Unsupervised machine learning classification for accelerating FE2 multiscale fracture simulations

被引:5
作者
Chaouch, Souhail [1 ]
Yvonnet, Julien [1 ]
机构
[1] Univ Gustave Eiffel, MSME, CNRS, UMR 8208, F-77454 Marne La Vallee, France
关键词
Multi-scale; Unsupervised machine learning; k-means; FE2; Cracks; Damage; CONSISTENT CLUSTERING ANALYSIS; COMPUTATIONAL HOMOGENIZATION; HYPER-REDUCTION; HETEROGENEOUS MATERIALS; MODEL; DAMAGE; BEHAVIOR; FAILURE;
D O I
10.1016/j.cma.2024.117278
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
An approach is proposed to accelerate multiscale simulations of heterogeneous quasi-brittle materials exhibiting an anisotropic damage response. The present technique uses unsupervised machine learning classification based on k-means clustering to select integration points in the macro mesh within an FE2 2 strategy to track redundant micro nonlinear problems and to avoid unnecessary Representative Volume Element (RVE) calculations. More specifically, a classification vector including strains and internal damage variables is defined for each macro integration point. The macro internal damage variables are constructed using harmonic analysis of damage. At each step of the macro iterations, the integrations points are grouped into clusters and only one nonlinear problem is solved for each cluster. As a result, the computations are accelerated within an FE2 2 scheme by reducing the total number of RVE problems to be solved. The developed algorithm includes a macro regularization and an arc-length technique to capture macro snap-back due to the softening. Applications are proposed to simulate the response of different heterogeneous quasi-brittle materials with strong anisotropic responses. speed-up factors of the order of 12 to 15 can be achieved without the need to build a database, and without reduced-order modeling approximations at the micro level. Estimates of structural strength can be obtained with Speed-up factors between 45 and 85.
引用
收藏
页数:24
相关论文
共 70 条
[61]   Localizing gradient damage model with micro inertia effect for dynamic fracture [J].
Wang, Zhao ;
Shedbale, Amit Subhash ;
Kumar, Sachin ;
Poh, Leong Hien .
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING, 2019, 355 :492-512
[62]   A recurrent neural network-accelerated multi-scale model for elasto-plastic heterogeneous materials subjected to random cyclic and non-proportional loading paths [J].
Wu, Ling ;
Van Dung Nguyen ;
Kilingar, Nanda Gopala ;
Noels, Ludovic .
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING, 2020, 369 (369)
[63]   A Multiscale Failure Modeling Framework for Strain Localization in Quasi-Brittle Materials [J].
Xing, Wenjin ;
Miller, Anthony D. .
INTERNATIONAL JOURNAL OF COMPUTATIONAL METHODS, 2021, 18 (10)
[64]   The reduced model multiscale method (R3M) for the non-linear homogenization of hyperelastic media at finite strains [J].
Yvonnet, J. ;
He, Q.-C. .
JOURNAL OF COMPUTATIONAL PHYSICS, 2007, 223 (01) :341-368
[65]  
Yvonnet J, 2019, SOLID MECH APPL, V258, P1, DOI 10.1007/978-3-030-18383-7
[66]   Numerically explicit potentials for the homogenization of nonlinear elastic heterogeneous materials [J].
Yvonnet, J. ;
Gonzalez, D. ;
He, Q. -C. .
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING, 2009, 198 (33-36) :2723-2737
[67]   Reducing internal variables and improving efficiency in data-driven modelling of anisotropic damage from RVE simulations [J].
Yvonnet, Julien ;
He, Qi-Chang ;
Li, Pengfei .
COMPUTATIONAL MECHANICS, 2023, 72 (01) :37-55
[68]   A data-driven harmonic approach to constructing anisotropic damage models with a minimum number of internal variables [J].
Yvonnet, Julien ;
He, Qi-Chang ;
Li, Pengfei .
JOURNAL OF THE MECHANICS AND PHYSICS OF SOLIDS, 2022, 162
[69]  
Yvonnet J, 2019, SOLID MECH APPL, V258, P161, DOI 10.1007/978-3-030-18383-7_9
[70]   COMPUTATIONAL HOMOGENIZATION METHOD AND REDUCED DATABASE MODEL FOR HYPERELASTIC HETEROGENEOUS STRUCTURES [J].
Yvonnet, Julien ;
Monteiro, Eric ;
He, Qi-Chang .
INTERNATIONAL JOURNAL FOR MULTISCALE COMPUTATIONAL ENGINEERING, 2013, 11 (03) :201-225