Automated discrete element method calibration using genetic and optimization algorithms

Huy Q. Do, Alejandro M. Aragón, Dingena L. Schott*

*Corresponding author for this work

Research output: Chapter in Book/Conference proceedings/Edited volumeConference contributionScientificpeer-review

17 Citations (Scopus)
54 Downloads (Pure)

Abstract

This research aims at developing a universal methodology for automated calibration of microscopic properties of modelled granular materials. The proposed calibrator can be applied for different experimental set-ups. Two optimization approaches: (1) a genetic algorithm and (2) DIRECT optimization, are used to identify discrete element method input model parameters, e.g., coefficients of sliding and rolling friction. The algorithms are used to minimize the objective function characterized by the discrepancy between the experimental macroscopic properties and the associated numerical results. Two test cases highlight the robustness, stability, and reliability of the two algorithms used for automated discrete element method calibration with different set-ups.

Original languageEnglish
Title of host publicationProceedings of the 8th International Conference on Micromechanics on Granular Media
Subtitle of host publicationPowders and Grains 2017
EditorsF. Radjai, S. Nezamabadi, S. Luding, J.Y. Delenne
Place of PublicationLes Ulis Cedex A, France
PublisherEDP Sciences
Number of pages4
DOIs
Publication statusPublished - 2017
Event8th International Conference on Micromechanics on Granular Media: Powders and Grains 2017 - Montpellier, France
Duration: 3 Jul 20177 Jul 2017

Publication series

NameEPJ Web of Conferences
PublisherEDP Sciences
Volume140
ISSN (Electronic)2100-014X

Conference

Conference8th International Conference on Micromechanics on Granular Media
Country/TerritoryFrance
CityMontpellier
Period3/07/177/07/17

Fingerprint

Dive into the research topics of 'Automated discrete element method calibration using genetic and optimization algorithms'. Together they form a unique fingerprint.

Cite this