A multi-level parallel solver for rarefied gas flows in porous media

Research output: Contribution to journalArticle

  • 1 Citations

Abstract

A high-performance gas kinetic solver using multi-level parallelization is developed to enable pore-scale simulations of rarefied flows in porous media. The Bhatnagar–Gross–Krook model equation is solved by the discrete velocity method with an iterative scheme. The multi-level MPI/OpenMP parallelization is implemented with the aim to efficiently utilize the computational resources to allow direct simulation of rarefied gas flows in porous media based on digital rock images for the first time. The multi-level parallel approach is analyzed in detail confirming its better performance than the commonly-used MPI processing alone for an iterative scheme. With high communication efficiency and appropriate load balancing among CPU processes, parallel efficiency of 94% is achieved for 1536 cores in the 2D simulations, and 81% for 12288 cores in the 3D simulations. While decomposition in the spatial space does not affect the simulation results, one additional benefit of this approach is that the number of subdomains can be kept minimal to avoid deterioration of the convergence rate of the iteration process. This multi-level parallel approach can be readily extended to solve other Boltzmann model equations.

LanguageEnglish
Pages14-25
Number of pages12
JournalComputer Physics Communications
Volume234
Early online date30 Aug 2018
DOIs
Publication statusPublished - 31 Jan 2019

Fingerprint

rarefied gases
gas flow
Flow of gases
Porous materials
Kinetic theory of gases
Resource allocation
Program processors
Deterioration
simulation
Rocks
Decomposition
Communication
Processing
deterioration
iteration
resources
communication
rocks
porosity
decomposition

Keywords

  • rarefied gas dynamics
  • porous media
  • multi-level parallel
  • permeability

Cite this

@article{766c94b7949f4b10a0bdd39c3dde193d,
title = "A multi-level parallel solver for rarefied gas flows in porous media",
abstract = "A high-performance gas kinetic solver using multi-level parallelization is developed to enable pore-scale simulations of rarefied flows in porous media. The Bhatnagar–Gross–Krook model equation is solved by the discrete velocity method with an iterative scheme. The multi-level MPI/OpenMP parallelization is implemented with the aim to efficiently utilize the computational resources to allow direct simulation of rarefied gas flows in porous media based on digital rock images for the first time. The multi-level parallel approach is analyzed in detail confirming its better performance than the commonly-used MPI processing alone for an iterative scheme. With high communication efficiency and appropriate load balancing among CPU processes, parallel efficiency of 94{\%} is achieved for 1536 cores in the 2D simulations, and 81{\%} for 12288 cores in the 3D simulations. While decomposition in the spatial space does not affect the simulation results, one additional benefit of this approach is that the number of subdomains can be kept minimal to avoid deterioration of the convergence rate of the iteration process. This multi-level parallel approach can be readily extended to solve other Boltzmann model equations.",
keywords = "rarefied gas dynamics , porous media, multi-level parallel, permeability",
author = "Ho, {Minh Tuan} and Lianhua Zhu and Lei Wu and Peng Wang and Zhaoli Guo and Zhi-Hui Li and Yonghao Zhang",
year = "2019",
month = "1",
day = "31",
doi = "10.1016/j.cpc.2018.08.009",
language = "English",
volume = "234",
pages = "14--25",
journal = "Computer Physics Communications",
issn = "0010-4655",

}

A multi-level parallel solver for rarefied gas flows in porous media. / Ho, Minh Tuan; Zhu, Lianhua; Wu, Lei; Wang, Peng; Guo, Zhaoli; Li, Zhi-Hui; Zhang, Yonghao.

In: Computer Physics Communications, Vol. 234, 31.01.2019, p. 14-25.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A multi-level parallel solver for rarefied gas flows in porous media

AU - Ho, Minh Tuan

AU - Zhu, Lianhua

AU - Wu, Lei

AU - Wang, Peng

AU - Guo, Zhaoli

AU - Li, Zhi-Hui

AU - Zhang, Yonghao

PY - 2019/1/31

Y1 - 2019/1/31

N2 - A high-performance gas kinetic solver using multi-level parallelization is developed to enable pore-scale simulations of rarefied flows in porous media. The Bhatnagar–Gross–Krook model equation is solved by the discrete velocity method with an iterative scheme. The multi-level MPI/OpenMP parallelization is implemented with the aim to efficiently utilize the computational resources to allow direct simulation of rarefied gas flows in porous media based on digital rock images for the first time. The multi-level parallel approach is analyzed in detail confirming its better performance than the commonly-used MPI processing alone for an iterative scheme. With high communication efficiency and appropriate load balancing among CPU processes, parallel efficiency of 94% is achieved for 1536 cores in the 2D simulations, and 81% for 12288 cores in the 3D simulations. While decomposition in the spatial space does not affect the simulation results, one additional benefit of this approach is that the number of subdomains can be kept minimal to avoid deterioration of the convergence rate of the iteration process. This multi-level parallel approach can be readily extended to solve other Boltzmann model equations.

AB - A high-performance gas kinetic solver using multi-level parallelization is developed to enable pore-scale simulations of rarefied flows in porous media. The Bhatnagar–Gross–Krook model equation is solved by the discrete velocity method with an iterative scheme. The multi-level MPI/OpenMP parallelization is implemented with the aim to efficiently utilize the computational resources to allow direct simulation of rarefied gas flows in porous media based on digital rock images for the first time. The multi-level parallel approach is analyzed in detail confirming its better performance than the commonly-used MPI processing alone for an iterative scheme. With high communication efficiency and appropriate load balancing among CPU processes, parallel efficiency of 94% is achieved for 1536 cores in the 2D simulations, and 81% for 12288 cores in the 3D simulations. While decomposition in the spatial space does not affect the simulation results, one additional benefit of this approach is that the number of subdomains can be kept minimal to avoid deterioration of the convergence rate of the iteration process. This multi-level parallel approach can be readily extended to solve other Boltzmann model equations.

KW - rarefied gas dynamics

KW - porous media

KW - multi-level parallel

KW - permeability

UR - https://www.sciencedirect.com/journal/computer-physics-communications

U2 - 10.1016/j.cpc.2018.08.009

DO - 10.1016/j.cpc.2018.08.009

M3 - Article

VL - 234

SP - 14

EP - 25

JO - Computer Physics Communications

T2 - Computer Physics Communications

JF - Computer Physics Communications

SN - 0010-4655

ER -