Optimization of surface free energy parameters for asphalt binder-aggregate system based on RBF neural network model

• Published in: Construction & building materials Vol. 357; p. 129382
• Main Authors: Sun, Mengkai, Li, Bo, Yi, Zhao, Cao, Kai, Li, Andi, Wang, Yongning
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 28.11.2022
• Subjects: Adhesion and cohesion; Asphalt binder-aggregate system; Moisture stability; Optimization; RBF neural network model; Surface free energy (SFE); Surface free energy (SFE); Moisture stability; Adhesion and cohesion; Optimization; Asphalt binder-aggregate system; RBF neural network model
• ISSN: 0950-0618; 1879-0526
• DOI: 10.1016/j.conbuildmat.2022.129382

•The SFE parameters of the asphalt binder-aggregate system are optimized based on the RBF neural network model.•The Lewis base component and Lifshitz nonpolar component of aggregates played a dominant role in the asphalt mixes’ resistance to water damage.•The optimized SFE parameters of asphalt binder-aggregate system are reasonable to replace the water stability test of asphalt mixture. Asphalt pavement is subject to water damage frequently, which reduces the level of service that it provides and contributes to many road diseases. The surface free energy (SFE) theory has proven to be a trustworthy approach to assess water damage in asphalt mixtures. The goal of this investigation was to optimize the SFE parameters of asphalt binder-aggregate systems using an RBF neural network model. In this research, the SFE parameters of asphalt binder-aggregate systems were calculated firstly. Then, the modified boiling water test and the immersion Marshall test, as well as the freeze–thaw splitting test were then used to assess the asphalt mixtures’ adhesive properties and water stability. Thereafter, the comprehensive analysis method of entropy weight was used to conduct a thorough evaluation of the asphalt mixtures’ resistance to water damage. Next, the comprehensive evaluation index, Wi, of the asphalt mixtures’ resistance to water damage was used as the output layer of the RBF neural network model to optimize the SFE parameters. Finally, the feasibility of the optimized the SFE parameters using RBF neural network model was verified by establishing the connection between the optimized SFE parameters and the asphalt mixtures’ resistance to water damage by the entropy weight comprehensive analysis method. The results showed that among the SFE parameters of asphalt binder-aggregate systems optimized using the RBF neural network model, the Lewis base component and the Lifshitz nonpolar component of the aggregate played a dominant role in the ability to resist water damage, followed by the VDW adhesion work and the VDW de-bonding work formed by van der Waals forces. Furthermore, the asphalt binder-aggregate system composed of limestone had the best adhesion properties among all aggregates.