Premium
Artificial neural network approach to predict the fracture parameters of the size effect model for concrete
Author(s) -
Yan Y.,
Ren Q.,
Xia N.,
Shen L.,
Gu J.
Publication year - 2015
Publication title -
fatigue and fracture of engineering materials and structures
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.887
H-Index - 84
eISSN - 1460-2695
pISSN - 8756-758X
DOI - 10.1111/ffe.12309
Subject(s) - artificial neural network , fracture (geology) , parametric statistics , yield (engineering) , regression analysis , fracture mechanics , materials science , experimental data , structural engineering , computer science , mathematics , engineering , statistics , machine learning , composite material
The fracture parameters (the fracture energy and the effective length of the fracture process zone for an infinitely large specimen) in the size effect model of concrete exhibit a large scatter as measured in most of the experimental studies. This phenomenon is ubiquitous and has presented a great challenge to characterize the structural failure over the last decade. In order to remove the perplexing issue, this paper develops two models to predict the two fracture parameters using the artificial neural network (ANN) methodology. The proposed models are verified by using 77 experimental data collected from the literature. The results demonstrate that the two ANN‐based size effect model models (ANN‐I and ANN‐II) are viable for predicting the fracture parameters and yield more accurate results than those obtained from the conventional regression formulations. Additionally, a parametric study is employed to evaluate the impact of each independent material parameter on the fracture parameters.