides accurate results for the particle source, but is inaccurate for the parallel momentum and ion energy source. Adding a parallel momentum equation gives maximum errors of about 9% for the momentum and 32% for the energy source. These errors are further reduced to respectively 6% and 14% by adding a separate neutral energy equation.
Self-consistent plasma-neutrals fluid modeling of edge and scrape-off layer turbulence in diverted tokamaks