Fast Aerodynamics Prediction of Wedge Tail Airfoils Using Multi-head Perceptron Network.

Wedge tail airfoils refer to the addition of a wedge section to the tail of an airfoil. This modification has demonstrated greater efficiency in increasing the lift forces compared to the base airfoil design. Such airfoils are particularly important in scenarios where increased lift forces are required, such as in the fast and efficient maneuvering of marine and aerospace vehicles. This study aimed to predict the flow fields and aerodynamic coefficients of wedge tail airfoils using a multi-head perceptron (MHP) network and classical machine learning (ML) algorithms, including k-nearest neighbors, decision tree, and random forest. These predictions were based on airfoil sections, x–y grid coordinates, Mach number, and angle of attack, eliminating the need to solve the Navier–Stokes (NS) equations. The database required for training the MHP network and the ML models were generated using the Ansys solver, which solved the NS equations. The results of the models were compared, and it was observed that the MHP network consistently outperformed the others in terms of prediction accuracy. Additionally, the prediction process demonstrated a significant speed improvement of 125 times compared to conventional computational fluid dynamics techniques. Once the training was completed, the flow fields and aerodynamic coefficients for a wedge tail airfoil could be obtained within seconds. [ABSTRACT FROM AUTHOR]