Abstracts for the March 27-29, 2006 Annual Meeting


by Terhi K. Pellinen and Shangzhi Xiao


This research was initiated from a need to assess marginal mixtures that do not meet the Superpave mix design criteria for the air voids content requirements for the construction of hot- mix asphalt pavements. The developed criteria is comprised of the base criteria derived from test data of as-placed field cores normalized to 7.5% air voids content, which provides the minimum stiffness and strength requirements for a mixture as a function of the traffic level. The stiffness of pavement cores is measured using the Simple Shear Tester to obtain the shear stiffness |G*| of the mix, and the strength is measured using indirect tensile strength testing. The asphalt mixture will meet the criteria if it meets both the stiffness and strength requirements. Additionally, the criteria include coefficients intended to adjust the base criteria for variable mix density. Given the need to assess the mix performance in the plant-mixing stage, an adjustment factor is provided to account for the compaction procedure effects for the gyratory-compacted asphalt plant mixtures relative to pavement cores compacted by rollers. Forensic study of failed interstate pavement near Indianapolis, Indiana suggests that for heavy traffic of more than 30 million design equivalent single axle load applications, the factor of safety that the developed criteria is providing against rutting is ranging from 2.2 to 2.8.


by Graham C. Hurley and Brian D. Prowell


Several new processes have been developed to reduce the mixing and compaction temperatures of hot mix asphalt without sacrificing the quality of the resulting pavement. Three potential warm mix asphalt processes were evaluated in this study. They were Aspha-min , Sasobit , and Evotherm . A laboratory study was conducted to determine the applicability of these processes to typical paving operations and environmental conditions commonly found in the United States, including the performance of the mixes in quick traffic turn-over situations and high temperature conditions. Superpave gyratory compactor results indicated that Aspha- min , Sasobit , and Evotherm may lower the optimum asphalt content, so they should be added during the mix design process.
All three processes were shown to improve the compactability of mixtures in both the SGC and vibratory compactor. Statistics indicated an overall reduction in air voids. Improved compaction was noted at temperatures as low as 190°F (88 C). The addition of Aspha-min , Sasobit , or Evotherm did not affect the resilient modulus of an asphalt mix nor did they increase the rutting potential of an asphalt mix as measured by the Asphalt Pavement Analyzer. The rutting potential did increase with decreasing mixing and compaction temperatures, which may be related to the decreased aging of the binder resulting from the lower temperatures. The lower compaction temperature used when producing warm asphalt may increase the potential for moisture damage. Overall, Aspha-min , Sasobit , and Evotherm appear to be viable tools for reducing mixing and compaction temperatures that can be readily added to hot mix asphalt. Reductions in mixing and compaction temperatures are expected to reduce fuel costs, reduce emissions, widen the winter paving window and facilitate niche applications such as airport runway construction where rapid open to traffic is essential.


by Khalid Al Shamsi, Louay N. Mohammad, Ph.D., Zhong Wu, Ph.D., P.E., Sam Cooper, P.E., and Chris Abadie, P.E.


This study analyzes the compaction and performance characteristics of asphalt concrete mixtures with aggregate structures that were designed using an analytical method of aggregate blending. Three aggregate types were considered in this study: limestone, sandstone, and granite. All the aggregates were crushed aggregates. Three different aggregate structures were designed for each aggregate type using the Bailey method of aggregate gradation evaluation. The Bailey method is a comprehensive gradation evaluation procedure that provides aggregate interlock as the backbone for the aggregate skeleton. The aggregate structures were also characterized using the power-law method of aggregate evaluation. All the aggregate structures had a nominal maximum aggregate size (NMAS) of 12.5mm, and were designed for high traffic level. The Binder type selected was PG 76-22. During compaction, force measurement was made using the pressure distribution analyzer (PDA). The compaction characteristics of the mixtures were analyzed using data from the PDA and the traditional Superpave Gyratory Compactor (SGC) results. Simulative (Hamburg Wheel Tracking Test) and fundamental (Indirect Tensile Strength Test, and Semi-Circular Fracture Test) tests were conducted to determine laboratory performance properties and evaluate the mixtures under different loading and environmental conditions. The effect of aging on the designed mixtures was also investigated. The compaction indices and gradation parameters from the Bailey and the power-law methods were correlated with the results from the laboratory tests. The data indicate that the designed mixtures considered were impermeable and presented superior resistance to rutting and moisture damage. The compaction indices were sensitive to the change in gradation parameters from both methods of aggregate evaluation. The results from the mechanistic tests are sensitive to the gradation parameters from both gradation evaluation methods, but not to the energy indices from the SGC and the PDA.


by Ignacio Artamendi and Hussain A Khalid


Research has been conducted on the properties of paving binders modified with waste tyre rubber in the wet process. The absorption of bitumen from different sources and grades into waste truck and car-tyre rubber has been studied by weight gain experiments. Rubber monoliths cut form waste tyre threads have been immersed in hot bitumen and the increased mass of the rubber specimens at different immersion times has been used to determine the sorption curves. The kinetics of bitumen absorption by the rubbers have been analysed in terms of the simple Fickian model, which is characteristic of absorption of solvents by rubbery polymers. Two penetration grade bitumens, 50 and 100 Pen, each from two different sources, have been used in the study. Results show that both equilibrium swelling weight increase and diffusion coefficient are related to the grade and source of the binders as well as the type of rubber. Furthermore, truck-tyre rubber absorbs more bitumen than car-tyre rubber does. Equilibrium swelling and diffusion coefficient increase when the grade of the binder used is increased, thus, when the viscosity decreases. For the same penetration grade, binders with high asphaltene content swell the rubbers markedly less and at lower rate than those with lower asphaltene content. The effect of temperature on the sorption parameters has also been investigated. Results indicate that as the temperature increases, the diffusion coefficients for the two rubbers also increase. Furthermore, the temperature dependence of the diffusion coefficient has been observed to follow an Arrhenius relationship characteristic of an activated process. It has been found that the activation energy for car-tyre rubber is greater than that of truck-tyre rubber. Moreover, diffusion coefficients determined for rubber monoliths assuming plane-sheet geometry have been used to estimate the kinetics of diffusion into crumb rubber particles of idealised spherical geometry.


