HGPTrans: Hierarchical Graph-Pooling Transolver for Automotive Aerodynamic Drag Coefficient Prediction

2026-09-24

AI Summary

This study addresses the computational bottleneck of CFD simulations in early-stage automotive design by proposing a graph neural network-based method for rapid vehicle drag coefficient prediction. Taking vehicle surface meshes as input, the model integrates graph isomorphism convolutions with a physics-aware sliced attention mechanism to achieve local geometric encoding and linear-complexity global interactions. An information-redundancy-aware hierarchical pooling strategy is introduced to preserve critical structural features, complemented by transfer learning to enhance generalization. Experiments demonstrate that the proposed method achieves the lowest error on the DrivAerNet dataset, yielding relative prediction errors of only 1.5%–2.1% on real-world vehicles. With a single inference time of 0.293 seconds, it accelerates evaluation by several orders of magnitude compared to CFD, effectively supporting the efficient assessment of large-scale geometric design variants.

Abstract

Accurate and rapid prediction of the aerodynamic drag coefficient ($C_D$) is essential for vehicle design, particularly during early-stage styling iterations where a large number of candidate geometries must be evaluated. Although computational fluid dynamics (CFD) provides reliable aerodynamic estimates, its high computational cost, typically requiring hours to days for a single configuration, limits its use in large-scale design exploration. This paper proposes HGPTrans, a hierarchical graph-pooling network with Transolver-based attention, to directly predict $C_D$ from vehicle surface meshes. Motivated by the fact that vehicle aerodynamics depends on both local geometric features and long-range interactions among spatially distant surface regions, HGPTrans integrates three complementary components. Graph isomorphism convolutions encode discriminative local geometry, physics-aware slice attention captures global interactions with linear computational complexity, and information-redundancy-aware hierarchical pooling progressively removes redundant nodes while preserving informative geometric structures. The model is trained and evaluated on the large-scale DrivAerNet and DrivAerNet++ datasets, where it achieves the lowest mean absolute error and mean squared error among the evaluated baselines. Its generalization capability is further assessed through transfer learning on a real-vehicle dataset containing both sedans and SUVs, achieving relative $L_1$ errors of 1.56% (sedans) and 2.12% (SUVs) with an inference time of approximately $0.293$ s per vehicle. This corresponds to an acceleration of several orders of magnitude relative to high-fidelity CFD while keeping the predicted drag coefficients within a few percent of the CFD reference. Ablation studies confirm each component's contribution and reveal the effects of depth and pooling ratio.

Problem

• Aerodynamic drag coefficient prediction
• Computational fluid dynamics
• Vehicle design
• Surrogate modeling

Innovation

• Hierarchical Graph Pooling
• Transolver Attention
• Aerodynamic Drag Prediction
• Graph Isomorphism Convolution
• Physics-aware Learning