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 2509.01293cs.LGphysics.comp-phphysics.flu-dyn
收藏

等变U形神经算子用于Cahn-Hilliard相场模型

Equivariant U-Shaped Neural Operators for the Cahn-Hilliard Phase-Field Model

Centre for Computational Science, Department of Chemistry, University College London, London, UK(计算科学中心,化学系,伦敦大学学院,伦敦,英国);Institute for Mechanics, Computational Mechanics Group, Technical University of Darmstadt, Germany(力学研究所,计算力学组,达姆施塔特技术大学,德国);Department of Physics, Eindhoven University of Technology, Eindhoven, The Netherlands(物理系,埃因霍温理工大学,埃因霍温,荷兰);Department of Mechanical Engineering, University College London, London, UK(机械工程系,伦敦大学学院,伦敦,英国);College of Civil Engineering and Mechanics, Lanzhou University, Lanzhou 730000, China(土木工程与力学学院,兰州大学,兰州,中国);Informatics Institute, University of Amsterdam, Amsterdam, The Netherlands(信息学院,阿姆斯特丹大学,阿姆斯特丹,荷兰);Centre for Advanced Research Computing, University College London, UK(高级研究计算中心,伦敦大学学院,英国)

Xiao Xue, Marco F. P. ten Eikelder, Tianyue Yang, Yiqing Li, Kan He, Shuo Wang, Peter V. Coveney

AI总结:

提出等变U形神经算子(E-UNO),结合全局谱卷积与多分辨率架构并调节平移等变性,高效准确预测Cahn-Hilliard相场演化,优于现有基线。

摘要:

由Cahn-Hilliard方程控制的二元混合物中的相分离,在材料科学和软物质中的界面动力学中起着核心作用。虽然数值求解器是准确的,但它们通常计算成本高昂,并且在不同初始条件和几何形状下缺乏灵活性。神经算子通过学习函数空间之间的解算子提供了一种数据驱动的替代方案,但当前的架构往往无法捕捉多尺度行为,并忽略了潜在的物理对称性。在这里,我们展示了等变U形神经算子(E-UNO)可以从过去的动力学短历史中学习相场变量的演化,实现跨空间和时间的准确预测。该模型将全局谱卷积与多分辨率U形架构相结合,并调节平移等变性以与底层物理对齐。E-UNO在细尺度和高频结构上尤其优于标准傅里叶神经算子和U形神经算子基线。通过编码对称性和尺度层次,该模型具有更好的泛化能力,需要更少的训练数据,并产生物理一致的动力学。这确立了E-UNO作为复杂相场系统的高效替代模型。

英文摘要:

Phase separation in binary mixtures, governed by the Cahn-Hilliard equation, plays a central role in interfacial dynamics across materials science and soft matter. While numerical solvers are accurate, they are often computationally expensive and lack flexibility across varying initial conditions and geometries. Neural operators provide a data-driven alternative by learning solution operators between function spaces, but current architectures often fail to capture multiscale behavior and neglect underlying physical symmetries. Here we show that an equivariant U-shaped neural operator (E-UNO) can learn the evolution of the phase-field variable from short histories of past dynamics, achieving accurate predictions across space and time. The model combines global spectral convolution with a multi-resolution U-shaped architecture and regulates translation equivariance to align with the underlying physics. E-UNO outperforms standard Fourier neural operator and U-shaped neural operator baselines, particularly on fine-scale and high-frequency structures. By encoding symmetry and scale hierarchy, the model generalizes better, requires less training data, and yields physically consistent dynamics. This establishes E-UNO as an efficient surrogate for complex phase-field systems.

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。

↑