NeuralFoil
A learned XFoil for airfoil design
Paper title NeuralFoil: An Airfoil Aerodynamics Analysis Tool Using Physics-Informed Machine Learning
Paper arXiv 2503.16323PDF
Networks trained on XFoil results for about 7.9 million generated cases (56% converged) map an 18-parameter CST airfoil shape, angle of attack, Reynolds number and transition settings to lift, drag, moment and boundary-layer distributions. Symmetries are built into the architecture, known limit cases are blended in analytically, and a learned 'analysis confidence' output flags inputs where XFoil itself would struggle.
Share on LinkedIn, X or Bluesky, or or .
Know related work that belongs here? Suggest a link
Something wrong on this page? Tell the editor