Hybrid PIC -MCC computational modeling of Hall thrusters.

This dissertation describes the development of a 2-D axisymmetric hybrid Particle-In-Cell Monte Carlo Collision (PIC-MCC) Hall thruster code and an investigation into the effects of the numerical parameters and physical models for this code. From the outset, it is clear that some of the necessary model boundary conditions have significant effects on both the spatial and temporal dynamics of the simulation. Therefore, judicious modeling choices must be taken to minimize interference in critical thruster physical processes. A study of the electron mobility term assesses the performance of various existing computational models of electron mobility. In the process, it is demonstrated that nearly identical thrust performance can be achieved by simulations with different plasma characteristics. As a result, robust validation against more than integrated performance data is necessary to truly validate simulation results. In this regard, none of the computational mobility models shows great success in capturing the details of the mean centerline potential profile. A semi-empirical electron mobility is developed which results in the successful validation of this code using data measured for the UM/AFRL P5 Hall thruster. The semi-empirical electron mobility is then used as a reference configuration against which to refine the computational models for electron mobility. An analysis of a dataset of UM/AFRL P5 internal plasma properties uncovered evidence of a strong magnetic self-field during thruster operation. (The existence of a magnetic self-field in this thruster has been identified only once before.) This self-field provides physically motivated corrections which drive the output, of the computational model for electron mobility towards the reference configuration. The original goal of this work was to better understand and extend the physical principles contained in existing computational simulation of Hall thrusters. In the process of validating the existing code with a mobility profile derived from experimental sources, it is discovered that electrostatic codes (such as this one) can produce results in good agreement with experimental data if tuned correctly with some knowledge of experimental conditions; however, fully self-consistent computational modeling of these thrusters will require an electromagnetic solver to properly resolve the correct magnetic configuration during thruster operation.Ph.D.Aerospace engineeringApplied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/124847/2/3163850.pd