A 10-moment fluid numerical solver of plasma with sheaths in a hall effect thruster

Electric propulsion can reach higher exhaust velocities compared to chemical systems and thus result in lower propellant mass requirements. Among the different electric propulsion systems, Hall effect thrusters are used for spatial propulsion since the 1970s. However inside a Hall thruster, complex physical phenomena such as erosion or electron anomalous transport which may lower thruster efficiency and lifetime, are not yet fully understood. Thanks to high performance computing, numerical simulations are now considered for understanding the plasma behavior. With the renewed interest for such electric propulsion to supply small satellites, numerical solvers able to predict accurately the real thruster efficiency have become crucial for industry. This paper presents the approach used and first validation tests of such a solver. The AVIP code solves plasma equations in complex industrial geometries using an unstructured parallel-efficient 3D fluid methodology. AVIP also includes a Particle-In-Cell (PIC) solver used as a reference for validation. While full 3D PIC simulations of a Hall thruster still require unaffor-dable computational time, fluid models provide in a reasonable time 3D results on the plasma behavior inside the discharge channel. In this category, standard drift-diffusion models [1–3] are fast and robust but at the cost of strong hypotheses and simplifications. In particular such models do not describe explicitly the sheath formation in the vicinity of walls and often use analytical models instead. They are limited to simple configurations and only provide a first insight into plasma complex phenomena. The present approach includes a more detailed two-fluid plasma model without drift-diffusion approximation. After the description of the formulation and main features of the solver, the paper focuses on wall boundary conditions which are crucial for the formation of sheaths. It is demonstrated in particular that a vacuum boundary condition is not adapted to fit PIC results. A boundary condition based on wall thermal fluxes is more realistic. The mesh resolution is also found to be critical. The simulation methodology is finally applied to a 2D simulation of a typical Hall effect thruster in order to observe the plasma properties inside the discharge chamber.