Open AccessNeural Network Method Based on Concrete Carbonation Depth Prediction
Author(s) -
Dongbo Wu,
Yuanrong Liu,
Yuxue Yin,
Zhiyong Deng,
Zhifu Liu
Publication year - 2021
Publication title -
iop conference series. earth and environmental science
Language(s) - English
Resource type - Journals
eISSN - 1755-1307
pISSN - 1755-1315
DOI - 10.1088/1755-1315/825/1/012020
Subject(s) - carbonation , durability , artificial neural network , cement , compressive strength , computer science , sample (material) , matlab , geotechnical engineering , environmental science , materials science , machine learning , engineering , composite material , database , chemistry , chromatography , operating system
Carbonation is a typical disease that affects the long-term durability of concrete. In this paper, neural network toolbox in MATLAB software was employed to analyze sample parameters such as CO 2 concentration, compressive strength, age and water-cement ratio in concrete carbonation research, and to predict the depth of carbonation. The results show that under the premise of setting reasonable parameters, the sample training results are satisfactory, the average error is about 7%∼14%, which basically meets the precision requirements of the preliminary identification of concrete carbonation depth.