INTRODUCTION
Problem Statement
In recent years, the United States has seen a rise in rutting of hot-mix asphalt (HMA) pavements, primarily attributed to the HMA layers themselves Contributing factors include changes in asphalt binder properties, increased traffic volumes and weights, and elevated tire pressures In response, the Strategic Highway Research Program (SHRP) launched a five-year, $50 million study in 1987 to tackle these performance issues, leading to the development of the Superpave volumetric mixture design method This innovative design approach has significantly reduced instances of premature permanent deformation; however, concerns about the long-term performance and durability of Superpave-designed mixtures remain.
The relationship between air voids and the density of a Hot Mix Asphalt (HMA) mixture is crucial, as increased density leads to fewer air voids, which significantly impacts pavement performance An adequately compacted HMA mixture must maintain a balanced amount of air voids to prevent permanent deformation, such as rutting and shoving, while avoiding excessive voids that can lead to moisture infiltration, oxidation, and subsequent issues like raveling and cracking Therefore, air voids play a vital role in determining permeability; as the number of in-place air voids rises, so does the permeability of the mixture.
Objectives and Scope of Study
Since HMA pavement performance and durability are directly related to the air voids contents and thereby permeability of the HMA pavement, the objectives of this study are to:
1 Better understand the increase in HMA pavement performance and durability that can be gained by increasing the initial pavement density; and
2 Better quantify the inter-relationship among HMA pavement density, permeability, and moisture-induced damage
To meet the specified objectives, the study evaluated various parameters of Hot Mix Asphalt (HMA) mixtures to determine their impact on performance and durability All analyses were conducted in relation to the air void content and initial density of the mixtures.
Research Approach and Methodology
Pavement density significantly affects the performance and durability of Hot Mix Asphalt (HMA), yet the extent to which increased initial density enhances long-term outcomes remains uncertain This study evaluated the long-term performance and durability of four HMA mixtures with varying air void contents in a laboratory setting Three key factors influencing HMA mixture performance were analyzed: density (expressed as a percentage of maximum theoretical specific gravity), nominal maximum aggregate size (NMAS), and gradation The density was assessed at four levels (90%, 92%, 94%, and 96%, corresponding to air void contents of 10%, 8%, 6%, and 4%), while NMAS was examined at two levels (9.5 mm and 19 mm) and gradation at two categories (fine- and coarse-graded).
Following the completion of four mixture designs, appropriate specimens were prepared and tested to explore the correlation between experimental factors and the long-term performance and durability of pavements, as assessed through laboratory tests The study utilized three main groups of testing methods: permeability and porosity testing, moisture susceptibility testing, and performance and durability testing.
Permeability testing was conducted using the falling head permeability and Corelok tests, employing equipment and methods developed by the Florida Department of Transportation (FDOT) The Corelok method was utilized to assess the porosity of the mixture samples, with research indicating that porosity serves as a more reliable indicator of permeability than air void content Both tests were performed on specimens prepared using the Superpave Gyratory Compactor (SGC).
Moisture sensitivity testing was conducted using the AASHTO T283 method and the Purdue University laboratory wheel tracking device, known as PurWheel AASHTO T283 is a recognized standard for assessing moisture susceptibility in Hot Mix Asphalt (HMA) The PurWheel test involves passing a loaded wheel over a compacted HMA slab while submerging the specimen in hot water to evaluate its moisture damage potential This method has proven to be effective and reliable in determining the moisture sensitivity of HMA mixtures.
The performance and durability of Hot Mix Asphalt (HMA) were evaluated using dynamic modulus and beam fatigue tests on both unconditioned and conditioned samples Two conditioning methods were employed: moisture conditioning, where samples were partially saturated and immersed in a 60°C water bath for 24 hours, and air conditioning, where specimens were placed in an 85°C forced draft oven for five days Moisture conditioning is known to promote moisture damage, allowing for a quantifiable decrease in performance and durability as indicated by the dynamic modulus and beam fatigue tests Conversely, oven conditioning tends to diminish durability and performance, particularly evident in the beam fatigue test results Notably, dynamic modulus samples were not tested post-oven conditioning, as this process stiffens the HMA mixture, thereby reducing the likelihood of permanent deformation.
This report outlines the testing details, results, and conclusions, beginning with a theoretical review of test methods and their parameters in Chapter 2 Chapter 3 discusses the materials used, including asphalt binder and aggregates, as well as the project's experimental design The subsequent chapters (4 to 7) provide comprehensive information on testing, results, and statistical analyses for permeability, moisture susceptibility, dynamic modulus, and beam fatigue tests Finally, Chapter 8 summarizes the findings and offers conclusions and recommendations.
LITERATURE REVIEW
Permeability
Permeability refers to the capacity of Hot Mix Asphalt (HMA) to allow water to flow through its pores under pressure or varying hydraulic head This property is quantified by measuring the volume of fluid that passes through a specific cross-sectional area of the material over a set period, reflecting the material's efficiency in fluid transmission.
