CHAOS: A multi-GPU PIC-DSMC solver for modeling gas and plasma flows

PIC;DSMC;Forest of Octree;GPU;CUDA;MPI;Morton encoding;linearization;permeability;porous media;plasma plume;neutralization;electron kinetics;charge-exchange collisions;ion backflow

Jambunathan, Revathi

关键词: PIC;    DSMC;    Forest of Octree;    GPU;    CUDA;    MPI;    Morton encoding;    linearization;    permeability;    porous media;    plasma plume;    neutralization;    electron kinetics;    charge-exchange collisions;    ion backflow;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/104740/JAMBUNATHAN-DISSERTATION-2019.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Numerical modeling of gas and plasma-surface interactions is critical to understanding the complex kinetic processes that dominate the extreme environments of planetary entry and in-space propulsion. However, simulations of these systems that evolve over multiple length- and time-scales is computationally expensive.Until recently, approximations were used to keep computational costs tenable, which in turn, increased the uncertainty in predictions and offered limited insights into the micro-scale flow properties and electron kinetics that dominate the macroscale processes.The need to perform high-fidelity physics-based gas and plasma simulations has led to the development of a three-dimensional, multi-GPU, Particle-in-cell (PIC)-direct simulation Monte Carlo (DSMC) solver called Cuda-based Hybrid Approach for Octree Simulations (CHAOS) that is presented in this work.This computational tool has been applied to candidate PICA-like TPS materials that consist of an irregular porous network of fibers to allow high-temperature boundary layer gases as well as pyrolysis by-products to penetrate in and flow out of the material. Quantifying bulk transport properties of these materials is essential for accurate prediction of the macroscopic ablation rate. The second application that CHAOS is being used with is the modeling of ion thruster plumes that consist of fast beam ions and slow neutrals that undergo charge-exchange (CEX) reactions to produce slow ions and fast neutrals. These slow CEX ions are strongly influenced by the electric field induced between the ion plume and the thruster surface, resulting in a backflow of ions towards the critical solar panel and thruster surfaces. Three backflow quantities, namely, ion flux, incidence angle, and incidence energy affect the macroscopic sputtering rate of the solar panel surfaces over extended operational times and are predicted from the PIC-DSMC simulations.