by Hilde Soenen, Joelle De Visscher, Tine Tanghe, Ann Vanelstraete, and Per Redelius


There is a need for true performance indicators for bituminous binders, which are equally applicable independent of whether the binder is modified or not. The purpose of this research is to evaluate performance indicators for permanent deformation both for modified and unmodified binders, and to select those indicators that relate best to permanent deformation in laboratory asphalt mix tests. The binder indicators considered are mainly rheological parameters, including the SHRP rutting parameter, the zero shear viscosity and the recently proposed repeated creep test. To evaluate the permanent deformation or rutting susceptibility of the corresponding mixes two devices were used: the French rut tester (large size device, according to EN12697-22) and the cyclic triaxial compression test, according to EN12697-25, part B. This study shows that rheological and even conventional properties of modified binders can be extremely dependent on the thermal history and preparation conditions prior to testing. A very strict control of these parameters is absolutely necessary in order to be able to determine repeatable and reproducible binder performance indicators. The thermal history effects are related to variations in crystallinity or to variations in micro-morphology. For unmodified binders, any of the performance indicators can be used to relate binder properties to their performance in asphalt mix tests. For polymer modified binders the behavior is more complex, only well selected parameters relate to the behavior observed in the asphalt mix tests, and only if the binders are tested after appropriate thermal preconditions. A detailed discussion is included on the relations between binder parameters and the respective mix tests, and on the importance of thermal history in these relations. Sample preparation recommendations for binder performance testing are also proposed.


by R. Michael Anderson and Michael Huner


In 2002, the Kentucky Transportation Cabinet (KTC) began evaluations of asphalt binders supplied in Kentucky to determine the efficacy of using the AASHTO MP-1a specification as an alternate means of specifying the low temperature grade of the asphalt binder. In the AASHTO MP-1a specification, the critical cracking temperature (CCT) is determined and compared to the required climatic grade. The KTC asked suppliers to provide monthly samples for testing for each grade supplied to Kentucky. Suppliers were asked to determine the CCT of their binders so that a direct comparison could be made between the KTC lab and the supplier's lab. Four asphalt binder grades were provided for this study by five different producers. The four asphalt binder grades were: PG 58-22, PG 64-22, PG 70-22, and PG 76-22.
Thirty-eight asphalt binder samples were evaluated in the study. In approximately 25% of the cases for each lab, the data indicated that the low temperature grade will change when using the CCT instead of the M320 grading. Considering the data from both labs, approximately 10% of all comparisons indicate a warmer grade than expected (i.e., -16 instead of -22) and approximately 10% of all comparisons indicate a colder grade than expected (i.e., -28 instead of -22).
The data also indicated that the average CCT of the PG 76-22 asphalt binders was lower than the average of the PG 58-22 asphalt binders, and considerably lower than the average of the PG 64-22 and PG 70-22 asphalt binders.
Before implementing the AASHTO MP-1a specification, users should consider testing issues with the bending beam rheometer (BBR) and direct tension tester (DTT). Quality data checks are important to achieving an accurate, repeatable CCT value for an asphalt binder.


by Dr. David A. Anderson, P.E., Vikas V. Thakar, and Dr. Charles E. Antle


The Superpave Performance-Graded binder specification is based upon material properties that are related to field performance. The test results for these material properties are intended to be used for acceptance and payment purposes. The conformance or non conformance of the material to the specification is decided through the procedure documented in the quality assurance plan that forms a part of the contract agreement for the construction of the pavement. The quality assurance plan is a component of the overall quality management system that needs to be enforced in order to achieve the intended performance of the pavement. This paper presents the essential components of a quality management system for Performance Graded (PG) binders, and a payment protocol developed to a stage that it can be incorporated into the specifications for the PG binders.
The overall process of binder acceptance and payment is very complicated due to the potential involvement of a number of different parties including the producer, brokers, terminals and shippers, the supplier (often the HMAC contractor), and the user agency. Acceptance of the asphalt binder is dependent upon the test results from the supplier and user. The potential effects of the variability and bias in these test results must be carefully considered in the acceptance and payment protocol. These considerations led the authors to the development of a statistically based acceptance and payment plan which considers laboratory specific testing variability and bias. This is a different approach than the traditional percent within limit (PWL) method, which requires a large number of samples and testing.
The proposed scheme is intended to be reasonable to all the parties by balancing their individual risks. In the future, this payment schedule, contingent upon field verification, can be integrated with performance prediction models as they are developed and verified. In this way, payment for the asphalt binder can be related to the expected performance of the pavement.


by Ioan Negulescu, Louay Mohammad, William Daly, Chris Abadie, Rafael Cueto, Codrin Daranga, and Ionela Glover


An SBS copolymer modified asphalt cement (PG 76-22) was subjected to accelerated aging using multiple PAV operations in the absence and in the presence of water. A similar polymer modified asphalt cement composition was sampled from a road after up to seven years of service. The extent of oxidation and changes in the molecular mass of the asphalt cement components of aged samples were estimated from FTIR and GPC analyses. Dynamic viscoelastic properties were determined using a high torque instrument. A characteristic point of interest was the temperature at which tand became unitary at a frequency of 10rad/s. Oxidative aging in the presence of water promoted an increase in the carbonyl content of aged samples, primarily as acid groups, but the high humidity aging reduced the extent of asphalt hardening. A correlation between laboratory and field results was obtained.


