For the simulation of the inductively coupled plasma in a micro-Newton Radio Frequency Ion Thruster (RIT), PlasmaPIC was developed from scratch in this thesis. It is based on the particle in cell (PIC) method and distinguishes itself with a fully three-dimensional simulation of plasma discharge including an electrodynamic part. PlasmaPIC features the investigation of arbitrary geometries, which can be imported from common computer aided design (CAD) tools.The very long simulation time of a three-dimensional simulation is dramatically reduced by incorporating a massive parallelization using the domain decomposition method. In this context, different software parallelization libraries as well as hardware architectures, CPU and GPU, are tested and compared. The CPU architecture performing the message passing interface (MPI) is the most promising concept due to the great scalability and is employed in PlasmaPIC . Depending on the mesh size, up to one thousand cores were efficiently used.Although PlasmaPIC was designed for the modeling of inductively coupled low pressure discharges, it supports the modeling of different low pressure discharges.The capabilities of PlasmaPIC are demonstrated by showing and analyzing the results of the electrodynamic plasma discharge simulation in a micro-Newton RIT and an electrostatic DC-discharge. In detail the impact of the power deposition and the neutral gas pressure is investigated for the micro-Newton RIT. For the latter, measurements for larger RITs are available and in good agreement with the simulation results.Beside the high spatial resolution in PlasmaPIC the temporal resolution of nanoseconds reveals the moving striations in a DC-discharge as well as a considerable oscillating plasma sheath in the micro-Netwon RIT during one rf-cycle, which is hard to obtain in experiments.The main advantage of PlasmaPIC is its ability to predict plasma and performance parameters for new thruster designs on a microscopic scale. By this means PlasmaPIC introduces a new way of understanding and optimizing micro-Newton RITs
Verknüpfung zu Publikationen oder weiteren Datensätzen
