AirCANS: CFD 2D Mesh Optimisation-based Airfoil Classification and Assessment using Neural Networks

2025/06/27 by Lushun Fan, Fan, Lushun, Xia, Yuqin +4
Engineering · Physics and Astronomy · #Biomimetic flight and propulsion mechanisms #Computational Fluid Dynamics and Aerodynamics #Computational Physics (physics.comp-ph) #FOS: Physical sciences #Model Reduction and Neural Networks

paper · pdf · doi:10.48550/arxiv.2506.22662

openalex publication_date 2025/06/27 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

This study explores the possibilities of automating the loading, classification and assessment of Computational Fluid Dynamics (CFD) mesh data by Convolutional Neural Networks (CNNs). The research aim is finding a feasible way to quickly make classification and assessment on airfoil mesh data. For this purpose, this study designed a new framework named CFD-based airfoil Classification and Assessment Network (AirCANS) for CFD mesh data which including the data loader and improved the CNN structure to achieve our target. In our research, we found that CNNs are fully adaptable as well as understandable to CFD airfoil mesh data structures, which suggests that our hypothesis is successful and that neural networks can be used to have a greater positive impact on the CFD industry, such as it can be used to refine the mesh and accelerate the solution. This could allow CFD to spend much less time.

Citations

• MeshCNN Fundamentals: Geometric Learning through a Reconstructable Representation
• Machine Learning-Based Optimal Mesh Generation in Computational Fluid Dynamics
• Machine Learning for Fluid Mechanics
• Deep Learning Methods for Reynolds-Averaged Navier–Stokes Simulations of Airfoil Flows
• Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations
• PointNet: Deep Learning on Point Sets for 3D Classification and\n Segmentation

Related

• DeepCFD: Efficient near-ground airfoil lift coefficient approximation with deep convolutional neural networks
• Accelerated Airfoil Design Using Neural Network Approaches
• Loop2Net: Data-Driven Generation and Optimization of Airfoil CFD Meshes from Sparse Boundary Coordinates
• Efficient aerodynamic coefficients prediction with a long sequence neural network
• A Hybrid CNN-Cheby-KAN Framework for Efficient Prediction of Two-Dimensional Airfoil Pressure Distribution