| In order to reasonably simulate the uniaxial compression dynamic modulus test of asphalt mixture,based on the morphological characteristics of aggregate,a meso virtual specimen of asphalt mixture including aggregates,asphalt mortars,ITZs and voids,whose simulation results are compared with the experimental data,is established.The results show that the dynamic moduli obtained from the simulation experiment are in good agreement with the experimental data.In order to study the influence of aggregate randomness on dynamic modulus of asphalt mixture,100 mesoscopic virtual specimens including random aggregate libraries in which aggregate profiles are reconstructed by morphology are generated by the random packing algorithm.The dynamic modulus master curves of asphalt mortar are constructed at different reference temperatures.The surface to surface contact pair model of ITZ uses the Coulomb friction.Based on the virtual experimental data under different temperatures and loading frequencies,values of the virtual dynamic modulus obtained by each group of temperature and frequency combination are approximately the normal distribution.Moreover,the randomness of aggregate morphology at both lower temperature and higher frequency can cause obvious dispersion of dynamic modulus.The experimental data of different literature show that the prediction master curve method of dynamic modulus for asphalt mixture in this paper is reasonable.For the integral kernel functions with the continuous relaxation spectrum of the storage modulus,the loss modulus,the dynamic modulus and the relaxation modulus,four numerical integration methods are respectively selected to obtain integral values.Based on five groups of experimental data of asphalt mixture,effects of the WLF and the Arrhenius shift factors to the 2S2P1D model are compared.The results show that the loss modulus has a certain influence on the dynamic modulus,and parameters of the continuous relaxation spectrum,i.e.H(τ),should be obtained based on the experimental data of dynamic modulus.By using the self-adaptive Simpson formula and tolerance of 1,numbers of integral intervals and calculation accuracies can be both considered.Parameter values of the 2S2P1D model obtained from experimental data of dynamic modulus can well deduce material parameters of the storage modulus and the loss modulus,but calculated values of the loss modulus show some slight discreteness.The Arrhenius formula is more suitable for analyzing dynamic modulus of asphalt mixture,while WLF formula is more suitable for analyzing the loss modulus of asphalt mortar.The 2S2P1D model can be used to fit a master curve of the dynamic modulus,whose parameters can be obviously effected by initial values and fitting ranges of the parameters.In order to obtain stable fitting solutions,parameter sensitivity analysis and classification of the 2S2P1D model are carried out.Based on the sensitivity classification results,experimental data of dynamic modulus of asphalt mortar and asphalt mixture are used to obtain values of the fixable parameters.The results show that E0,β,K and h of asphalt mortar can be 0,1247,0.233 and 0.655;E0,β,K,h andτref of asphalt mixture can be 182MPa,2258,0.189,0.530 and 0.053.If need,only E∞of the 2S2P1D model should be recognized to describe material properties of asphalt mortar and asphalt mixture,whereδandτref of the former can be 4.818 and 6.61×10-4andδof the latter can be 4.818.This paper analyzes the feasibility of substituting part of the laboratory test with simulation experiment on the dynamic modulus of asphalt mixture under uniaxial compression and gives suggestions on how to use the 2S2P1D model to establish a reasonable master curve of the dynamic modulus. |