by Christopher Robinette and R. Christopher Williams


The Superpave Simple Performance Test (SPT) has been gaining use within the asphalt pavement research community. Testing has been conducted to better understand the response of materials to the proposed testing scheme. This paper presents the affects of testing history on the materials response with particular emphasis on the dynamic modulus and flow number tests. Also examined is the method of specimen preparation in terms of sawed/cored or compacted test geometries. This work has shown that the testing history of the specimen does not affect the mean response, but does tend to increase the variability. Also shown is that the method of specimen preparation does not affect the materials response to loading and this finding is confirmed through pavement design with the aid of the 2002 Design Guide software.


by Nam H. Tran and Kevin D. Hall


Recent published studies clearly emphasize the importance of dynamic modulus (|E*|) testing of hot-mix asphalt concrete (HMA). The dynamic modulus is one of the fundamental inputs in the mechanistic-empirical (M-E) Design Guide developed through National Cooperative Highway Research Program (NCHRP) Project 1-37A. The current AASHTO TP62- 03 specifies that the dynamic stress be adjusted to obtain: (1) axial recoverable strains between 50 and 150 microstrain; and (2) cumulative permanent strain of less than 1,500 microstrain at the end of the tests at each temperature level. The requirement is to maintain the test specimen behavior in its linear viscoelastic range during the |E*| testing regime. However, it is suspected that axial dynamic strains between 100 and 150 microstrain may produce cumulative permanent strain levels which affect the linear viscoelastic response of the material, which in turn affects the |E*| results that influence predicted pavement performance. In order to investigate the effect of the dynamic strains, a study of four |E*| tests was conducted. Two |E*| tests were performed on a set of three replicate specimens using axial dynamic strains between 50 and 100 microstrain. Another two |E*| tests were performed on a second set of three replicate specimens using axial dynamic strains between 100 and 150 microstrain. The two sets of test specimens were fabricated using the same HMA mixture. The results showed that the |E*| determined using the two dynamic strain levels were significantly different at high temperatures (38 and 54C). Based on the study results, it is recommended that axial dynamic strain levels used in E* testing be re-evaluated and possibly reduced to account for possible viscoelastic response effects.


by Mansour Solaimanian, David Fedor, Ramon Bonaquist, Ali Soltani, and Vivek Tandon


Developing a highly reliable laboratory test to predict the susceptibility of hot mix asphalt (HMA) concrete to moisture damage remains a challenge to asphalt industry. Pennsylvania State University in cooperation with Advanced Asphalt Technologies, LLC, University of Texas at El Paso, and PaveTex Engineering and Testing, Inc. conducted an extensive research study to develop an improved testing procedure for prediction of moisture damage in HMA. The research, sponsored by National Cooperative Highway Research Program (NCHRP,) was conducted in two phases and included eleven different mixes from various states with known field performance. The simple performance tests, developed by NCHRP projects 9-19 and 9-29, were utilized with the Environmental Conditioning System (ECS) and the combined system was used to test moisture damage sensitivity of these mixes.
Phase I included the design and execution of an experiment where unconditioned and ECS conditioned specimens were tested using the three Project 9-19 simple performance tests: flow time, flow number, and dynamic modulus. Tests were conducted on three mixtures made from aggregates with good, marginal, and poor resistance to moisture damage based on past research. In addition to the combined ECS/simple performance tests, ASTM D4867 and the Hamburg Wheel Tracking Device were included in the Phase I experiment. The primary conclusion from the Phase I experiment was that the dynamic modulus test was the most suited of the three simple performance tests for possible use with the ECS in an improved moisture sensitivity test.
Phase II included eight mixtures with known field performance in regard to moisture damage. Three were good performing mixes and five were poor performing. These mixes were tested with the combined ECS/dynamic modulus system. The mixes were also tested according to ASTM D4867 and the Hamburg Wheel Tracking Device. The results indicated that the dynamic modulus test utilized with the ECS has the potential to discriminate between good and poor performing mixes in regard to stripping and moisture damage.


by Javed Bari and Matthew W. Witczak


The main purpose of this paper is to present the development of a new revised version of the widely known Witczak E* Predictive Model. The model development was aimed at overcoming the limitations of current models available to the pavement community, which are used for predicting the dynamic modulus (E*) of hot mix asphalt mixtures (HMA). A comprehensive study was completed at Arizona State University to conduct numerous E* testing and finalize a huge E* database, containing 7400 data points from 346 HMA mixtures. This database was used to develop the new E* predictive model. The model is capable of accurately estimating changes in E* of HMA mixture as a function of changes in mixture volumetrics, material properties, temperature and loading frequency (or time) for the entire E* database used. The model has been found to be rational, unbiased, accurate, and statistically sound. The new mechanistic-empirical pavement design guide entitled "Guide for Mechanistic-Empirical Design of New and Rehabilitated Pavement Structures" developed under National Cooperative Highway Research Program (NCHRP) Project 1-37A uses the current version of Witczak E* Predictive Model in its input levels 2 and 3. It is hypothesized that due to its similar sigmoidal structure as used in the guide, the newly developed E* model can be easily incorporated in a future revision of this pavement design guide.


