Search papers, labs, and topics across Lattice.
This paper introduces the Multi-scale Feature Enhanced Graph Neural Network (ME-GNN), designed to enhance fluid dynamics predictions in complex geometries while addressing the limitations of traditional graph neural networks (GNNs). By employing a two-step message-passing mechanism and integrating an Attention U-Net with uniform grid discretization, ME-GNN effectively captures both local and global features in large-scale meshes. The model achieves state-of-the-art performance on benchmark datasets, significantly reducing prediction errors in velocity and surface pressure fields compared to existing methods.
Achieving state-of-the-art accuracy in fluid dynamics predictions, ME-GNN dramatically reduces computational costs associated with complex geometries.
Industrial design in fields such as vehicle and aerospace engineering often relies on large-scale numerical simulations to evaluate fluid dynamics performance, which can incur substantial computational costs. Deep neural networks have shown promise in improving simulation efficiency, especially graph neural networks (GNNs), which demonstrate great potential due to their flexibility with unstructured data. However, GNNs face challenges when dealing with tasks involving complex geometries and large-scale meshes. In this paper, we propose the Multi-scale Feature Enhanced Graph Neural Network (ME-GNN) to tackle these challenges. ME-GNN employs a graph neural network with a two-step message-passing mechanism to capture detailed local features effectively. Additionally, it integrates an Attention U-Net with uniform grid discretization, enabling the extraction of both fine and coarse features. The model also utilizes K-hop sampling to construct subgraphs, facilitating efficient training on large datasets while preserving detailed local features. We evaluated ME-GNN on three benchmark datasets and achieved state-of-the-art results: a relative L2 error of 0.0196 for the velocity field and 0.0556 for the surface pressure on ShapeNet-Car, a normalized mean squared error of 0.0033 for the flow field on AirfRANS, and a relative L2 error of 0.1416 for the surface pressure on DrivAerNet.
