Asphalt Concrete: Simulation, Modeling, and Experimental Characterization

This paper summarizes the findings of modeling the micromechanical behavior of asphalt mastics and asphalt mixtures under various loading conditions. A commercial discrete element code called Particle Flow Code in 2-Dimensions (PFC2D) is used for this purpose. Asphalt mastics are simulated using an assembly of stiff particles randomly dispersed in a medium of soft particles, representing the aggregate fillers and the asphalt binder, respectively. The stiffening effect of the aggregate fillers on the micromechanical behavior of asphalt mastics is investigated at different filler volume fractions. These results are compared to Dynamic Shear Rheometer (DSR) measurements on actual mastics. These mastic models are used to simulate the micromechanical behavior of hot mix asphalt (HMA) concretes. The behavior of these HMA models was investigated at high and low temperatures under loading conditions similar to those applied in the Simple Performance Test (SPT) and the Indirect Tension Test (IDT), respectively.

This paper presents a finite element (FE) micromechanical model for viscoelastoplastic behavior of asphalt mixtures. Asphalt mixture is composed of highly irregular aggregates, mastic (asphalt plus fine aggregate and fines), and air voids. In this paper, the microstructural model of particulate asphalt materials incorporates an equivalent lattice network structure whereby intergranular load transfer is simulated through an effective asphalt mastic zone. The finite element model integrates the ABAQUS user material subroutine with continuum elements for the effective asphalt mastic and rigid body elements for each aggregate. A FE incremental algorithm with a recursive relationship for three-dimensional (3D) viscoelastic behavior was developed. Chaboche's plastic model was applied, and the constitutive equations were solved using a predictor-corrector scheme. These algorithms were defined in a 3D user material model for the asphalt mastic to predict global rate-independent permanent deformation of asphalt materials. The effect of loading rates on the material viscoelastic and viscoelasto-plastic behavior was investigated using FE numerical simulations for an ideal asphalt mixture specimen.

Many researchers recognized that micromechanical models have tremendous potential in the field of asphalt technology, for reducing or eliminating costly tests to characterize asphalt-aggregate mixtures for the design and control of flexible pavement structures and materials. The objective of this study is to develop micromechanical based finite element (FE) model to capture the microstructure of asphalt mixture and to predict mixture properties. In this approach, various material phases (aggregates and mastic) are modeled with a number of fine finite elements. Aggregate geometry, shape, orientation, and gradation are considered in the modeling. Furthermore, high-resolution images are used to study the microstructure of asphalt mixture and to prepare geometry input for the FE model. In addition, the complex modulus (E*) of the sand mastic (asphalt plus fine aggregate) are measured by an experimental program and used to compare the prediction of FE model. The E* of the asphalt mixture are measured and used to compare the prediction of FE model. The developed FE approach has the ability to predict asphalt mixture complex moduli in compression across a range of loading frequencies at low and intermediate temperatures (–20, –10, and 0°C).

Simple Performance Tests (SPT) including Indirect Tensile (IDT) Test and Dynamic Modulus Test have been widely used in the evaluation of the performance of asphalt concrete. The so-called SPT tests typically apply uniform stresses on the boundary and therefore obtain the stress-strain relation with convenience. Nevertheless, asphalt concrete is a heterogeneous material composed of asphalt binder, aggregates and air voids. The three constituents have drastically different stiffness. Even under a uniform boundary stress, the internal stress and strain distributions are not uniform. This paper presents a comparison between the stress distribution based on heterogeneous material properties and that based on homogeneous material properties using X-ray Computed Tomography (XCT) and Finite Element (FE) simulation. The comparison indicates that material heterogeneity is an important factor that must be considered in the characterization of asphalt concrete.

