Simulation and comparison between multi-fluid and kinetic models of impurity transport

Abstract

Impurity transport is a highly significant research topic in international fusion plasma simulations, which are mainly simulated by numerical codes at present. Most of the numerical simulation codes for impurity transport adopt multi-fluid or kinetic model to treat impurity particles. Therefore, it is necessary to select a suitable transport model for the simulation process. For impurity particles, if the mean free path of particles λ is much smaller than the gradient scale length of particles λ g, it is sufficient to treat the particles by the multi-fluid model. However, under some conditions, λ will be much larger than λ g. The applicability of the fluid model is limited when λ is larger than or equal to λ g. A comparison with the simulations on impurity transport treated with multi-fluid and kinetic models is necessary, respectively. In this study, the simulation results of carbon (C) impurity transport in the EAST scrape-off layer with the 2D edge plasma fluid code SOLPS-ITER and the 2D Monte Carlo impurity transport code DIVIMP are compared. The comparison between the distributions of carbon impurities ( C 0 ∼ C + 6) in the different ionization states and the CIII emissivity predicted by SOLPS-ITER and DIVIMP shows that the density distributions of carbon atoms C 0 predicted by the SOLPS-ITER and DIVIMP codes are similar. However, for carbon ions in different ionization states, the variations between the density distributions simulated from the SOLPS-ITER and DIVIMP codes can become larger with the increase in ionization states. DIVIMP performs slightly better than SOLPS-ITER in reproducing the shape of the CIII profile when drifts are switched off in SOLPS-ITER, but the difference is extremely small in terms of the uncertainties involved in these calculations.