Modeling the 3D fracture-associated behavior of viscoelastic asphalt mixtures using 2D microstructures

摘要

Asphalt concrete is a highly heterogeneous mixture with complicated microstructures. This heterogeneity strongly affects the overall three-dimensional (3D) mechanical behavior and damage-associated performance of asphaltic mixtures, which makes accurate modeling a challenge. This paper presents a computational microstructure model using multiple two-dimensional (2D) finite element microstructures as an alternative and efficient approach to predict the actual 3D fracture-associated response of asphaltic mixtures. To simulate crack initiation and propagation, an autonomous algorithm was developed to efficiently generate multiple 2D microstructures using image processing of scanned microstructures, a phase-based segmentation module to separate particles and the surrounding matrix phase, and a finite element meshing module that allows cohesive zone elements to be embedded within the mixture microstructure. Aggregates were considered elastic, but viscoelastic and fracture properties were used to model the binding matrix phase. The validity of the multiple 2D microstructures approach was statistically investigated by comparing the model simulation results with the corresponding experimental results from two cases: (1) a three-point bending beam testing of a gap-graded mixture: and (2) a semicircular bending beam testing of a conventional dense-graded mixture. The 2D simulation results of multiple microstructures could generally capture 3D viscoelastic fracture behavior, but the prediction power of the modeling was reduced when volume fraction, distribution of coarse aggregates became high and complex. With some limitations to be further resolved, this study implies that multiple 2D microstructures can appropriately represent the complex viscoelastic-fracture behavior of 3D mixtures, which can significantly reduce the experimental and computational costs for laboratory mixture tests and 3D microstructure simulations with fracture, respectively. (c) 2017 Elsevier Ltd. All rights reserved.