Deep–learning framework for aircraft aerodynamics prediction

The present paper aims to develop a deep--learning framework able to predict distributed quantities of aircrafts flying in transonic regime, which are critical for the determination of aerodynamic loads and aeroelastic analysis. Angle of attack and Mach number are chosen as the two independent parameters for the reduced--order models. A comparative assessment of the proposed non--linear model is made with Proper Orthogonal Decomposition approach in order to highlight strengths and weaknesses of each method. The accuracy of the data--driven machine--learning method in modelling steady--state aerodynamics is assessed with three benchmark cases of 3D--wings in transonic regime. Despite the challenges of the analyzed scenarios, promising results are obtained for each test case, showing the effectiveness of the model implemented.