Multidisciplinary Robust Design Optimization Incorporating Extreme Scenario in Sparse Samples

Wei Li, Yuzhen Niu, Haihong Huang*, Akhil Garg, Liang Gao

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

Robust design optimization (RDO) is a potent methodology that ensures stable performance in designed products during their operational phase. However, there remains a scarcity of robust design optimization methods that account for the intricacies of multidisciplinary coupling. In this article, we propose a multidisciplinary robust design optimization (MRDO) framework for physical systems under sparse samples containing the extreme scenario. The collaboration model is used to select samples that comply with multidisciplinary feasibility, avoiding time-consuming multidisciplinary decoupling analyses. To assess the robustness of sparse samples containing the extreme scenario, linear moment estimation is employed as the evaluation metric. The comparative analysis of MRDO results is conducted across various sample sizes, with and without the presence of the extreme scenario. The effectiveness and reliability of the proposed method are demonstrated through a mathematical case, a conceptual aircraft sizing design, and an energy efficiency optimization of a hobbing machine tool.

Original languageEnglish
Article number091701
JournalJournal of Mechanical Design
Volume146
Issue number9
DOIs
Publication statusPublished - 1 Sept 2024
Externally publishedYes

Keywords

  • design automation
  • extreme scenario
  • linear moment estimation
  • multidisciplinary design
  • multidisciplinary robust design optimization
  • optimization
  • sparse samples

Fingerprint

Dive into the research topics of 'Multidisciplinary Robust Design Optimization Incorporating Extreme Scenario in Sparse Samples'. Together they form a unique fingerprint.

Cite this