A posteriori performance-based comparison of three new path-following constraints for damage analysis of quasi-brittle materials

A. Fayezioghani, B. Vandoren*, L. J. Sluys

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

5 Citations (Scopus)

Abstract

Using a path-following algorithm to analyze a quasi-static nonlinear structural problem involves selecting an appropriate constraint function. This function should improve the desired performance targets of the path-following algorithm such as robustness, speed, accuracy, and smoothness. In order to be able to draw a fair objective selection of a constraint function, it is necessary to collect adequate constraint equations as well as to define the performance of nonlinear methods. In this paper, three new path-following constraints applicable for damage analysis of quasi-brittle materials are proposed. Additionally, performance criteria and their numerical measures for a posteriori assessment of robustness, smoothness, accuracy, and speed of solving nonlinear problems by a path-following method are proposed. Based on the proposed criteria, the performance of the three new constraints and two existing ones is compared for two example problems. As a result, the performance measures are shown to possess an ability to clearly explore the strengths of each constraint. They establish a firm basis for the assessment of not only path-following methods but also other methods for solving nonlinear structural problems.

Original languageEnglish
Pages (from-to)746-768
Number of pages23
JournalComputer Methods in Applied Mechanics and Engineering
Volume346
DOIs
Publication statusPublished - 2019

Keywords

  • Path-following method
  • Performance criterion
  • Quasi-brittle materials
  • Snap-back

Fingerprint

Dive into the research topics of 'A posteriori performance-based comparison of three new path-following constraints for damage analysis of quasi-brittle materials'. Together they form a unique fingerprint.

Cite this