Influence of interface transition zone on effective elastic property of heterogeneous materials with an artificial neural network study

被引:1
|
作者
Xue, Jing [1 ]
Cao, Yajun [2 ]
Burlion, Nicolas [1 ]
Shao, Jianfu [1 ,3 ]
机构
[1] Univ Lille, CNRS, Cent lille, LaMcube, Lille, France
[2] Hohai Univ, Key Lab Minist Educ Geomech & Embankment Engn, Nanjing, Peoples R China
[3] Univ Lille, CNRS, Cent Lille, LaMcube, F-59000 Lille, France
关键词
artificial neural network; concrete; cement-based material; effective elastic properties; interface transition zone; multi-scale modeling; NUMERICAL-METHOD; CEMENT PASTE; BULK MODULUS; CONCRETE; AGGREGATE; PREDICTION; STRENGTH; MODEL; MICROSTRUCTURE; DIFFUSIVITY;
D O I
10.1002/nag.3508
中图分类号
P5 [地质学];
学科分类号
0709 ; 081803 ;
摘要
In this study, the influence of interface transition zone (ITZ) on the elastic properties of concrete and rock like heterogeneous materials is investigated. Direct simulations of a representative volume element are realized by using a fast Fourier transform based method and the obtained results are used as the reference solutions. Some widely used analytical homogenization models are evaluated by comparing the theoretical predictions with the reference solutions. Based on this evaluation, an artificial neural network (ANN) model is constructed in order to improve the analytical models. The proposed ANN model is trained, tested and validated against the reference solution. Its efficiency and good accuracy are clearly demonstrated.
引用
收藏
页码:1134 / 1151
页数:18
相关论文
共 38 条