Modeling Reynolds stress anisotropy invariants via machine learning

作     者：Xianglin Shan Xuxiang Sun Wenbo Cao Weiwei Zhang Zhenhua Xia 单湘淋;孙旭翔;曹文博;张伟伟;夏振华

作者机构：School of AeronauticsNorthwestern Polytechnical UniversityXi’an710072China Department of Engineering MechanicsZhejiang UniversityHangzhou310027China 

基　　金：supported by the National Natural Science Foundation of China(Grant No.92152301) 

出 版 物：《Acta Mechanica Sinica》 (力学学报（英文版）)

年 卷 期：2024年第40卷第6期

页      码：50-63页

摘      要：The presentation and modeling of turbulence anisotropy are crucial for studying large-scale turbulence structures and constructing turbulence ***,accurately capturing anisotropic Reynolds stresses often relies on expensive direct numerical simulations(DNS).Recently,a hot topic in data-driven turbulence modeling is how to acquire accurate Reynolds stresses by the Reynolds-averaged Navier-Stokes(RANS)simulation and a limited amount of data from *** existing studies use mean flow characteristics as the input features of machine learning models to predict high-fidelity Reynolds stresses,but these approaches still lack robust generalization *** this paper,a deep neural network(DNN)is employed to build a model,mapping from tensor invariants of RANS mean flow features to the anisotropy invariants of high-fidelity Reynolds *** the aspects of tensor analysis and input-output feature design,we try to enhance the generalization of the model while preserving invariance.A functional framework of Reynolds stress anisotropy invariants is derived *** irreducible invariants are then constructed from a tensor group,serving as alternative input features for ***,we propose a feature selection method based on the Fourier transform of periodic *** results demonstrate that the data-driven model achieves a high level of accuracy in predicting turbulence anisotropy of flows over periodic hills and converging-diverging ***,the well-trained model exhibits strong generalization capabilities concerning various shapes and higher Reynolds *** approach can also provide valuable insights for feature selection and data generation for data-driven turbulence models.

主 题 词：Reynolds stress Anisotropy invariant Tensor analysis Machine learning 

学科分类：080704[080704] 080103[080103] 08[工学] 0807[工学-电子信息类] 0801[工学-力学类] 

核心收录：

D　O　I：10.1007/s10409-024-23629-x

馆 藏 号：203128204...