Quasi Monte Carlo simulations for stochastic failure analysis in composites

Emil J. Pitz, Kishore V. Pochiraju

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Scopus citations

Abstract

Due to the heterogeneity in their microstructure composite materials exhibit considerable spatial variations in properties, which strongly impact their mechanical performance, cause variability in stress and strain fields, and influence evolution of damage in the material. Recent research has focused on considering material behavior uncertainties and quantifying the impact on the structural response. In this paper eXtended Finite Element Method (XFEM) and Continuum Damage Mechanics (CDM) models are employed for damage modeling in composite materials, where variations of material strength are taken into account using spatial random fields to model the material strength parameters. Using samples of the property fields, Monte Carlo (MC) and Quasi Monte Carlo (QMC) techniques are utilized to quantify the influence of strength variability on the structural response. As MC technique can become prohibitively expensive, research using MC technique for stochastic model analysis in composite structures is sparse. QMC technique has been shown to provide faster convergence compared to MC method, hence reducing the computational expense, though application to composite analysis has not been found. The current methodology allows examination of the stochastic progressive damage behavior of composite materials. The influence of input variability on the mean crack length in open hole angle-ply tensile tests is examined using MC and QMC techniques, and the convergence behavior of the methods is compared. The results show faster convergence when using QMC.

Original languageEnglish
Title of host publicationProceedings of the American Society for Composites - 34th Technical Conference, ASC 2019
EditorsKyriaki Kalaitzidou
ISBN (Electronic)9781605956022
DOIs
StatePublished - 2019
Event34th Technical Conference of the American Society for Composites, ASC 2019 - Atlanta, United States
Duration: 23 Sep 201925 Sep 2019

Publication series

NameProceedings of the American Society for Composites - 34th Technical Conference, ASC 2019

Conference

Conference34th Technical Conference of the American Society for Composites, ASC 2019
Country/TerritoryUnited States
CityAtlanta
Period23/09/1925/09/19

Fingerprint

Dive into the research topics of 'Quasi Monte Carlo simulations for stochastic failure analysis in composites'. Together they form a unique fingerprint.

Cite this