Skip to main navigation Skip to search Skip to main content

Adaptive model refinement with batch bayesian sampling for optimization of bio-inspired flow tailoring

  • SUNY Buffalo

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

6 Scopus citations

Abstract

This paper presents an advancement of a model-independent surrogate based optimization method with adaptive sampling, known as Adaptive Model Refinement. The primary advancement relates to the capability to identify optimum location to add new samples. To this end, a new variable fidelity optimization method is proposed, where the timing of refinement and the number of infill points are determined based on the existing work in adaptive model refinement, and the location of the infill points are uniquely decided by integrating a GP-based criteria (“q-EI”) adopted from Bayesian optimization. Preliminary optimization runs showcase the achievement of up to 9% of drag reduction.

Original languageEnglish
Title of host publicationAIAA Aviation 2019 Forum
PublisherAmerican Institute of Aeronautics and Astronautics Inc, AIAA
Pages1-20
Number of pages20
ISBN (Print)9781624105890
DOIs
StatePublished - 2019
EventAIAA Aviation 2019 Forum - Dallas, United States
Duration: Jun 17 2019Jun 21 2019

Publication series

NameAIAA Aviation 2019 Forum

Conference

ConferenceAIAA Aviation 2019 Forum
Country/TerritoryUnited States
CityDallas
Period06/17/1906/21/19

Keywords

  • Bayesian Optimization
  • Bio-inspired Fluid Dynamics
  • Passive Flow Control
  • Riblets
  • Surrogate-Based Optimization
  • Variable-Fidelity Optimization

Fingerprint

Dive into the research topics of 'Adaptive model refinement with batch bayesian sampling for optimization of bio-inspired flow tailoring'. Together they form a unique fingerprint.

Cite this