Prediction and sensitivity analysis of long-term skid resistance of epoxy asphalt mixture based on GA-BP neural network

•GA-BP neural network was used to model skid resistance of epoxy asphalt mixture.•A comprehensive sensitivity analysis method is proposed based on neural network.•The designed GA-BP model has shown good agreement with experimental results.•Skid resistance has a significant negative correlation with...

Full description

Saved in:
Bibliographic Details
Published in:Construction & building materials 2018-01, Vol.158, p.614-623
Main Authors: Zheng, Dong, Qian, Zhen-dong, Liu, Yang, Liu, Chang-bo
Format: Article
Language:English
Subjects:
Analysis
Artificial neural networks
Asphalt pavements
Chemical properties
Epoxy asphalt mixture
Epoxy resins
GA-BP neural network
Logistic chaotic mapping
Mechanical properties
Sensitivity analysis
Skid-resistant prediction model
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•GA-BP neural network was used to model skid resistance of epoxy asphalt mixture.•A comprehensive sensitivity analysis method is proposed based on neural network.•The designed GA-BP model has shown good agreement with experimental results.•Skid resistance has a significant negative correlation with BC and κ-value.•Bulk specific gravity is the most important skid-resistant influencing factor. The objective of this study is to investigate the relationship between long-term skid resistance of epoxy asphalt mixture (EAM) and multiple engineering parameters involving mixture design parameters, construction parameters and operation parameters. Firstly, a database of 124 data sets was obtained, including optimal binder content, aggregate gradation characteristics, bulk specific gravity, air-void content and load repetitions for input parameters, and long-term skid resistance of EAM simulated by an accelerated pavement test for output. Secondly, using the database, an optimized GA-BP neural network model (i.e. GA-BP model) was established to predict the long-term skid resistance, and then a comprehensive sensitivity analysis was conducted to explore the effect of input parameters on the skid-resistant evolution based on the trained neural network. Results show that the optimized GA-BP model can effectively predict the long-term skid resistance of EAM, and the long-term skid resistance has a significant negative correlation with binder content and shape characteristic of aggregate gradation. In addition, bulk specific gravity is the most important factor influencing the long-term skid resistance, and also has the most remarkable interaction effect with other input parameters.
ISSN:0950-0618
1879-0526
DOI:10.1016/j.conbuildmat.2017.10.056