Going with the Speed of Sound: Pushing Neural Surrogates into Highly-turbulent Transonic Regimes

TL;DR

The paper addresses the lack of public, diverse 3D transonic wing data and the challenge of out-of-distribution generalization in neural surrogates for aerospace. It introduces Emmi-Wing, a public dataset of approximately 30,000 high-fidelity RANS cases with both volumetric and surface flow fields, and benchmarks several surrogates, highlighting AB-UPT's superior performance on unseen geometries and inflow conditions. Key contributions include the first public 3D transonic wing dataset that enables lift, drag, and drag–lift analysis, and evidence that AB-UPT can approximate drag–lift Pareto fronts for unseen geometries, suggesting practical utility for rapid aerodynamic design. The dataset and findings support efficient exploration of the aerodynamic design space and potential anomaly detection to improve data quality, with open-source access provided.

Abstract

The widespread use of neural surrogates in automotive aerodynamics, enabled by datasets such as DrivAerML and DrivAerNet++, has primarily focused on bluff-body flows with large wakes. Extending these methods to aerospace, particularly in the transonic regime, remains challenging due to the high level of non-linearity of compressible flows and 3D effects such as wingtip vortices. Existing aerospace datasets predominantly focus on 2D airfoils, neglecting these critical 3D phenomena. To address this gap, we present a new dataset of CFD simulations for 3D wings in the transonic regime. The dataset comprises volumetric and surface-level fields for around $30,000$ samples with unique geometry and inflow conditions. This allows computation of lift and drag coefficients, providing a foundation for data-driven aerodynamic optimization of the drag-lift Pareto front. We evaluate several state-of-the-art neural surrogates on our dataset, including Transolver and AB-UPT, focusing on their out-of-distribution (OOD) generalization over geometry and inflow variations. AB-UPT demonstrates strong performance for transonic flowfields and reproduces physically consistent drag-lift Pareto fronts even for unseen wing configurations. Our results demonstrate that AB-UPT can approximate drag-lift Pareto fronts for unseen geometries, highlighting its potential as an efficient and effective tool for rapid aerodynamic design exploration. To facilitate future research, we open-source our dataset at https://huggingface.co/datasets/EmmiAI/Emmi-Wing.

Figures (15)

• Figure 1: A visual representation of the parameterized 3D wing geometry alongside the four key geometric design parameters and the two inflow condition parameters with their respective sampling ranges used for the design of experiments. The geometry parameters include the span $(b)$, taper ratio $(\lambda)$, sweep angle $(\Lambda)$, and chord root $(c_r)$; inflow parameters are the inflow velocity $(U_{\infty})$ and angle of attack $(\alpha)$.
• Figure 2: Left: Pareto frontier of drag $C_D$ versus lift $C_l$ coefficients for all cases in the dataset. Middle:$C_D$ as a function of angles of attack present in the dataset. Right:$C_l$ as a function of the angles of attack present in the dataset.
• Figure 3: 3D effects on wing.
• Figure 4: Comparison between surface field coefficients on the wing's surface of the CFD (left), AB-UPT surrogate (center) and the error between them (right). The case presented is from the the extrapolation test set with geometry and inflow conditioning parameters within the training range. Corresponding surface pressure and friction profile plots at a certain span length are shown in Figure \ref{['fig:profile_plot_test_extrapol']}.
• Figure 5: Correlation of predicted $C_D$ and $C_l$ from AB-UPT to the ground truth for all cases in the OOD test set. AB-UPT's predictions closely match the ground truth.