FSAN：用于气动预测的流态注意力网络
FSAN: Flow State Attention Network for Aerodynamic Prediction
- School of Computer Science, Shanghai Jiao Tong University（上海交通大学计算机学院）
- Shanghai Key Laboratory of Scalable Computing and Systems（上海市可扩展计算与系统重点实验室）
机构由 AI 辅助整理，请以论文原文为准。
AI总结：
针对现有深度学习气动预测模型忽略局部流动差异的问题，提出流态注意力网络（FSAN），通过软分配划分流态并交互几何与流动信息，在两个基准上分别降低REL-L2误差超20%和10%。
AI中文摘要：
准确的气动预测对于设计燃油效率高且安全的运输系统（如飞机和汽车）至关重要，然而传统的计算流体力学（CFD）模拟仍然计算成本高昂且需要专业知识，严重限制了其在迭代设计和实时分析中的应用。现有的深度学习替代模型存在两个主要局限：（i）它们仅在流动条件范围狭窄的数据集上进行评估，其在复杂流动条件下的性能尚未得到验证；（ii）它们将全局流动条件视为一个单一向量，均匀地注入到所有表面点上，忽略了不同几何区域经历不同的局部流动现象，这降低了在复杂流动条件下的预测精度。为了解决这些局限，我们提出了流态注意力网络（FSAN）。FSAN分别编码点云和流动条件，然后通过可学习的软分配将几何体划分为多个流态，并使用流动特征更新这些状态表示，进而通过状态变化影响点云特征。这实现了几何与流动信息之间细粒度、特定状态的交互。在两个公认的气动基准上的大量实验表明，FSAN在本文比较的方法中达到了最高精度，但计算成本更高。在Emmi-Wing上，与最强基线（Transolver）相比，FSAN将相对L2（REL-L2）误差降低了超过20%；在DrivAerNet++上，与最强基线（AdaField）相比，降低了10%。这些结果确立了FSAN作为在具有多样流动条件和几何形状的公共基准上的一种有前景的神经替代模型。
英文摘要：
Accurate aerodynamic prediction is critical for designing fuel-efficient and safe transportation systems such as aircraft and automobiles, yet traditional computational fluid dynamics (CFD) simulations remain computationally expensive and expertise-intensive, severely limiting their use in iterative design and real-time analysis. Existing deep learning surrogates suffer from two major limitations: (i) they are evaluated on datasets with narrow flow-condition ranges, leaving their performance under complex flow conditions undemonstrated; (ii) they treat global flow conditions as a single vector injected uniformly across all surface points, ignoring that different geometric regions experience distinct local flow phenomena, which degrades prediction accuracy under complex flow conditions. To address these limitations, we propose the Flow State Attention Network (FSAN). FSAN separately encodes point cloud and flow conditions, then partitions the geometry into multiple flow states via learnable soft assignments, and uses flow features to update these state representations, which in turn influence point cloud features through state changes. This enables fine-grained, state-specific interaction between geometry and flow information. Extensive experiments on two well-recognized aerodynamic benchmarks demonstrate that FSAN achieves the highest accuracy among the methods compared in this work at a higher computational cost. On Emmi-Wing, FSAN reduces the Relative L2 (REL-L2) error by over 20\% compared to the strongest baseline (Transolver), and on DrivAerNet++, it achieves a 10\% reduction compared to the strongest baseline (AdaField). These results establish FSAN as a promising neural surrogate on public benchmarks with diverse flow conditions and geometries.