Permanent deformation is one of the primary distresses in hot mix asphalt (HMA). It is influenced by the properties of the mix constituents including microstructural features such as aggregate directional distribution (anisotropy), and damage. Conventional continuum constitutive models do not explicitly consider the influence of microstructure distribution on the macroscopic response. The main objective of this paper is to develop an elasto-visco-plastic continuum model that accounts for aggregate directional distribution and damage. This is accomplished by modifying the Drucker-Prager yield function to account for these microstructural features. This yield function is incorporated into Perzyna's viscoplastic flow function. The model is implemented in finite element (FE) using an implicit numerical integration algorithm. FE analysis is conducted to study the influence of key factors such as anisotropy and damage on the model response. In addition, the FE model is used to simulate the response of HMA to compressive and extension triaxial tests at different loading rates and confining pressures, and the numerical results are compared to experimental measurements.

When traffic load is imposed on a pavement, a non-uniform displacement field develops giving rise to a multitude of triaxial states of stress. Triaxiality has been known to significantly influence the response of asphaltic materials. In this contribution the extension of the Desai plasticity based material model for reversed cyclic loading is presented. Hardening of the material is simulated via a mixed hardening formulation. In order to enable a smooth elastic-plastic transition, the bounding surface concept proposed by Dafalias and Popov is utilized.

The algorithmic aspects and the numerical implementation of a new multi-dimensional hyperelastic-viscoplastic-damage model for asphalt concrete are presented. A predictor-corrector algorithm is developed, the algorithmic moduli are computed and the model is implemented into the finite element environment ABAQUS® and is used for the simulation of several experimental results as well as for the analysis of a section of a pavement. The proposed strain energy function is expressed in terms of the invariants of the deviatoric left Cauchy-Green strain tensor as well as of the volume ratio J. A unique feature of the paper is that experiments are treated both as homogeneous (local) and as boundary value problems (global) and the model's simulations for each case are compared.

The applicability of the Huet-Sayegh (H&S) model is shown for different asphalt mixes. The H&S model with only six parameters simulates in an excellent way the behavior of asphalt mixes in cycling tests over a very wide range of frequencies. Examples are given of these perfect fits for three completely different asphalt mixes. The only disadvantage is that the original model does not contain a viscous element for simulating the permanent deformation in contrast with the more familiar Burger's model. However, by adopting a linear dashpot in series with the H&S model an attractive alternative for the Burger's model is obtained which can be used over a large frequency range. Special attention is given to the mathematical operations connected to the use of the H&S model.

Earlier findings in fatigue tests using a repeated sequence of a fixed number of cycles with a high deflection and a fixed number of cycles with a small deflection confirms the supposition that healing of the stiffness already occurs during the fatigue test and does not depend on a rest period of no loading at all. It was shown by theory and in practice that the decrease in temperature due to the decrease in dissipated energy per cycle when smaller strains were applied could not explain the amount of increase in the stiffness modulus. Further investigation of these findings and results has lead to a new material response model that describes both the evolution of the modulus and the phase lag of the complex stiffness model during fatigue testing. It should be marked that healing in this paper is only related to the complex stiffness modulus. This partial healing (PH) model is a material model for which the response in a point depends on the stress/strain state in that point (the rate of dissipated energy). Therefore the application of the PH model for the determination of the response (evolution of stiffness) of a finite specimen in a test will depend on the geometry and loading configuration of the specimen and the load mode. The PH model is applied on four point bending beam tests (4PB) in controlled deflection mode. The obtained parameters for the PH material model were used for the prediction of the stiffness evolution in uni-axial push-pull fatigue tests (UPP) on the same asphalt mix. A fair comparison was observed between the measured and predicted evolution of both the modulus and phase lag of the complex stiffness modulus in the UPP tests.

Asphalt is a complex civil engineering material. Asphalt when mixed with granular aggregates exhibits wide range of responses. "Healing" is one such mechanical response wherein if the material is allowed to rest after certain number of load applications, the material exhibits a "beneficial internal structure change". In this investigation, repeated triaxial tests were carried out on sand asphalt mixtures with varied confinement conditions and rest periods to quantify healing of asphalt mixtures in the laboratory. Two parameters were identified to quantify healing. The first parameter corresponds to the percentage change in instantaneous deformation before and after rest period. The second parameter corresponds to the percentage change in strain recovery at a fixed time period after unloading. The influence of various testing conditions on these two parameters is investigated.

