cs.LGOct 24, 2026

FG-ConvNO: A Geometry-Aware Neural Operator for Propeller CFD Prediction

Authors: Yichen Di, Valentin Duruisseaux, Di Zhou, Xinyi Li, Daniel Leibovici, Jean Kossaifi, Anima Anandkumar

Organizations: Tsinghua University · California Institute of Technology · University of Tennessee · NVIDIA

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Computational Fluid Dynamics (CFD) is essential in contemporary engineering design, yet it remains expensive to compute and can take thousands of CPU/GPU hours or more for a single high-fidelity simulation. Recent developments in machine learning for scientific modeling have led to powerful surrogate frameworks that can approximate flow fields and physical quantities by using a fraction of the computational cost. However, most of these methods require largescale homogeneous datasets or simple geometries, exhibiting severe generalization gaps when applied to complex and data-scarce configurations. To address this, we propose the Factorized Geometry Convolutional Neural Operator (FG-ConvNO) for robust geometry-aware propeller pressure prediction under limited training data. We introduce geometryaware encoding, efficient interpolation-based decoding, and auxiliary global quantity supervision. In addition, we incorporate DISCO blocks that use discrete continuous convolutions to make the model discretization agnostic. In our experiments, FG-ConvNO achieves the highest accuracy for both surface pressure and global thrust prediction on a procedurally generated propeller CFD dataset. Geometry-aware encoding and architectural simplification improve stability, while random vertex subsampling enhances robustness. Overall, adding geometric priors to grid-based convolutional operators reduces overfitting on complex propeller geometries and improves generalization in low-data regimes.

Figures & tables

Figure 1: Visualization of procedurally generated propeller geometries. Each sample is colored by time-averaged pressure obtained from simulations, illustrating flow variations across different blade designs.
Figure 2: The Factorized Geometry Convolutional Neural Operator (FG-ConvNO) for propeller pressure prediction. Each propeller geometry is first represented as a centroid-based surface point cloud, where each centroid corresponds to a local patch on the propeller surface. FG-ConvNO then projects vertex features onto multiple latent grids, performs hierarchical convolutional aggregation across the grids, and decodes the resulting latent fields back to the vertex domain for pressure prediction. Compared with FIGConv, FG-ConvNO incorporates geometry-aware conditioning (using boundary conditions, SDFs, and normals), interpolation-based feature decoding, an explicit global prediction head to infer quantities such as thrust and torque, and hierarchical DISCO blocks for discretization convergence, which preserves consistent local support across different grid resolutions.
Figure 3: Comparison of aerodynamic and propulsive forces. (a) Aircraft wings generate lift perpendicular to the inflow. (b) Propellers produce thrust opposite to the inflow. This illustrates how global quantities such as thrust and drag integrate local pressure distributions.
ModelSurface ↓\downarrowGlobal ↓\downarrow
GraphSAGE42.47%8.32×10−38.32\times 10^{-3}
Transolver++27.96%2.41×10−32.41\times 10^{-3}
GINO29.16%3.15×10−33.15\times 10^{-3}
FG-ConvNO (surface only)23.68%1.76×10−31.76\times 10^{-3}
FG-ConvNO (surface + global)24.07%3.43×10−4\boldsymbol{3.43\times 10^{-4}}
Table 1: Performance comparison with existing architectures on the propeller dataset. Both FG-ConvNO variants achieve lower surface error than the graph-based baseline (GraphSAGE) and the recent operator models (GINO and Transolver++), while explicit global supervision substantially improves the accuracy of global quantity prediction.
Figure 4: Ablation on dataset size and random vertex subsampling for training (tsub) . Increasing training samples reduces both surface and global losses, confirming the importance of data scale. Random subsampling consistently improves global prediction and provides stronger gains on surface accuracy when training data are limited.
Figure 5: Quantitative and qualitative ablation of U-Net hierarchy. (a) Comparison of surface and global losses and GPU memory usage across six configurations of nlevelsn_{\text{levels}} and block counts ( nup,ndownn_{\text{up}},n_{\text{down}} ). (b) Example predictions and error maps for representative models (1,1)(1,1) , (1,2)(1,2) , and (3,2)(3,2) .
ConfigurationSurface ↓\downarrowGlobal ↓\downarrow
PointConv (w/ Grid SDF)45.07%7.67×10−37.67\times 10^{-3}
Interp (w/o Grid SDF)26.85%9.72×10−49.72\times 10^{-4}
Proposed24.07%3.43×10−4\boldsymbol{3.43\times 10^{-4}}
Table 2: Ablation on decoder formulation and grid SDF conditioning. Interpolation enhances smoothness and stability, while SDF features improve geometric consistency.
More in this topic

Neural Operators

Neural operators learn mappings between functions, enabling fast surrogates for PDEs and other continuous physical systems.

See all 313 papers

Explore similar work

CardsList

  1. Airfoil2Vec: Spectral Geometry-Conditioned Neural Surrogate Models for Airfoil Aerodynamics and a Downforce-Generating CFD Dataset

    Sep 25, 2026Haitz Sáez de Ocáriz Borde, Flavio Savarino, Andrei Cristian Popescu +3Neural Surrogate ModelingPDE Surrogate Modeling

    Ratio Labs

  2. FlowForge: A Staged Local Rollout Engine for Flow-Field Prediction

    Apr 21, 2026Xiaowen Zhang, Ziming Zhou, Fengnian Zhao +1Efficient InferencePDE Surrogate Modeling

    Shanghai Jiao Tong University · University of Michigan

  3. FSAN: Flow State Attention Network for Aerodynamic Prediction

    Sep 6, 2026Wenxuan Jin, Jianguo Yao, Haibing Guan +1Neural Surrogate ModelingPDE Surrogate Modeling

    Shanghai Key Laboratory of Scalable Computing and Systems · School of Computer Science, Shanghai Jiao Tong University