Feature identification in complex fluid flows by convolutional neural networks

作     者：Shizheng Wen Michael W.Lee Kai M.Kruger Bastos Ian K.Eldridge-Allegra Earl H.Dowell Shizheng Wen;Michael W.Lee;Kai M.Kruger Bastos;Ian K.Eldridge-Allegra;Earl H.Dowell

作者机构：Department of MathematicsETH ZurichZürich 8092Switzerland NASA Langley Research CenterHamptonVA 23666USA RivianSan FranciscoCA 94080USA Department of Mechanical Engineering and Materials ScienceDuke UniversityDurhamNC 27708USA 

出 版 物：《Theoretical & Applied Mechanics Letters》 (力学快报（英文版）)

年 卷 期：2023年第13卷第6期

页      面：447-454页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 080704[工学-流体机械及工程] 081104[工学-模式识别与智能系统] 080103[工学-流体力学] 08[工学] 0807[工学-动力工程及工程热物理] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 0801[工学-力学（可授工学、理学学位）] 

主　　题：Subsonic buffet flows Feature identification Convolutional neural network Long-short term memory 

摘      要：Recent advancements have established machine learning s utility in predicting nonlinear fluid dynamics,with predictive accuracy being a central motivation for employing neural ***,the pattern recognition central to the networks function is equally valuable for enhancing our dynamical insight into the complex fluid *** this paper,a single-layer convolutional neural network(CNN)was trained to recognize three qualitatively different subsonic buffet flows(periodic,quasi-periodic and chaotic)over a high-incidence airfoil,and a near-perfect accuracy was obtained with only a small training *** convolutional kernels and corresponding feature maps,developed by the model with no temporal information provided,identified large-scale coherent structures in agreement with those known to be associated with buffet *** to hyperparameters including network architecture and convolutional kernel size was also *** coherent structures identified by these models enhance our dynamical understanding of subsonic buffet over high-incidence airfoils over a wide range of Reynolds numbers.