Computational speed up techniques for particle-in-cell-Monte Carlo collision simulations of an ion engine discharge chamber

摘要

Next generation ion thrusters such as NASA's Evolutionary Xenon Thruster (NEXT) are being considered for in-space propulsion applications to meet the future space mission needs of travel to asteroids and for satellite maneuvers. NASA has been actively pursuing new designs of ion thrusters to address the high power and high-thrust propulsion needs. Computational simulations of ion engine discharge chamber plasmas will help to understand the operation and performance of these high power ion thrusters. In this work, we describe a two-dimensional Particle-in-cell Monte Carlo Collision (PIC-MCC) model developed to simulate the ion engine discharge chamber plasma processes. We utilize VSimPD, a simulation package for plasma discharges based on the Vorpal computing engine. In this model, the electrons, singly charged xenon ions, doubly charged xenon ions and xenon neutrals are tracked as kinetic particles which includes the effects of both electric and magnetic fields. Also, the model solves the electric fields every time step based on the charge particle distributions. This detailed PIC-MCC model was benchmarked on NASA's NEXT ion engine discharge chamber and the simulation results are in good agreement with experimental plasma measurements. Recently we have focused on improving the numerical algorithms in VSimPD to speed up this discharge chamber model. New particle splitting and merging procedures are implemented. These procedures preserve the charge, energy, momentum and also the electron energy distribution functions (EEDF). We will discuss these numerical algorithms and compare the accuracies of these simulation results and provide speed up results. In addition, we will also provide speed up results from the parallel processing option and the convergence procedures developed with the numerical parameters considered in these simulations.