BlendedNet Multi-Fidelity Extension Dataset: HF-LF Aerodynamics for BWB Aircraft

This dataset contains 7,871 paired aerodynamic samples for Blended Wing Body (BWB) aircraft, designed to support multi-fidelity machine learning and surrogate modeling research. It extends the original BlendedNet high-fidelity dataset by adding corresponding low-fidelity evaluations. Dataset Contents: Input: 14-dimensional vector (10 geometric variables, 4 flight conditions). Output: 4 scalar aerodynamic coefficients per sample: Lift (CL) and Drag (CD) for both fidelities (CLHF, CDHF, CLLF, CDLF). High-Fidelity (HF): RANS CFD results (NASA FUN3D) from the original BlendedNet dataset. Low-Fidelity (LF): Vortex Lattice Method (VLM) results computed using OpenVSP (VSPAERO). Usage: This resource is intended for researchers developing multi-fidelity algorithms (e.g., DeepONets, Cokriging, Transfer Learning) who require a large-scale, physics-aligned benchmark.