Skip to main navigation Skip to search Skip to main content

A model for large scale near-real time simulation of granular material flow

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

This paper describes ongoing development of a methodology to model large-displacement behavior of non-cohesive soils, such as sands and gravels. A set of artificial neural networks are trained to learn the underlying knowledge behind the large displacement behavior of granular media. Data from a large number of discrete element simulations is used to train the neural networks. Particularly, the NN's are used to relate the stresses at any point inside the soil to the history of stress at some surrounding points as well as soil property and configuration in vicinity of that point. A numerical framework is developed to utilize the trained NN's in simulation environments of granular material flow. The results of 2D and 3D simulations are presented. Copyright ASCE 2006.

Original languageEnglish (US)
Title of host publicationEarth and Space 2006 - Proceedings of the 10th Biennial International Conference on Engineering, Construction, and Operations in Challenging Environments
Pages30
Number of pages1
DOIs
StatePublished - 2006
EventEarth and Space 2006 - 10th Biennial International Conference on Engineering, Construction, and Operations in Challenging Environments - League City/Houston, TX, United States
Duration: Mar 5 2006 → Mar 8 2006

Publication series

NameEarth and Space 2006 - Proceedings of the 10th Biennial International Conference on Engineering, Construction, and Operations in Challenging Environments
Volume2006

Other

OtherEarth and Space 2006 - 10th Biennial International Conference on Engineering, Construction, and Operations in Challenging Environments
Country/TerritoryUnited States
CityLeague City/Houston, TX
Period3/5/06 → 3/8/06

ASJC Scopus subject areas

  • General Engineering

Fingerprint

Dive into the research topics of 'A model for large scale near-real time simulation of granular material flow'. Together they form a unique fingerprint.

Cite this