arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2511.01535physics.flu-dynstat.ML
收藏

HFNO:一种面向湍流的可解释数据驱动分解策略

HFNO: an interpretable data-driven decomposition strategy for turbulent flows

Sorbonne Université(索邦大学);KAUST(阿卜杜拉国王科技大学);Université Gustave Eiffel(古斯塔夫·埃菲尔大学);Conservatoire National des Arts et Métiers(国立工艺院)

Marco Cayuela, Vincent Le Chenadec, Peter Schmid, Taraneh Sayadi

AI总结:

提出HFNO架构,通过并行处理波数区间分离跨尺度流体行为,提升湍流降阶建模的可解释性,并在多个复杂动力学系统上验证了其多尺度分解能力。

摘要:

Fourier Neural Operators(FNOs,傅里叶神经算子)通过利用傅里叶变换建立与底层物理原理的联系,在映射函数空间方面展现出了卓越的准确性。然而,其不透明的内部机制往往构成了物理可解释性的障碍。本研究引入了 Hierarchical Fourier Neural Operators(HFNOs,分层傅里叶神经算子),这是一种专为湍流降阶建模量身定制的基于 FNO 的新型架构,旨在通过显式地分离跨尺度的流体行为来增强可解释性。所提出的架构并行处理波数区间,从而能够近似色散关系和非线性相互作用。输入被提升到更高维空间,进行傅里叶变换,并被划分为多个波数区间。每个区间由一个 Fully Connected Neural Network(FCNN,全连接神经网络)处理,随后将输出进行填充、求和,并逆变换回物理空间。最终的变换通过以下架构之一作为修正模型来细化物理空间中的输出:Convolutional Neural Network(CNN,卷积神经网络)和 Echo State Network(ESN,回声状态网络)。我们在一系列复杂度递增的动力学系统上评估了所提出的模型:首先是一维 Kuramoto-Sivashinsky 方程,其次是二维 Kolmogorov 流,最后是在给定近壁速度场的情况下预测湍流槽道流的壁面切应力。在所有测试案例中,该模型都展现了其跨各种尺度分解湍流的能力,为提高此类流体的可解释性和多尺度建模开辟了可能性。

英文摘要:

Fourier Neural Operators (FNOs) have demonstrated exceptional accuracy in mapping functional spaces by leveraging Fourier transforms to establish a connection with underlying physical principles. However, their opaque inner workings often constitute an obstacle to physical interpretability. This work introduces Hierarchical Fourier Neural Operators (HFNOs), a novel FNO-based architecture tailored for reduced-order modeling of turbulent fluid flows, designed to enhance interpretability by explicitly separating fluid behavior across scales. The proposed architecture processes wavenumber bins in parallel, enabling the approximation of dispersion relations and non-linear interactions. Inputs are lifted to a higher-dimensional space, Fourier-transformed, and partitioned into wavenumber bins. Each bin is processed by a Fully Connected Neural Network (FCNN), with outputs subsequently padded, summed, and inverse-transformed back into physical space. A final transformation refines the output in physical space as a correction model, by means of one of the following architectures: Convolutional Neural Network (CNN) and Echo State Network (ESN). We evaluate the proposed model on a series of increasingly complex dynamical systems: first on the one-dimensional Kuramoto-Sivashinsky equation, then on the two-dimensional Kolmogorov flow, and finally on the prediction of wall shear stress in turbulent channel flow, given the near-wall velocity field. In all test cases, the model demonstrates its ability to decompose turbulent flows across various scales, opening up the possibility of increased interpretability and multiscale modeling of such flows.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