The Prediction Model of Cold Recycled Materials Mechanical Properties based on the Grey Neural Network

Abstract

This paper aims at predicting the mechanical performance of emulsified asphalt cold recycled material to improve road performance. Measured the data of cold recycled materials mechanical properties under different pavement structures, and using Genetic Algorithm (GA) to optimize the grey neural network model for data analysis and forecast, analyze cold recycled materials mechanical properties influence factors (thickness and modulus of cold regeneration layer, thickness and modulus of the cement stable macadam mixture, and modulus of soil base) by using Grey Correlation theory. Results confirmed that the outputs of Grey Neural Network (GNN) showed that the errors between predicted results and measured results were lower than 6.281%, which means can predict the emulsified asphalt cold recycled materials mechanical properties effectively and quantify the cold regeneration mechanical properties under different factor prediction analysis. It shows that modulus of cement stabilized macadam has great influence on the mechanical properties of cold regeneration materials by using Grey Correlation theory, this study has great significance to improve the road performance cold recycled mixture and chose pavement structure better.