Validation of the DRACO Particle-in-Cell Code using Busek 200W Hall Thruster Experimental Data; Master's thesis

摘要

The thesis discusses developments to the electric propulsion plume code DRACO as well as validation and sensitivity analysis of the code. DRACO is a PIC code that models particles kinematically while using finite differences schemes to solve the electric potential and field. The DRACO ode has been recently modified. A particle source has been added to use the Hall Thruster device code HPHall as input to model Hall Thrusters. The code is capable of using a non-uniform mesh that uses any combination of uniform, linear and exponential stretching schemes in any of three directions. A stretched mesh can be used to refine simulation results in certain areas or improve performance by reducing the number of cells in a mesh. Finally, DRACO has the capability of using a DSMC collision scheme as well as performing recombination collisions. A sensitivity analysis of the upgraded DRACO code was performed to test experimental data gathered using a Busek 200 W Hall Thruster. A simulation was created that attempts to numerically recreate the experiment; the validation is performed by comparing the plasma potential, polytropic temperature, ion number density of the thruster plume as well as Faraday and ExB probe results. The study compares the HPHall source with older source models and also compares variations of the HPHall source. The field solver and collision model used are compared to determine how to achieve the best results. Finally, both uniform and non-uniform meshes are tested to determine if a non-uniform mesh can be properly implemented to improve simulation results and performance. Results from the validation and sensitivity study show the DRACO code can be used to recreate a vacuum chamber simulation using a Hall Thruster. The best results occur when the HPHall source is used with a MCC collision scheme using a projected background neutral density and CEX collision tracking. A stretched mesh was tested and proved results that are as accurate as a uniform mesh.