Weak Scaling of the Parallel Immersed-Finite-Element Particle-In-Cell (PIFE-PIC) Framework with Lunar Plasma Charging Simulations
Weak scaling performance of a recently developed fully kinetic, 3-D parallel immersed-finite-element particle-in-cell framework, namely PIFE-PIC, was investigated. A nominal 1-D plasma charging problem, the lunar photoelectron sheath at a low Sun elevation angle, was chosen to validate PIFE-PIC against recently derived semi-analytic solutions of a 1-D photoelectron sheath. The weak scaling performance test shows that the overall efficiency of PIFE-PIC is insensitive to the number of macroparticles and, counterintuitively, more domain decomposition iterations in the field-solve of PIC may lead to faster computing due to better convergence of field solutions at early stages of PIC iteration. The PIFE-PIC framework was then applied to simulate plasma charging of a wavy lunar surface with 324,000 cells and 150 million macroparticles demonstrating the capability of PIFE-PIC in resolving local-scale plasma environment near the surface of the Moon.
D. Lund et al., "Weak Scaling of the Parallel Immersed-Finite-Element Particle-In-Cell (PIFE-PIC) Framework with Lunar Plasma Charging Simulations," Computational Particle Mechanics, Springer, Mar 2022.
The definitive version is available at https://doi.org/10.1007/s40571-022-00470-0
Mathematics and Statistics
Mechanical and Aerospace Engineering
Center for High Performance Computing Research
Keywords and Phrases
Immersed finite element; Particle in cell; Plasma charging; Weak scaling
International Standard Serial Number (ISSN)
Article - Journal
© 2022 Springer, All rights reserved.
07 Mar 2022
This work was partially supported by a NASA Space Technology Graduate Research Opportunity (D.L.), NASA-Missouri Space Grant Consortium through NASA-EPSCoR-Missouri (X.H. and D.H.) and graduate scholarships (D.L.), as well as NSF through grants DMS-2005272 and DMS-2110833 (X.Z.), DMS-2111039 (X.H. and D.H.), and CBET-2132655 (D.H.). The simulations presented here were performed with computing resources provided by the Center for High Performance Computing Research at Missouri University of Science and Technology through an NSF Grant OAC-1919789.