Sampling method in probabilistic S-version finite element analysis for initial flaw size

M. R M Akramin, Ahmad Kamal Ariffin Mohd Ihsan, Masanori Kikuchi, Shahrum Abdullah

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

The aim of this paper is to demonstrate the accuracy of the sampling method in probabilistic fatigue surface crack growth analysis. A finite thickness plate with surface crack subjected to a fatigue load is considered for fracture analysis using the newly developed Probabilistic S-version Finite Element Model (ProbS-FEM). Two sampling methods, Monte Carlo and Latin Hypercube are employed in the ProbS-FEM. Two different sampling methods are used to generate the random variables in the analysis. The results from the strategies are compared to highlight the advantages of each sampling strategy. The distribution of the initial surface crack length and depth are extracted from the fatigue tests. The accuracy and consistency between the two sampling methods are evaluated. The Latin Hypercube sampling (LHS) shows advantages compared to the Monte Carlo. Then, the developed ProbS-FEM is demonstrated on a four-point bending model. The developed ProbS-FEM shows that the code is capable of modelling the uncertainty in the fatigue surface crack analysis with better accuracy and consistency of sampling process.

Original languageEnglish
Pages (from-to)357-365
Number of pages9
JournalJournal of the Brazilian Society of Mechanical Sciences and Engineering
Volume39
Issue number1
DOIs
Publication statusPublished - 1 Jan 2017

Fingerprint

Sampling
Finite element method
Defects
Fatigue of materials
Cracks
Random variables
Crack propagation
Monte Carlo methods

Keywords

  • Crack growth
  • Finite element model
  • Probabilistic
  • Stress intensity factor
  • Surface crack

ASJC Scopus subject areas

  • Mechanical Engineering

Cite this

@article{ebb695a44d544ba8b1849f421a45490c,
title = "Sampling method in probabilistic S-version finite element analysis for initial flaw size",
abstract = "The aim of this paper is to demonstrate the accuracy of the sampling method in probabilistic fatigue surface crack growth analysis. A finite thickness plate with surface crack subjected to a fatigue load is considered for fracture analysis using the newly developed Probabilistic S-version Finite Element Model (ProbS-FEM). Two sampling methods, Monte Carlo and Latin Hypercube are employed in the ProbS-FEM. Two different sampling methods are used to generate the random variables in the analysis. The results from the strategies are compared to highlight the advantages of each sampling strategy. The distribution of the initial surface crack length and depth are extracted from the fatigue tests. The accuracy and consistency between the two sampling methods are evaluated. The Latin Hypercube sampling (LHS) shows advantages compared to the Monte Carlo. Then, the developed ProbS-FEM is demonstrated on a four-point bending model. The developed ProbS-FEM shows that the code is capable of modelling the uncertainty in the fatigue surface crack analysis with better accuracy and consistency of sampling process.",
keywords = "Crack growth, Finite element model, Probabilistic, Stress intensity factor, Surface crack",
author = "Akramin, {M. R M} and {Mohd Ihsan}, {Ahmad Kamal Ariffin} and Masanori Kikuchi and Shahrum Abdullah",
year = "2017",
month = "1",
day = "1",
doi = "10.1007/s40430-016-0549-z",
language = "English",
volume = "39",
pages = "357--365",
journal = "Journal of the Brazilian Society of Mechanical Sciences and Engineering",
issn = "1678-5878",
publisher = "Brazilian Society of Mechanical Sciences and Engineering",
number = "1",

}

TY - JOUR

T1 - Sampling method in probabilistic S-version finite element analysis for initial flaw size

AU - Akramin, M. R M

AU - Mohd Ihsan, Ahmad Kamal Ariffin

AU - Kikuchi, Masanori

AU - Abdullah, Shahrum

PY - 2017/1/1

Y1 - 2017/1/1

N2 - The aim of this paper is to demonstrate the accuracy of the sampling method in probabilistic fatigue surface crack growth analysis. A finite thickness plate with surface crack subjected to a fatigue load is considered for fracture analysis using the newly developed Probabilistic S-version Finite Element Model (ProbS-FEM). Two sampling methods, Monte Carlo and Latin Hypercube are employed in the ProbS-FEM. Two different sampling methods are used to generate the random variables in the analysis. The results from the strategies are compared to highlight the advantages of each sampling strategy. The distribution of the initial surface crack length and depth are extracted from the fatigue tests. The accuracy and consistency between the two sampling methods are evaluated. The Latin Hypercube sampling (LHS) shows advantages compared to the Monte Carlo. Then, the developed ProbS-FEM is demonstrated on a four-point bending model. The developed ProbS-FEM shows that the code is capable of modelling the uncertainty in the fatigue surface crack analysis with better accuracy and consistency of sampling process.

AB - The aim of this paper is to demonstrate the accuracy of the sampling method in probabilistic fatigue surface crack growth analysis. A finite thickness plate with surface crack subjected to a fatigue load is considered for fracture analysis using the newly developed Probabilistic S-version Finite Element Model (ProbS-FEM). Two sampling methods, Monte Carlo and Latin Hypercube are employed in the ProbS-FEM. Two different sampling methods are used to generate the random variables in the analysis. The results from the strategies are compared to highlight the advantages of each sampling strategy. The distribution of the initial surface crack length and depth are extracted from the fatigue tests. The accuracy and consistency between the two sampling methods are evaluated. The Latin Hypercube sampling (LHS) shows advantages compared to the Monte Carlo. Then, the developed ProbS-FEM is demonstrated on a four-point bending model. The developed ProbS-FEM shows that the code is capable of modelling the uncertainty in the fatigue surface crack analysis with better accuracy and consistency of sampling process.

KW - Crack growth

KW - Finite element model

KW - Probabilistic

KW - Stress intensity factor

KW - Surface crack

UR - http://www.scopus.com/inward/record.url?scp=85003029580&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85003029580&partnerID=8YFLogxK

U2 - 10.1007/s40430-016-0549-z

DO - 10.1007/s40430-016-0549-z

M3 - Article

AN - SCOPUS:85003029580

VL - 39

SP - 357

EP - 365

JO - Journal of the Brazilian Society of Mechanical Sciences and Engineering

JF - Journal of the Brazilian Society of Mechanical Sciences and Engineering

SN - 1678-5878

IS - 1

ER -