Assessment of existing micro-mechanical models for asphalt mastic considering inter-particle and physico-chemical interaction. (20th November 2019)

• Record Type: Journal Article
• Authors: Ma, Xiaoyan; Chen, Huaxin; Yang, Pingwen; Xing, Mingliang; Niu, Dongyu; Wu, Shujuan
• Is Part Of: Construction & building materials. Volume 225(2019)
• Page Start: 649
• Page End: 660
• Publication Date: 2019-11-20
• Subjects: Asphalt mastic -- Micromechanical model -- Physico-chemical interaction -- Particle interaction; Building materials -- Periodicals; 624.18
• DOI: 10.1016/j.conbuildmat.2019.07.227
• Languages: English
• ISSNs: 0950-0618

Abstract

Highlights: Three micromechanical models was selected to predict the asphalt mastic behavior. GSCG gave the most accuracy modulus prediction of mastic with low filler fraction. J-C model cannot predict modulus of mastic with high filler fraction. Phy-C gave a good modulus prediction for mastic with moderate filler fraction. Prediction accuracy deteriorated severely with the increase of filler fraction. Abstract: Micromechanical models have been used since the 1990s to predict the properties of asphalt mastic. However, most of these models are found to be unsatisfactory in predicting the mastic's properties because the models were derived from the research of particle-filled composites that did not take into account the asphalt-filler physico-chemical interactions and particle interactions. In this paper, three micro-mechanical models for predicting the complex shear modulus master curve of asphalt mastic are evaluated: the generalized self-consistent scheme model, the four-phase micro-mechanical model, and the particle interaction model. All three micro-mechanical models are based on the mechanical properties of their constituent materials as well as the filler-asphalt physiochemical interactions and the particle interactions. Two virgin asphalt binders and two polymer modified asphalt binders were selected to fabricate 16 mastics of four different filler volume fractions. The accuracy of prediction was evaluated by comparing the relative differences between the experimental complex shear modulus master curves and that predicted by the models. The results suggest that (1) the generalized self-consistent scheme model have a satisfactory prediction at a low filler volume fraction, but their accuracy is significantly affected by frequency; (2) the particle interaction model cannot calculate the complex shear modulus of mastic with a filler volume fraction of more than 0.53; (3) the four-phase model shows a precise forecast of complex shear modulus for mastic with moderate and high filler volume fractions; and (4) for all of the three models, the deviations increase severely with the increase of the filler volume fraction.