by Syed Waqar Haider and Karim Chatti


The research described herein was conducted as a part of NCHRP Project 20-50 (10/16), LTPP Data Analysis: Influence of Design and Construction Features on the Response and Performance of New Flexible and Rigid Pavements. While there is considerable research on factors affecting flexible pavement rutting, they are limited in scope because of lack of field validation and no consideration for interactions of structural and site factors. This research addresses the relative influence of design and site factors on structural rutting of in-service flexible pavements. The data from the SPS-1 experiment of the LTPP program were used. This experiment was designed to investigate the effects of asphalt surface layer thickness, base type, base thickness, and drainage on performance of flexible pavements constructed in different site conditions (subgrade type and climate). Since there has been no comprehensive study of the SPS-1 experimental data, a thorough methodology involving analysis of variance (ANOVA), logistic regression and discriminant analysis was developed.
Although most of the findings from the analysis of the SPS-1 experiment data support the existing understanding of pavement rutting performance, the results from this study provide an outline of the interactions between design and site factors. The results showed that while pavement structural and asphalt mix designs are crucial, pavement construction quality plays a vital role in achieving better rutting performance for flexible pavements. This was clearly apparent in the SPS-1 experiment where 56 out of 212 sections had construction-related problems leading to premature rutting failures. Among the sections that did not show premature rutting, the effects of design and site factors on structural rutting were marginal. This is mainly due to the fact that these sections are still fairly young.
In general, pavements built on fine-grained soils have shown more rutting. Also, sections in warmer climates have shown slightly more rutting. Also, in general, regardless of site conditions, providing a thicker surface layer or an asphalt treated base layer will improve the structural rutting performance provided the materials are handled appropriately in design and during construction. Performance will be further enhanced if in-pavement drainage is provided, especially when the pavement structure is to be constructed on fine-grained soils.
Also, analysis of transverse surface profiles showed that most of the rutting in the test sections is confined to the upper layers of the pavement structure. Therefore, using the mechanistic-empirical prediction models that incorporate the contribution of each pavement layer to surface rutting is a better approach to characterize the rutting mechanism in flexible pavements.


by Lubinda F. Walubita, Amy Epps Martin, Charles J. Glover, Sung H. Jung, Gregory S. Cleveland, Robert L. Lytton, and Eun Sug Park


In this paper, a new Calibrated Mechanistic approach with Surface Energy (CMSE) measurements for fatigue characterization of hot-mix asphalt concrete (HMAC) mixtures is presented. The CMSE is a continuum micromechanics fatigue analysis approach based on Schapery's work potential theory, the visco-elastic correspondence principle, Paris' Law of fracture mechanics, and the energy concepts developed at Texas A&M University. The second objective of the study was to demonstrate the applicability of the CMSE approach for fatigue characterization of HMAC mixtures including investigating the effects of binder oxidative aging. To achieve these objectives, two Texas HMAC mixtures with different mix-design parameters and material characteristics subjected to three accelerated laboratory aging exposure conditions were evaluated. Analysis of the results indicated that the CMSE approach provides a promising methodology for fatigue characterization of HMAC mixtures based on fundamental mixture properties obtained from relatively simple laboratory material characteristic tests and a realistic fatigue failure criterion. The results obtained were comparable with other existing mechanistic- empirically based fatigue analysis approaches, and in fact exhibited significantly lower statistical variability. For the materials and test conditions considered in the study, binder oxidative aging reduced HMAC mixture fatigue resistance and its potential to heal. This finding signifies the importance of incorporating aging effects into the fatigue design and analysis of HMAC mixtures to ensure adequate fatigue performance. However, more research and HMAC mixture fatigue characterization is strongly recommended to further validate the CMSE approach.


by Rongzong Wu, John T. Harvey, and Carl L. Monismith


For highway rehabilitation and resurfacing projects, asphalt concrete (AC) overlay is the most common structure type for both old AC pavements and old Portland cement concrete (PCC) pavements. Yet despite the fact that reflective cracking is the dominant failure mode for AC overlays, current design guides either do not account for it or they treat it in a purely empirical manner. A mechanistic model for reflective cracking in AC overlays would not only help develop better design guides, but it would also help identify working solutions for retarding reflective cracking. This paper describes a research that takes a two-level approach to reach such a mechanistic model. At the material level, the research identified a constitutive model to describe cracking in AC mixes. At the structural level, the research used Finite Element Method to account for the effects of boundary conditions that define the forces driving crack development in the overlay. The paper first reviews existing models and identifies non-local continuum damage mechanics (CDM) as the most suitable framework for developing constitutive models for cracking in AC mixes. The non-local CDM was then customized to account for the special material behaviors of AC mixes. The paper then discusses the FEM implementation, the model parameter calibration, and field verification of the proposed mechanistic model. The research demonstrates that the implicit gradient non-local CDM, implemented in FEM, provides a promising mechanistic model for simulating reflective cracking.


by Samer Dessouky, Eyad Masad, and Dallas Little


Aggregate characteristics play a major role in the resistance of hot mix asphalt (HMA) to permanent deformation. Aggregate texture and angularity influence the frictional properties of the aggregate structure. In addition, the aggregate's non-spherical shape dictates the level of inherent anisotropy that develops due to the preferred orientation of aggregate particles under compaction forces.
The majority of previous studies on the influence of aggregate characteristics have focused on relating the mechanical response of HMA, measured using laboratory tests, to aggregate physical characteristics. A different approach is adopted in this study in which a mechanistic elasto-visco-plastic model is developed and implemented in finite element (FE) formulation to predict permanent deformation of HMA. The model's parameters are then related to aggregate shape characteristics. The main advantage of this approach is marked by its ability to identify aggregate characteristics that are required in order to optimize the mixture's resistance to permanent deformation.
The efficacy of the model is demonstrated by analyzing the permanent deformation response of three mixes that include aggregates with distinct characteristics. Aggregate characteristics are measured using the Aggregate Imaging System (AIMS). The model's parameters are obtained from compressive and extension triaxial tests conducted at different confining pressures and strain rates. Finite element analyses of pavement sections with these three mixes are conducted under uniform and non-uniform tire pressure distributions.
The relationships among the model's parameters, FE simulations of permanent deformation, and aggregate characteristics clearly illustrate the advantages of using the approach presented in this study to directly link aggregate characteristics to asphalt pavement performance.


