Data-driven model for Fourier transform infrared spectrum characteristics of the aged modified bio-asphalt binder. (October 2022)

• Record Type: Journal Article
• Authors: Ren, Jiaolong; Zhang, Xiangquan; Zhao, Hongbo; Zhang, Lin; Yang, Jiale; Tang, Gangting; Chen, Jingchun; Wang, Jian; Li, Min
• Is Part Of: Measurement. Volume 202(2022)
• Publication Date: 2022-10
• Subjects: Modified bio-asphalt binder -- Ageing behaviour -- Fourier transform infrared -- Ensemble learning -- Reduced order model; Weights and measures -- Periodicals; Measurement -- Periodicals; Measurement; Weights and measures; Periodicals; 530.8
• DOI: 10.1016/j.measurement.2022.111879
• Languages: English
• ISSNs: 0263-2241

Abstract

Highlights: Develop a novel data-driven framework by combing ROM model and ensemble learning. Built the ROM model to capture the characteristics of FTIR spectrum. Improve the understanding of Fourier transform infrared spectrum characteristics. Verify and illustrate the developed framework using the aged modified bio-asphalt asphalt binder. Abstract: In this study, a novel data-driven framework was developed to capture the characteristics of FTIR spectrum based on the reduced-order model (ROM) and ensemble learning, owing to the advantages of the ROM in high computational efficiency and low dependence of data volume. The vector base and the corresponding coefficients of the ROM were determined based on the proper orthogonal decomposition (POD). The ensemble learning was used to approximate the relationship between the FTIR features of the modified bio-asphalt binder and the POD coefficients and determine the ROM which was established based on the FTIR data (snapshots). The developed framework was applied to predict the whole FTIR spectrum of the aged modified bio-asphalt asphalt binder. The results show that the developed framework provides a high-precision and high-efficiency tool for predicting the FTIR spectrum of asphalt-based materials.