Large-scale parallel discrete element simulations of granular flow

Purpose - The purpose of this paper is to present large-scale parallel direct numerical simulations of granular flow, using a novel, portable software program for discrete element method (DEM) simulations.Design methodology approach - Since particle methods provide a unifying framework for both disc...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Engineering computations Ročník 26; číslo 6; s. 688 - 697
Hlavní autoři: Walther, Jens H., Sbalzarini, Ivo F.
Médium: Journal Article
Jazyk:angličtina
Vydáno: Bradford Emerald Group Publishing Limited 21.08.2009
Témata:
Boundary conditions
Communication
Decomposition
Discrete element method
Engineering
Libraries
Mathematical models
Measurement
Simulation
Software
Studies
Viscoelasticity
Meshes
Particle size measurement
Simulation
Computer software
Programming and algorithm theory
ISSN:0264-4401, 1758-7077
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Purpose - The purpose of this paper is to present large-scale parallel direct numerical simulations of granular flow, using a novel, portable software program for discrete element method (DEM) simulations.Design methodology approach - Since particle methods provide a unifying framework for both discrete and continuous systems, the program is based on the parallel particle mesh (PPM) library, which has already been demonstrated to provide transparent parallelization and state-of-the-art parallel efficiency using particle methods for continuous systems.Findings - By adapting PPM to discrete systems, results are reported from three-dimensional simulations of a sand avalanche down an inclined plane.Originality value - The paper demonstrates the parallel performance and scalability of the new simulation program using up to 122 million particles on 192 processors, employing adaptive domain decomposition and load balancing techniques.
Bibliografie:SourceType-Scholarly Journals-1
ObjectType-Feature-1
content type line 14
ISSN:0264-4401
1758-7077
DOI:10.1108/02644400910975478