by B. Shane Underwood, Y. Richard Kim and Murthy Guddati


This paper presents an argument for the utilization of constitutive models to predict fatigue performance. The argument is made within the framework of the viscoelastoplastic continuum damage (VEPCD) model. Four asphalt concrete mixtures, each a part of an ongoing study at the Federal Highway Administration's Accelerated Loading Facility (FHWA ALF), are characterized using this model. Polymer-modified asphalt binders are used with three of the four mixtures. Comparisons of linear viscoelastic, viscoelastic damage and viscoplastic properties are provided. Further, it is shown that the time-temperature superposition principle with growing damage is not limited to mixtures with simple asphalt binders and applies even to those with polymer-modified asphalt binders. The VEPCD model is shown to accurately predict material behavior over a range of conditions different than those used to characterize the model. Finally, experimental results are presented which show that fatigue performance is a complicated phenomenon dependent on many factors that may not be taken into account easily in current empirical procedures.


by Jianlin Wang, Bjorn Birgisson, Reynaldo Roque


Although the rheological behavior of hot mix asphalt (HMA) mixtures has been well recognized, it has rarely been taken into account in predicting the pavement response and evaluating its cracking performance. In this paper, a detailed analysis of the rheological behavior of asphalt concrete (AC) and its effect on the pavement response is carried out using a viscoelastic boundary element method developed recently by the authors. The research findings show that the stress distribution in flexible pavements is continuously changing as the pavement is loaded, because viscoelastic bituminous materials relax stresses. Generally, both tensile stresses at the top and compressive stresses at the bottom of the AC layer reduce with increased loading time or number of repeated loads. Continuous repeated loads may lead to accumulation of residual tensile stresses at the surface of the pavement and residual compressive stresses at the bottom of the AC layer. The presence of significant tensile stresses at the top of the pavement layer results in an accumulation of dissipated creep strain energy (DCSE) over time (or with number of loads) and may eventually lead to the formation of a crack. As time increases, the tension at the bottom of AC layer reduces significantly and may finally turn into compression. At the same time, a considerable level of tension builds up at the surface of the AC layer through loading and unloading cycles. This could possibly explain why bottom-up cracking does not appear as frequently as top-down cracking. Interestingly, these findings also appear to explain why top-down cracks have not been observed in parking lots. In summary, this paper clearly shows that the rheological behavior of the HMA mixture results in load-induced stress redistributions that may dominate the failure mode of pavement structure. The results presented imply a wide range of consequences for pavement engineers. The nature of the viscoelastic stress redistributions depends on the rheology of the binder, mixtures, pavement structure, and the loads applied. Therefore, successful optimization of a pavement structure and mixtures for enhanced cracking performance may require the inclusion of viscoelastic effects. The research findings reported in this paper open a door for the evaluation of the cracking performance of asphalt pavements considering the rheology of asphalt mixtures and for the subsequent adjustment of the current criteria for selection and design of mixtures and binder types.


by E. A. Ossa, H. Taherkhani and A. C. Collop


The compressive deformation behaviour of DBM and HRA asphalt mixtures is investigated over a wide range of strain-rates, stresses and temperatures under uniaxial and triaxial monotonic and cyclic compressive conditions. The response of the mixtures was found to be described well by a constitutive phenomenological model recently proposed by Ossa et al (1, 2). The steady- state monotonic behavior of the mixtures followed the modified Cross model with the mixtures exhibiting linear and nonlinear viscous behaviour at low and high stresses, respectively. Both loading and recovery responses were observed to be temperature dependent with the WLF relation capturing the temperature behaviour over the range of temperatures tested. The dilation gradient s was found to play an important role in the deformation behaviour of the mixtures studied. A higher dilation gradient increases the stiffening effect of the confining pressure and reduces the strain to reach tertiary creep. The recovery behaviour of the DBM mixtures was found to be different compared to the HRA mixtures which is thought to be due to the development of internal damage. Research into the causes and implications of this behaviour is proposed as a topic which merits further research.


by Eyad Masad, Corey Zollinger, Rifat Bulut, Dallas Little, and Robert Lytton


Moisture damage in asphalt mixtures can be defined as loss of strength and durability due to the presence of moisture at the binder-aggregate interface (adhesive failure) or within the binder (cohesive failure). This study focuses on evaluating the moisture susceptibility of asphalt mixtures through understanding the mechanisms that influence the adhesive bond between aggregate and asphalt, the cohesive strength of the binder and/or mastic (binder and mineral filler), and fracture of viscoelastic materials. Adhesive and cohesive bond energies are calculated from surface energy measurements. The viscoelastic and fracture properties are obtained using cyclic, torsional strain-controlled loading on cylindrical specimens of sand asphalt mixes until failure occurs. The Dynamic Mechanical Analyzer (DMA) apparatus was used to apply this loading.
A methodology is developed in this study to provide an index that is directly related to crack growth in asphalt mixtures subjected to dynamic loading. This index is a function of bond energy, viscoelastic properties, and fracture properties. The developed methodology was used to evaluate a number of asphalt mixtures that exhibited good or poor performance in the field. The resistance of the field mixtures to moisture damage was found in most cases to be related to the mixtures bond energies, the accumulated damage in DMA, and the crack growth index.


