arXivDaily arXiv每日学术速递 周一至周五更新
北美退伍军人节(Veterans Day),12 日无更新
arXiv 2209.04285physics.flu-dyn
收藏

利用物理信息神经网络生成湍流态

Generation of Turbulent States using Physics-Informed Neural Networks

Sofia Angriman, Pablo Cobelli, Pablo Mininni, Martín Obligado, Patricio Clark Di Leoni

摘要:

在模拟湍流时,通常情况下,有关强迫项或边界条件的信息要么不可用,要么过于复杂,实施起来成本高昂。相反,一些流动特征,例如平均速度分布或其统计矩,可以通过实验或观察来获得。我们提出了一种基于物理信息的神经网络生成湍流状态受一组给定的条件的方法。物理通知方法确保最终状态近似于有效流。我们展示了不同的统计条件,可用于准备状态,实验和大气问题的动机的例子。最后,我们展示了两种缩放准备状态的分辨率的方法。一种是通过使用多个并行神经网络。另一种方法是使用nudging,这是一种基于同步的数据同化技术,利用了专业数值求解器的力量。

英文摘要:

When modelling turbulent flows, it is often the case that information on the forcing terms or the boundary conditions is either not available or overly complicated and expensive to implement. Instead, some flow features, such as the mean velocity profile or its statistical moments, may be accessible through experiments or observations. We present a method based on physics-informed neural networks to generate turbulent states subject to a set of given conditions. The physics-informed method ensures the final state approximates a valid flow. We show examples of different statistical conditions that can be used to prepare states, motivated by experimental and atmospheric problems. Lastly, we show two ways of scaling the resolution of the prepared states. One is through the use of multiple and parallel neural networks. The other uses nudging, a synchronization-based data assimilation technique that leverages the power of specialized numerical solvers.

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。

↑