BlendedNet: A Blended Wing Body Aircraft Dataset and Surrogate Model for Aerodynamic Predictions

2025-09-08

AI Summary

The scarcity of publicly available high-fidelity aerodynamic data for blended-wing-body (BWB) aircraft severely hinders data-driven design. Method: This paper introduces the first large-scale, open-source BWB aerodynamic dataset, comprising 999 distinct configurations and 8,830 Spalart–Allmaras RANS simulation results, fully characterizing lift, drag, and spatial distributions of surface pressure and skin-friction coefficients. We further propose an end-to-end point-cloud surrogate model that employs PointNet for permutation-invariant geometric representation and incorporates Feature-wise Linear Modulation (FiLM) to jointly condition predictions on flight conditions and geometric parameters. Contribution/Results: Experimental evaluation demonstrates high-accuracy prediction of surface aerodynamic fields across diverse BWB configurations, achieving a mean relative error below 5%. The dataset and model collectively alleviate the aerodynamic data bottleneck for unconventional configurations, establishing a reproducible benchmark and a novel paradigm for data-driven, rapid BWB design.

Technology Category

• Application Domains: Mobility, Driving & Flight
• Search and Optimization: Sampling/Simulation-based Search
• Planning, Routing, and Scheduling: Model-Based Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Web Mining and Content Analysis: Web data generation and simulation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs

Abstract

BlendedNet is a publicly available aerodynamic dataset of 999 blended wing body (BWB) geometries. Each geometry is simulated across about nine flight conditions, yielding 8830 converged RANS cases with the Spalart-Allmaras model and 9 to 14 million cells per case. The dataset is generated by sampling geometric design parameters and flight conditions, and includes detailed pointwise surface quantities needed to study lift and drag. We also introduce an end-to-end surrogate framework for pointwise aerodynamic prediction. The pipeline first uses a permutation-invariant PointNet regressor to predict geometric parameters from sampled surface point clouds, then conditions a Feature-wise Linear Modulation (FiLM) network on the predicted parameters and flight conditions to predict pointwise coefficients Cp, Cfx, and Cfz. Experiments show low errors in surface predictions across diverse BWBs. BlendedNet addresses data scarcity for unconventional configurations and enables research on data-driven surrogate modeling for aerodynamic design.

Problem

• Providing aerodynamic dataset for blended wing body aircraft geometries
• Enabling surrogate modeling for pointwise aerodynamic prediction
• Addressing data scarcity in unconventional aircraft configurations

Innovation

• Public aerodynamic dataset with 8991 RANS cases
• PointNet regressor predicts geometric parameters from point clouds
• FiLM network conditioned on parameters predicts surface coefficients