pp. 3263-3283
S&M4117 Research paper of Special Issue https://doi.org/10.18494/SAM5598 Published: July 31, 2025 Application of AI Deep Learning in Aircraft Shape Analysis and Validation for High-Performance Numerical Wind Tunnel [PDF] Ching-Huei Huang, Kun-Lin Tsai, and Shih-Ting Tseng (Received February 27, 2025; Accepted May 8, 2025) Keywords: aircraft shape design, AI, backpropagation neural network, deep learning, numerical wind tunnel
Aircraft represent one of the most significant technological advancements in human history. Aircraft not only provide convenient and rapid transportation, but also serve as critical tools for exploring the natural world. During the early stages of aircraft design, aerodynamic coefficients must undergo meticulous calculations and analysis to ensure the development of safe, reliable, and efficient aircraft. Generally, aerodynamic coefficients are obtained through numerical simulations using computational fluid dynamics and wind tunnel testing. These methods are effective and accurate; however, they often consume significant amounts of time and resources. With advancements in AI, deep learning techniques have been increasingly applied in aerodynamics research. To reduce the time and cost spent on wind tunnel testing during the aircraft design phase, we apply deep learning techniques to high-performance numerical wind tunnels to analyze aerodynamics in the aviation field. A backpropagation neural network model with an error compensation mechanism is created to enhance the efficiency of analyzing and validating aerodynamic coefficients. This approach minimizes the reliance on physical wind tunnel testing, thereby reducing overall development costs.
Corresponding author: Ching-Huei Huang![]() ![]() This work is licensed under a Creative Commons Attribution 4.0 International License. Cite this article Ching-Huei Huang, Kun-Lin Tsai, and Shih-Ting Tseng, Application of AI Deep Learning in Aircraft Shape Analysis and Validation for High-Performance Numerical Wind Tunnel, Sens. Mater., Vol. 37, No. 7, 2025, p. 3263-3283. |