by Y. Kim, H.J. Lee, D.N. Little, and Y.R. Kim


This paper describes the development and validation of a simple and reasonable fatigue testing apparatus and analysis method that may be able to either supplement or replace traditional fatigue testing and analysis. In this method, three mechanical tests with cyclic torsional loading are performed to characterize basic linear viscoelastic material properties and the fatigue behavior of the asphalt matrix, which is, in this case, defined as a composite of asphalt mastic (binder and mineral filler) and the fine aggregate fraction. The testing induced damage in asphalt matrix that was sensitive to the materials tested. Damage accumulation within each asphalt mixture tested was due to microcrack development and propagation and ultimately to fatigue failure as a result of fatigue crack damage. This damage was monitored using image analysis. Testing results from asphalt matrix mixtures generally agreed well with testing data on asphalt binders and with fatigue testing results from asphalt concrete mixtures. Furthermore, the mechanically-developed fatigue life prediction model can be applied to various asphalt mixture systems by simply performing fundamental material property tests such as frequency sweep tests. Successful development of this testing protocol has the potential to become a specification-type test method because it is fast, repeatable, and accurate.


by Sungho Kim, Reynaldo Roque, and Bjorn Birgisson


Coarse aggregate structure or interlock is critical for resistance to rutting, and recent work has shown that it can also play a significant role in resistance to damage and fracture. Therefore, large enough aggregates should engage dominantly in the structure for good mixture performance. This study focused on the development of a conceptual and theoretical approach to evaluate coarse aggregate structure based on gradation.
In soil mechanics, it has been well established that the porosity of granular materials in the loose state is approximately constant between 45% and 50%, regardless of size or gradation. This implies that one can use porosity as a criterion to assure contact between large enough particles within the mixture to provide suitable resistance to deformation and fracture. Calculations performed for gradations associated with typical dense graded mixtures indicated that the porosity of particles retained on any single sieve was significantly greater than 50%, even for gradations associated with the maximum density line. Since many dense graded mixtures are known to provide suitable resistance to deformation and fracture, then there must be a range of contiguous coarse aggregate particle sizes that form a network of interactive particles with a porosity of less than 50%.
A theoretical analysis procedure was developed to calculate the center to center spacing between specific size particles within a compacted assemblage of particles of known gradation. Calculations performed with this procedure indicated that the relative proportion of two contiguous size particles, as defined by the standard arrangement of Superpave sieves, can be no greater than 70/30 in order to form an interactive network. Thus, the 70/30 proportion can be used to determine whether particles on contiguous Superpave sieves can form an interactive network of particles in continuous contact with each other. The range of particle sizes determined to be interactive was referred to as the dominant aggregate size range (DASR) and its porosity must be no more than 50% for the particles to be in contact with each other.
It was concluded that porosity of the DASR may provide a good criterion for determining the suitability of gradation for dense-graded asphalt mixture.


by Michael P. Wagoner, William G. Buttlar, Glaucio H. Paulino, and Philip Blankenship


The calibration of pavement models to field observations can be an important step in refining the model to accurately predict the pavement response. A major challenge with model calibration has been obtaining the material properties with laboratory test geometries that have the capability to use specimens fabricated from standard field cores. Unless complicated fabrication procedures are used, field cores limit the specimen geometry to cylindrical shapes and thin cross sections due to the relatively thin pavement lifts typically used during pavement construction. Recent developments in fracture testing of HMA concrete have enabled the development of a testing suite that captures both continuum and material separation properties from field cores. The testing suite utilizes two specimen geometries (indirect tension and disk- shaped compact tension) and was developed to investigate reflective cracking mechanisms.
The main objective of this paper is to illustrate the application of the testing suite to provide material properties for material constitutive models. For this purpose, a single pavement with three test sections was selected. The test sections consist of similar pavement structures, but different asphalt mixture designs, namely different grades of asphalt binder. The testing suite was successful in determining the continuum properties of interest, namely complex modulus, creep compliance, and coefficient of thermal expansion, using the indirect tension geometry. Two parameters were required to describe the material separation. The tensile strength was obtained using the indirect tension geometry while the fracture energy was determined using the disk-shaped compact tension geometry. As expected, the continuum properties are influenced by the asphalt binder grade where both the complex modulus and creep compliance show the same trends. The material separation properties also appear to be influenced by the asphalt binder grade. The fracture energy appears to distinguish the fracture resistance of the material better than the tensile strength. Future work includes integrating the material properties with finite element analysis to provide a detailed analysis of the pavement structure to determine the mechanisms that initiate the reflective cracks.


by Adrian Andriescu, Nelson H. Gibson, Simon A.M. Hesp , Xicheng Qi, and John S. Youtcheff


