Modeling Flows through Porous Media with a Kinetic Hybrid CPU-GPU Computational Tool

Deborah Levin, University of Illinois at Urbana-Champaign

Usage Details

The main objective of our Blue Waters proposal is to characterize the back flow contamination environment due to plasma created by electric-propulsion (EP) plumes, their interaction with the spacecraft environment, and neutralizer sources, using state-of-the-art high performance petascale computations. In terms of modeling and simulation, we will build on our earlier work where we have developed an object-oriented C++ Direct Simulation Monte Carlo (DSMC) code that uses AMR/Octree grids to capture the vast length scales inherent in supersonic expansions to vacuum for neutral-neutral and neutral-ion collisions. We have shown that we can accurately capture the Charge-Exchange (CEX) process but to model spacecraft material erosion and material changes due to the back flow plasma environment we need to generalize our present particle-in-cell approach to include a full solution of Poisson's equation, which can be notoriously expensive. As is well known, modeling electrons is challenging due to their high speed and almost fluid-like behavior and we will consider both a fluid and kinetic approach in the proposed work. A key aspect of the proposed computational work is to take advantage of our unique, recent advances in GPU multi-thread parallelization applied to tree-based computational strategies. Furthermore, to efficiently capture the multiple length scales as well as weighting schemes required for multiple species with different timesteps, we propose to extend our Morton Z-ordered linear octree algorithms for flows through porous media to the modeling of collisional plasmas.

Having completed the material permeability calculations using the Cuda-based Hybrid Approach for Octree Simulations (CHAOS) DSMC solver, and demonstrating high scaling efficiency, our next target problem is to model the Hall-Effect thruster plume characteristics and the neutralization process of the plasma plume due to electrons from the hollow cathode. We will also predict the number density and energy of CEX ions in the back flow region and model the erosion and material changes to the solar panel due to the impact of CEX ions.