<p>This study develops an automated end-to-end airfoil-optimization framework that couples parametric modeling, genetic algorithms (GA), and CFD simulation to improve the aerodynamic performance of RAE and NACA series airfoils across multiple angles of attack. Airfoil geometry is parametrized using the Class-Shape Transformation (CST) method, with ten Bernstein-polynomial control points per surface and two shape parameters (N,M), giving 22 design variables encoded in 16-bit binary strings. The GA evolves airfoil populations to maximize the lift-to-drag ratio (<InlineEquation ID="IEq1"> <EquationSource Format="TEX">$C_L/C_D$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mi>C</mi> <mi>L</mi> </msub> <mo stretchy="false">/</mo> <msub> <mi>C</mi> <mi>D</mi> </msub> </mrow> </math></EquationSource> </InlineEquation>), while a Python-driven workflow automates geometry generation, CFD meshing, simulation, and aerodynamic-coefficient extraction, reducing manual data transfer and improving procedural consistency. Optimization over eight angles of attack (<InlineEquation ID="IEq2"> <EquationSource Format="TEX">$0^{\circ }$</EquationSource> <EquationSource Format="MATHML"><math> <msup> <mn>0</mn> <mo>∘</mo> </msup> </math></EquationSource> </InlineEquation>–<InlineEquation ID="IEq3"> <EquationSource Format="TEX">$17.5^{\circ }$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <mn>17</mn> <mo>.</mo> <msup> <mn>5</mn> <mo>∘</mo> </msup> </mrow> </math></EquationSource> </InlineEquation>) shows consistent <InlineEquation ID="IEq4"> <EquationSource Format="TEX">$C_L/C_D$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mi>C</mi> <mi>L</mi> </msub> <mo stretchy="false">/</mo> <msub> <mi>C</mi> <mi>D</mi> </msub> </mrow> </math></EquationSource> </InlineEquation> gains, with different geometry trends under low, moderate, and high angles of attack. The results indicate larger and more stable benefits under low-to-moderate angles and diminishing gains under high-angle conditions. Pearson-correlation sensitivity analysis further identifies CST control regions that have stronger influence on <InlineEquation ID="IEq5"> <EquationSource Format="TEX">$C_L/C_D$</EquationSource> <EquationSource Format="MATHML"><math> <mrow> <msub> <mi>C</mi> <mi>L</mi> </msub> <mo stretchy="false">/</mo> <msub> <mi>C</mi> <mi>D</mi> </msub> </mrow> </math></EquationSource> </InlineEquation>, providing design-space information for possible parameter reduction. The results show that the proposed workflow can support multi-airfoil, multi-angle optimization and provide a basis for future extensions, such as surrogate-model-assisted or machine-learning-driven multi-objective optimization.</p>

An Airfoil Optimization Framework Based on Genetic Algorithm and Computational Fluid Dynamics

期刊论文
出版模式：

开放获取
发表日期：

2026年6月22日

查看全文

Marco Orban,
Po-Chao Chuang,
Zun-Hong Yu,
Wei-Hsiang Wang

摘要

This study develops an automated end-to-end airfoil-optimization framework that couples parametric modeling, genetic algorithms (GA), and CFD simulation to improve the aerodynamic performance of RAE and NACA series airfoils across multiple angles of attack. Airfoil geometry is parametrized using the Class-Shape Transformation (CST) method, with ten Bernstein-polynomial control points per surface and two shape parameters (N,M), giving 22 design variables encoded in 16-bit binary strings. The GA evolves airfoil populations to maximize the lift-to-drag ratio ( $C_L/C_D$ $C_{L} / C_{D}$ ), while a Python-driven workflow automates geometry generation, CFD meshing, simulation, and aerodynamic-coefficient extraction, reducing manual data transfer and improving procedural consistency. Optimization over eight angles of attack ( $0^{\circ }$ $0^{\circ}$ – $17.5^{\circ }$ $17 . 5^{\circ}$ ) shows consistent $C_L/C_D$ $C_{L} / C_{D}$ gains, with different geometry trends under low, moderate, and high angles of attack. The results indicate larger and more stable benefits under low-to-moderate angles and diminishing gains under high-angle conditions. Pearson-correlation sensitivity analysis further identifies CST control regions that have stronger influence on $C_L/C_D$ $C_{L} / C_{D}$ , providing design-space information for possible parameter reduction. The results show that the proposed workflow can support multi-airfoil, multi-angle optimization and provide a basis for future extensions, such as surrogate-model-assisted or machine-learning-driven multi-objective optimization.