Intelligent data-driven aerodynamic analysis and optimization of morphing configurations

被引：17

作者：

Magalhaes Junior, Jose M. ^{[1
]}

Halila, Gustavo L. O. ^{[2
]}

Kim, Yoobin ^{[1
]}

Khamvilai, Thanakorn ^{[1
]}

Vamvoudakis, Kyriakos G. ^{[1
]}

机构：

[1] Georgia Inst Technol, Atlanta, GA 30332 USA

[2] Inst Aeronaut & Espaco, BR-12228904 Sao Jose Dos Campos, SP, Brazil

来源：

AEROSPACE SCIENCE AND TECHNOLOGY | 2022年 / 121卷

基金：

美国国家科学基金会;

关键词：

Aerodynamic shape optimization; Data-driven aerodynamic design; Neural networks; Model predictive control; Aerodynamic design; DESIGN;

D O I：

10.1016/j.ast.2022.107388

中图分类号：

V [航空、航天];

学科分类号：

08 ; 0825 ;

摘要：

In this paper, we develop an online, data-based framework for the aircraft airfoil to be able to optimally morph vertically. The proposed framework combines data-driven analysis, optimization, and control theoretic tools to optimally morph airfoils while guaranteeing efficiency and safety. It incorporates a surrogate model that is based on a deep neural network that is used to predict the aerodynamic coefficients while a meta-heuristic optimization algorithm is employed to find shapes with reduced value of drag coefficient that fulfill the geometric and lift constraints. Finally, a data-driven shape controller is used to morph the airfoil while following smooth trajectories and small aerodynamic coefficient variations. Experimental numerical results show the efficacy of the proposed framework for different flight conditions.(c) 2022 Elsevier Masson SAS. All rights reserved.

引用

页数：15

共 43 条

[1]

Barr A. H., 1984, Computers & Graphics, V18, P21

[2] Data-Driven Model Predictive Control With Stability and Robustness Guarantees [J].