Laboratory fatigue characterization of HMAC mixtures constitutes a fundamental component of pavement design and analysis to ensure adequate performance. In this study, the traditional mechanistic empirical (ME) approach and a continuum micromechanics based calibrated mechanistic approach with surface energy (CMSE) measurements were comparatively utilized to characterize the fatigue resistance of two HMAC mixtures in the laboratory, including investigating the effects of aging. Although the results were comparable, the CMSE approach exhibited greater flexibility and potential to discretely account for most of the fundamental material properties (including fracture, aging, healing, visco-elasticity, anisotropy, crack initiation, and crack propagation) that affect HMAC pavement fatigue performance. Compared to the mechanistic-empirically based ME approach, the CMSE approach is based on the fundamental concepts of continuum micromechanics and energy theory; and utilizes the visco-elastic correspondence principle, Paris' Law of fracture mechanics, and Schapery's work potential theory to monitor cumulative fracture damage in HMAC mixtures, measured in terms of dissipated pseudo strain energy (DPSE) under repeated uniaxial tensile tests. Additionally, the CMSE results exhibited relatively lower statistical variability. For the materials and test conditions considered in the study, aging reduced HMAC mixture fatigue resistance and its ability to heal. Thus aging plays a significant role in HMAC mixture fatigue performance and should be incorporated in fatigue design and analysis.

Thermal cracking results in both structural and functional problems in asphalt concrete (AC) pavements. A critical parameter affecting the buildup of thermal stresses that ultimately lead to cracking in AC pavements is the coefficient of thermal contraction (CTC) of the AC mixture. The objective of this study is to assess the sensitivity of CTC of AC mixture on thermal cracking. Values of CTC of AC mixtures were identified through literature review, while other relevant data needed for thermal cracking prediction were obtained for three AC mixtures from the new Mechanistic-Empirical Pavement Design Guide (MEPDG) documentation. Finally, the amount of thermal cracking, in terms of length and percent damage, was predicted using the new (MEPDG) software. Results showed that the sensitivity of CTC of AC mixtures on thermal cracking is highly material dependent. The greatest sensitivity was observed with medium strength and ductile AC mixture.

The research presented in this paper developed a test protocol for using the flexural beam fatigue test to evaluate the moisture sensitivity of hot mix asphalt (HMA). The following test parameters were determined: vacuum intensity and duration for pre-saturating specimens, preconditioning temperature and duration before the fatigue test, and temperature and strain level during the fatigue test. An experiment was designed to include two aggregates, two binders, and one antistripping additive for testing by this procedure. The indirect tensile strength ratio (TSR) test and the Hamburg Wheel Tracking Device (HWTD) test were also performed on the same mixes. Generally, the three test methods did not give consistent results in terms of relative ranking of performance. The flexural beam fatigue test showed potential capability to evaluate moisture sensitivity of HMA, but further refinement is needed.

In this paper, the responses of two different pavement structures (10 and 20 cm HMA layer), were analyzed under the steering, driving and trailer axles of an eighteen-wheeler truck during braking on a 6% downhill grade. The response of the asphalt pavement to the decelerated truck during the braking period was estimated at three different traveling speeds: 64, 32, and 3.2 km/h using the computer code 3D-Moving Load Analysis (3D-MOVE). Significant load redistribution, which occurs between the truck axles under deceleration, has been accounted for in the analysis. The time dependent behavior of the HMA layer as the truck approaches the stopping point is incorporated by using the complex shear modulus and the internal damping as a function of loading frequency. The base course and subgrade layers are treated as linear elastic materials with an internal damping assumed to be 5%. The non-uniform stress distributions at the tire-pavement interface were interpolated from measured contact stress distributions at various speeds made with the Kistler MODULAS Quartz Sensor Array by the Nevada Automotive Test Center (NATC). Braking forces at each tire were included as interface shear stresses with a distribution that was estimated by multiplying the vertical stress distribution by the calculated coefficient of friction between each tire and the pavement surface. The study reveals that rutting in the HMA layer is more prone under the steering single tire while shoving is mainly caused by the dual tandems driving tires.