Fully kinetic PIC-DSMC-MCC approach for modeling of electric thruster plasma plumes and their interaction with solar panel surfaces, using multi-GPU CHAOS solver

Deborah Levin, University of Illinois at Urbana-Champaign

Usage Details

The main objective of our Blue Waters proposal is to study the influence of electrons that are ejected from an electric-propulsion (EP) neutralizer device on the backflow contamination plasma environment using state-of-the-art high performance petascale computations. Using the previous Blue Waters' allocations, we extended our in-house multi-GPU CHAOS solver to include an electric-field Particle-In-Cell (PIC) module, capable of solving the Poisson's equation as well as extended the DSMC module to include the crucial CEX reactions between neutrals and ions. From our previous work in the group, we have shown that the use of octrees is computationally efficient for expansion flows with multiple length scales. Additionally, we developed the implementation for multi-timestep and species weighting factors required for dense slow neutrals and fast-moving ions. Our ultimate goal is to develop a coupled DSMC-PIC approach that will enable us to accurately model the plume dynamics as well as the plasma sheath at the surface of the solar panels, i.e., develop a physics-based computational tool to model and predict spacecraft material erosion and charge. As is well known, kinetic modeling of electrons is challenging due to their high speed, low mass, and small Debye lengths. Except for the work of Wang et al, where only collisionless mesothermal plasma were modeled, their behavior is mostly modeled using a fluid approach. Using our last year's Blue Waters project, we demonstrated that our multi-GPU hybrid parallelization applied to tree-based computational strategies could be generalized to model collisional plasma. We extended our Morton Z-ordered linear octree algorithms for flows through porous media to include weighting schemes required for heavy particle and electron species with different time step values. These computational savings have made it is possible to model the behavior of electrons kinetically, thereby achieving the highest fidelity physics of space plasmas. Our goal this year is to build on this exciting development by including electron-neutral ionization and electron-ion recombination reactions in a Monte Carlo Collision formalism to obtain a coupled PIC-DSMC-MCC approach.