This paper considers the essential work of fracture approach to asphalt binder grading for fatigue performance. Earlier published test results (Andriescu et al., Journal of the Transportation Research Board, No. 1875, 2004, pp. 1-8 and Andriescu et al., Proceedings of the Canadian Technical Asphalt Association, Vol. 49, 2004, pp. 93-122), are further analysed and subsequently validated with fatigue cracking data from the US Federal Highway Administration's pavement testing facility (FHWA PTF), which has recently become available.
Samples were tested in double-edge-notched tension (DENT) while the ligament length was varied. Duplicate measurements were generally found to be highly reproducible. Shapes of the load versus displacement records were found to be self-similar for all ligament lengths yet distinctly different between binders. The specific total work of fracture was plotted versus the ligament length to provide nearly straight lines. These observations validate the essential work of fracture approach to asphalt binder testing. The analysis of the DENT test results provides: (1) the essential work of fracture, we, a measure of the work done in the fracture process zone, (2) the plastic work of fracture, wp, associated with energy absorbing processes away from the process zone, (3) the net section stress, ån, which provides a good approximation for the tensile yield stress, åty, and (4) the critical crack tip opening displacement, ët (= we/åty), all in ductile failure. Binder parameters were compared with cracking rates in the respective lanes of the FHWA PTF. Only the ët property was able to correctly rank performance and provide a high correlation with cracking distress for all five PTF sections studied. A system containing fibres, showed surprisingly high essential and plastic works of fracture compared to other materials that were evaluated. This ability for damage delocalisation may well explain the fact that the corresponding PTF lane shows almost no cracking to date.
Given the fact that ët is a proven fracture mechanics based parameter that provides a measure of strain tolerance in the presence of cracks in the brittle-to-ductile and ductile states, it deserves further investigation for fatigue ranking of asphalt binders. Eventually the authors envision the need for a comprehensive specification that includes limits on binder and perhaps mixture we, wp, and ët in the ductile state.


by A.A.A. Molenaar, A.J.J. Meerkerk, M. Miradi, and T. van der Steen


Because of its noise reducing capacity, porous asphalt concrete is commonly used nowadays in the Netherlands for wearing courses on highways. Traditionally a single layer of porous asphalt concrete (PAC) (gradation 6/16 mm) is applied with a thickness of 50 mm. Investigations however have shown that double layer PAC, consisting of a 30 mm top layer of fine graded (2/6 or 4/8 mm) PAC on top of 50 mm coarse graded (11/16 mm) PAC, have better noise reducing capabilities. These mixtures are therefore going to be used on a large scale on highways in densely populated areas. The lifetime of PAC shows a large variation and in order to keep the maintenance costs under control, measures are sought to extend the mean lifetime and to reduce the variation therein. More and more it is believed that the lifetime of PAC and its variation are strongly influenced by the construction process. A study was therefore undertaken to determine to what extent the development of the main defect type, being raveling, is determined by the composition of the mixture as laid and factors like traffic and climate. For this, the PAC sections in the SHRP-NL data base were analyzed by means of Artificial Neural Network techniques (the SHRP-NL data base only contained information on single layer PAC). This study revealed that the quality of the mixture as laid has a very large influence on the performance of the mixture and it was shown e.g. that bitumen contents lower than 4% should not be accepted. Also siliceous river gravel should not be used.
Furthermore a study was undertaken to determine to what extent the initial quality of single layer PAC was influenced by the working conditions. For this a particular paving job was intensively monitored and it appeared that the composition of the mixture as laid is highly variable. From the data an equation could be developed that relates the void content, being the key parameter of PAC, to mixture composition, temperature of the mixture and compaction effort.
Finally research was done on double layer PAC. This double layer offers a better reduction in traffic noise than the conventional single layer PAC but problems may arise with the permeability because of a low void content in the interface between top and bottom layer. It will be shown that selection of the gradation is of prime importance for noise reduction and permeability purposes and an optimal gradation for both layers will be proposed.


by Chun Chen, Akhtarhusein A. Tayebali, and Detlef R. U. Knappe


Litmus and colorimetric tests were used to quantify the contents of amine-based antistrip additives in asphalt binders and mixes. In addition, the effect of prolonged heating on antistrip additive content was evaluated for both asphalt binders and mixes. Results of this study indicate that both litmus and colorimetric tests are capable of detecting and quantifying amine-based antistrip additives in asphalt binders and mixes. Also, both test methods were able to validate the antistrip additive contents in field samples with known additive contents. When subjected to prolonged heating periods, the antistrip additive content decreased substantially for both asphalt binders and mixes. For asphalt binders, no antistrip additive content was detectable after 24 to 48 hours of extended heating; for mixes, the measured antistrip additive content approached zero percent after 6 to 12 hours of extended heating.


by Michael Heitzman, PhD, PE


The objective of this study is to develop a more accurate measure of film thickness (FT). The concept of FT was introduced in the 1950's and assumes that all the asphalt exists as uniform films. The equation was expressed in English units and uses total binder volume, weight of aggregate, aggregate surface area, and a constant to convert to microns. The equation was later improved to account for absorbed asphalt binder.
Numerous studies over the years examined FT as a durability parameter, but lack of strong correlation usually eliminated FT as a primary parameter. Most studies applied the FT equation without any question about its accuracy.
The procedure for determining surface area is part of the Hveem mix design method. The surface area values are not a direct expression of particle surface area and do not account for differences in aggregate specific gravity. This is a critical limitation. Studies that examine differences in FT values in an attempt to identify trends in HMA mixture performance are not comparing equivalent FT values when the mixtures have aggregates with dissimilar specific gravities.
In 2003, a new procedure introduced a three-dimensional mixture model. The approach models the random orientation of size-graded spheres and distributes the binder into the remaining available space. It takes into account that FT is not a uniform coating of each particle.
This study generates two different film thickness values. The first is an extension of the past practice of a uniform coating "index" and the second is based on the "virtual" three- dimensional model. The proposed Index Model uses the fundamental principles of weight, volume, specific gravity, and particle geometry to calculate a surface area of each aggregate particle and accounts for the differences between aggregate sources. However, it still uses the simplification that the surface area coated by the asphalt is a flat surface and each particle is separately and uniformly coated.
The Virtual Model examines the spatial relationship between particles and requires knowledge of the mixture volumetrics, binder, and aggregate proportions. This three dimensional model achieves a film thickness value that approaches the true film thickness. If our primary interest in film thickness is to define the durability of the mixture, then the nominal thickness of the asphalt between the aggregate surface and the void space represents the minimal depth of exposed coating where binder aging occurs.
The study ran a sensitivity analysis on a family of generic mixes and examined the differences between the Standard Model, the Index Model, and the Virtual Model. Each analysis isolated a mixture parameter to check the reasonableness of the proposed models. The analysis included gradation, aggregate specific gravity, amount and gradation of the mineral filler, particle shape, level of mixture compaction, and multiple source blended gradations. The sensitivity analysis demonstrates that the Standard Model does not accurately express FT. The proposed Index Model is a better two-dimensional approach. The Virtual Model is a new concept and can reflect differences in mixture density.


