Machine learning-enhanced optimization of rotor blades for rotary-wing Mars UAVs through coupled CFD simulation

Zhou Zhang; Pengyue Zhao; Zhixiang Zhang; Qinghan Sun; Jiaxuan Zhang; Zongquan Deng

doi:10.1016/j.ast.2024.109858

Machine learning-enhanced optimization of rotor blades for rotary-wing Mars UAVs through coupled CFD simulation

Zhou Zhang
, Pengyue Zhao^*
, Zhixiang Zhang
, Qinghan Sun
, Jiaxuan Zhang
, Zongquan Deng

^*Corresponding author for this work

Harbin Institute of Technology
Ministry of Industry Information Technology
Harbin Institute of Technology

Research output: Contribution to journal › Article › peer-review

8 Scopus citations

Abstract

This study introduces a comprehensive method for optimizing rotary-wing rotor blades for Mars UAV by combining 2D airfoil CFD simulations with machine learning methods. The algorithms such as ANN, Ada-Boost, SVM-L, and SVM-G were applied to analyze key input and output parameters. The SVM-G algorithm showed high accuracy for C_l/C_d and both SVM-G and Ada-Boost for C_l^1.5/C_d. The optimized blade design aimed to maximize C_l^1.5/C_d spanwise and was validated through Martian atmospheric ground simulations. The results showed that with a power input of 100W, the rotor system generated a thrust of 5.02N, achieving a figure of merit of 0.7437 and a power loading of 0.0487N/W, indicating the rotor system's efficiency and potential for Martian exploration missions.

Original language	English
Article number	109858
Journal	Aerospace Science and Technology
Volume	158
DOIs	https://doi.org/10.1016/j.ast.2024.109858
State	Published - Mar 2025

Keywords

Aerodynamic performance
CFD simulation
Machine learning
Mars UAV
Rotor blade

Access to Document

10.1016/j.ast.2024.109858

Cite this

@article{714fc81ce4974d6ba316884143ae03fd,

title = "Machine learning-enhanced optimization of rotor blades for rotary-wing Mars UAVs through coupled CFD simulation",

abstract = "This study introduces a comprehensive method for optimizing rotary-wing rotor blades for Mars UAV by combining 2D airfoil CFD simulations with machine learning methods. The algorithms such as ANN, Ada-Boost, SVM-L, and SVM-G were applied to analyze key input and output parameters. The SVM-G algorithm showed high accuracy for Cl/Cd and both SVM-G and Ada-Boost for Cl1.5/Cd. The optimized blade design aimed to maximize Cl1.5/Cd spanwise and was validated through Martian atmospheric ground simulations. The results showed that with a power input of 100W, the rotor system generated a thrust of 5.02N, achieving a figure of merit of 0.7437 and a power loading of 0.0487N/W, indicating the rotor system's efficiency and potential for Martian exploration missions.",

keywords = "Aerodynamic performance, CFD simulation, Machine learning, Mars UAV, Rotor blade",

author = "Zhou Zhang and Pengyue Zhao and Zhixiang Zhang and Qinghan Sun and Jiaxuan Zhang and Zongquan Deng",

note = "Publisher Copyright: {\textcopyright} 2024 Elsevier Masson SAS",

year = "2025",

month = mar,

doi = "10.1016/j.ast.2024.109858",

language = "英语",

volume = "158",

journal = "Aerospace Science and Technology",

issn = "1270-9638",

publisher = "Elsevier Masson s.r.l.",

}

TY - JOUR

T1 - Machine learning-enhanced optimization of rotor blades for rotary-wing Mars UAVs through coupled CFD simulation

AU - Zhang, Zhou

AU - Zhao, Pengyue

AU - Zhang, Zhixiang

AU - Sun, Qinghan

AU - Zhang, Jiaxuan

AU - Deng, Zongquan

PY - 2025/3

Y1 - 2025/3

N2 - This study introduces a comprehensive method for optimizing rotary-wing rotor blades for Mars UAV by combining 2D airfoil CFD simulations with machine learning methods. The algorithms such as ANN, Ada-Boost, SVM-L, and SVM-G were applied to analyze key input and output parameters. The SVM-G algorithm showed high accuracy for Cl/Cd and both SVM-G and Ada-Boost for Cl1.5/Cd. The optimized blade design aimed to maximize Cl1.5/Cd spanwise and was validated through Martian atmospheric ground simulations. The results showed that with a power input of 100W, the rotor system generated a thrust of 5.02N, achieving a figure of merit of 0.7437 and a power loading of 0.0487N/W, indicating the rotor system's efficiency and potential for Martian exploration missions.

AB - This study introduces a comprehensive method for optimizing rotary-wing rotor blades for Mars UAV by combining 2D airfoil CFD simulations with machine learning methods. The algorithms such as ANN, Ada-Boost, SVM-L, and SVM-G were applied to analyze key input and output parameters. The SVM-G algorithm showed high accuracy for Cl/Cd and both SVM-G and Ada-Boost for Cl1.5/Cd. The optimized blade design aimed to maximize Cl1.5/Cd spanwise and was validated through Martian atmospheric ground simulations. The results showed that with a power input of 100W, the rotor system generated a thrust of 5.02N, achieving a figure of merit of 0.7437 and a power loading of 0.0487N/W, indicating the rotor system's efficiency and potential for Martian exploration missions.

KW - Aerodynamic performance

KW - CFD simulation

KW - Machine learning

KW - Mars UAV

KW - Rotor blade

UR - https://www.scopus.com/pages/publications/85214334172

U2 - 10.1016/j.ast.2024.109858

DO - 10.1016/j.ast.2024.109858

M3 - 文章

AN - SCOPUS:85214334172

SN - 1270-9638

VL - 158

JO - Aerospace Science and Technology

JF - Aerospace Science and Technology

M1 - 109858

ER -

Machine learning-enhanced optimization of rotor blades for rotary-wing Mars UAVs through coupled CFD simulation

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this