Trust region based moving asymptotes method: A stabilized optimizer for stress-constrained topology optimization

Xueyan Hu, Erik Lund*, Zonghao Li, Weiqiu Chen*

*Corresponding author for this work

Research output: Contribution to journalJournal articleResearchpeer-review

3 Citations (Scopus)

Abstract

Inconsistent topologies and oscillations during iterations are commonly observed when solving stress-constrained problems, indicating inadequate stability in topology optimization. The trust region based moving asymptotes (TRMA) method is proposed in this paper. To comprehensively investigate the stability of the TRMA method, the Trinity Evaluation Framework is proposed in this paper. It focuses on three types of dependencies in topology optimization, i.e. mesh dependence, parameter dependence, and initial topology dependence. The corresponding evaluation vector is proposed to measure the stability of the TRMA method. Numerical examples are given, indicating strong stability of the TRMA method in topology optimization of stress-constrained problems.

Original languageEnglish
JournalMechanics of Advanced Materials and Structures
Number of pages18
ISSN1537-6494
DOIs
Publication statusE-pub ahead of print - 30 Jul 2024

Bibliographical note

Publisher Copyright:
© 2024 Taylor & Francis Group, LLC.

Keywords

  • augmented Lagrange function
  • Method of Moving Asymptotes
  • stress-constrained problems
  • topology optimization
  • trust region

Fingerprint

Dive into the research topics of 'Trust region based moving asymptotes method: A stabilized optimizer for stress-constrained topology optimization'. Together they form a unique fingerprint.

Cite this