by Sunghwan Kim and Brian J. Coree


Recently, the Iowa Department of Transportation requested the Department of Civil, Construction and Environmental Engineering (CCEE) at Iowa State University to evaluate the possibility of using 150mm diameter Superpave samples instead of 4" (100mm) Marshall samples in an HMA moisture sensitivity test. The CCEE Department was asked to look into the possibility of a test using their Nottingham Asphalt Tester (NAT). This required that the team re- evaluate current test protocols, and re-think (re-engineer) the approach to moisture sensitivity testing. A test protocol is proposed and was used in the DOT research project.
Test results were analyzed using three parameters based on performance characteristics: the retained flow number based on critical permanent deformation failure (RFNP), the retained flow number based on cohesion-failure (RFNC) and the energy ratio (ER).
Analysis demonstrated that the energy ratio of elastic strain (EREE ) at a flow number at cohesion failure (FNC) showed a higher potential to discriminate HMA moisture sensitivity.
It was also shown that the vacuum saturation used in AASHTO T 283 and in the proposed test may unreasonably damage some specimens before testing, leading to "false" results.


by Baoshan Huang, Ph.D., P.E., Jingyao Cao, Ph.D., Xingwei Chen, Xiang Shu, and Wenbin He


An asphalt mixture generally behaves as an insulator. The addition of electrically conductive additives may produce conductive asphalt mixtures. Traditional technology of applying carbon-based additives compromises the mechanical properties of hot-mix asphalt (HMA) mixtures and prevents the wide applications of electrically conductive asphalt. The present study investigated into various options of producing electrically conductive HMA mixtures. Electrical conductivity was evaluated for asphalt mixtures containing micron-scale steel fiber, aluminum chips, and graphite as conductive fillers. Further investigation was conducted relating the electrical conductivity to other engineering properties of HMA mixtures. The results from this study indicated that the micron-scale steel fiber provided the most efficient way to increase the electrical conductivity of HMA mixtures. The percolation point for the micron-scale steel fiber was 0.2% (by volume of asphalt binder); whereas, those points for graphite and aluminum powder were much higher. A strong relationship was discovered between the electrical resistivity and mechanical properties for asphalt mixture specimens under indirect tensile strength and beam fatigue testing. Further studies are recommended to explore the many applications related to the electrical conductivity of HMA mixtures.


by Ghazi Al-Khateeb, Aroon Shenoy, Nelson Gibson, and Thomas Harman


This paper presents a new mechanistic empirical model for predicting dynamic modulus of asphalt paving mixtures at a wider range of temperatures and loading frequencies, which can be shifted into a master curve for characterizing asphalt concrete. Available predictive models seem to be not capable of predicting dynamic modulus of asphalt mixtures at higher temperatures and lower loading frequencies; these models over-predict dynamic modulus by a significant deviation from laboratory-measured values. The new model is capable of accurately predicting measured dynamic modulus at a broader range of temperatures and loading frequencies.
The proposed new model was derived from the law of mixtures where composite materials are modeled in a combination of parallel and series phases. For a system of purely parallel phases, the combined mechanical behavior is simply the addition of the responses from these phases.
To develop the new model, asphalt mixtures with different performance grades covering highly modified and unmodified asphalt binders were tested in one of the marketed simple performance test (SPT) devices to measure dynamic modulus.
The new model is simple in its formulation, and needs no more than one input from the asphalt binder; the dynamic shear modulus (|G*|), and one input from the mixture; the voids in mineral aggregate (VMA).


by Carl L. Monismith, Lorina Popescu, John Harvey


This paper presents a methodology for estimating rut depths in asphalt concrete pavement structures for use in mechanistic-empirical pavement design and analysis. The procedure utilizes stiffness and plastic strain versus repetitions data obtained from the SHRP-developed repeated load simple shear test conducted at constant height (RSST-CH). To use this methodology, the pavement is assumed to behave as a multilayer elastic system for determination of key stresses and strains to permit rutting estimates to be made at the pavement surface for specific mix types. Influences of temperature and time of loading on the asphalt concrete stiffness are considered in an ad-hoc manner.
The paper includes a discussion of the approach used as well as calibrations of the methodology for selected WesTrack pavement sections (those with rutting but no visible fatigue cracking). Specific performance models for three California dense-graded asphalt concrete mixes obtained from calibrations obtained from Heavy Vehicle Simulator tests under controlled temperature conditions are also presented.
Results of the use of two of these models to estimate rutting in an in-service pavement near Sacramento, California for which condition data were available are also presented.
The methodology is then used to evaluate 17 different pavement sections designed according to the current Caltrans procedure for a range in traffic loadings and checked to ensure that the asphalt concrete thicknesses were adequate to resist fatigue cracking. Results of these analyses performed for three climate regions reflect current Caltrans experience with performance of the mixes used.
Appendices to the paper include the description of the development of the performance model from RSST-CH tests on WesTrack mixes and the step by step procedure used to calculate rut depth for a specific pavement structure.