Airfoil​

​​​​

Task: Predict Surface Pressure and Skin Friction Distributions for varying airfoils and flow conditions.

​

This benchmark database provides high-fidelity surface aerodynamic data for modeling pressure and skin friction distributions across airfoil geometries, supporting surrogate modeling challenges in applied aerodynamics. Using the well-established RAE2822 transonic airfoil as a baseline, the dataset incorporates airfoil variations generated through Class Shape Function Transformations (CST), defining nine design parameters to explore diverse aerodynamic shapes.

​

The dataset includes 597 Reynolds-Averaged Navier-Stokes (RANS) simulations performed with the DLR TAU solver and Spalart-Allmaras turbulence model, covering a broad range of Mach numbers, angles of attack, and Reynolds numbers. A Sobol sequence-based Design of Experiment (DoE) ensures efficient coverage of the design space. Each simulation file contains surface coordinates, normal vectors, pressure coefficients, and skin friction coefficients, along with global aerodynamic parameters.

​

Designed as a benchmark for machine learning applications, this dataset enables researchers to develop and validate data-driven aerodynamic models, optimizing airfoil performance predictions within a representative civil aircraft operating envelope. The dataset is publicly available and structured to facilitate further data expansion and surrogate model training.