Previous research has highlighted the critical parameters influencing the performance and durability of Hot Mix Asphalt (HMA) mixtures under various climatic and load conditions Key findings suggest that air void content is a significant factor affecting performance, with some studies indicating that permeability may be more influential, viewing air voids as an indicator of permeability Brown emphasizes that the quantity of air voids, or the degree of compaction, is paramount for HMA pavement performance, recommending that initial in-place air voids should be below 8 percent and terminal air voids above 3 percent Ford supports this by asserting that HMA mixtures must maintain a minimum terminal air voids content of 2.5 percent to prevent excessive rutting and hydroplaning Additional studies indicate that significant rutting occurs when air voids reach around 3 percent Harvey et al argue for a balance in air void size to ensure structural stability while avoiding degradation under load Conversely, McLaughlin and Goetz suggest that a mixture's ability to prevent water ingress is more indicative of pavement performance than density alone.
2.1.1 Effect of Air Voids Content
Increases in air voids content within a mixture lead to higher permeability, with a critical threshold identified at approximately 8 percent air voids, where mixtures become permeable to both air and water Research by Zube and others indicates that once air voids exceed this percentage, permeability escalates significantly Studies conducted by Brown et al and Santucci et al further support the finding that maintaining air voids below 8 percent is essential to prevent rapid oxidation, cracking, and raveling in dense-graded mixtures Thus, 8 percent air voids content serves as a crucial benchmark for differentiating between permeable and impermeable hot mix asphalt (HMA) mixtures.
Kanitpong et al (11) identified air voids content as the most significant factor affecting permeability, while also noting that effective air voids (EAV), or the percentage of porosity comprised of water-permeable voids, plays a crucial role In a subsequent study (12), they found that EAV serves as a superior indicator of permeability compared to air voids, attributing the differing results between the two studies to variations in the saturation levels of the samples analyzed.
2.1.2 Effect of Aggregate Gradation and Size
Various studies have shown that aggregate characteristics, particularly gradation and nominal maximum aggregate size (NMAS), significantly affect permeability Research following the introduction of the Superpave mixture design method indicates that coarse-graded Superpave mixtures exhibit higher permeability compared to traditional hot mix asphalt (HMA) mixtures when air void content is consistent Additionally, Cooley et al identified that both NMAS and gradation play crucial roles in influencing permeability, noting that larger NMAS correlates with increased void sizes within the mixture They determined that to meet a maximum field permeability specification of 100×10 -5 cm/s, the tested mixtures would require specific air void contents of 7.7%, 5.5%, and 4.4% for 9.5-, 12.5-, 19.0-, and 25.0-mm mixtures, respectively.
Achieving low air void content during construction is crucial, as interconnected voids can lead to increased permeability in mixtures, allowing water and air to flow through them Research indicates that coarse-graded mixtures, which fall below the maximum density line (MDL), exhibit higher permeability compared to fine-graded mixtures above the MDL This finding is supported by studies, including one by Choubane et al., which highlighted the permeability of coarse-graded Superpave mixtures The gradation of materials significantly influences the size of air voids in compacted hot mix asphalt (HMA), with a higher percentage of coarse particles leading to more interconnected voids Both larger nominal maximum aggregate sizes (NMAS) and coarser gradations contribute to increased permeability due to insufficient fines to occupy void spaces Additionally, Prowell's research confirms that fine-graded mixtures are generally less permeable than their coarse-graded counterparts.
Research by Choubane et al revealed that the size and interconnectivity of air voids significantly affect pavement permeability They found that coarse-graded Superpave mixtures could experience permeability issues even with air void content below 8 percent, while no significant permeability changes were observed below 7 percent Pavements became nearly impermeable at air voids below 6 percent, but a slight increase above 7 percent led to a substantial rise in permeability In a related study by FDOT on fine-graded Marshall designed mixtures, those with air void contents exceeding 7 percent were found to be impermeable, indicating less interconnectivity of voids in fine-graded mixtures compared to coarse-graded ones.
Masad et al (17) conducted research on the significance of aggregate shape and size distribution in relation to permeability They formulated an empirical equation (2.1) to predict the permeability of mixtures, utilizing air void content and aggregate surface area as key parameters.
Va = air voids of porous media (%); m = regression coefficient; c = constant that depends of the idealized shape of the air voids;
Sagg = average specific surface area of given gradation and NMAS
(1/mm 2 ); γ = unit weight of water at 20C (9.79 kN/m 3 ); and à = viscosity of water (10 -3 kg/m-s)
The permeability of Hot Mix Asphalt (HMA) mixtures is influenced by both the air voids content and the size of the voids Contrary to the common assumption that permeability is solely proportional to air voids, Hudson and Davis found that non-interconnected voids can lead to higher air voids while still offering resistance to moisture intrusion.
2.1.3 Comparison of In-Service and Laboratory Results
Research work comparing in-service (field) permeability to the results of laboratory permeability tests has been completed by Cooley et al (19) and
Maupin's report reveals that five out of six tested mixtures showed that the permeability of field cores closely matched that of laboratory-compacted specimens However, in one instance, he attributed the discrepancy to significant variability in the field core results.
Cooley et al (19) successfully compared the permeabilities of field cores and laboratory-prepared specimens across five mixtures, finding that only one mixture (9.5-mm NMAS) did not align well, while the other four showed strong correlation They suggest that with more data, the comparisons could improve significantly Furthermore, they analyzed laboratory permeability test results against field permeability tests, concluding that field permeability values below a certain threshold indicate a notable distinction in performance.
Field and laboratory tests produced comparable results at a permeability rate of 500×10^-5 cm/s Given that the recommended field permeability criterion is significantly lower than this threshold, the laboratory test outcomes can be considered dependable indicators of field performance.
Moisture Susceptibility…
Moisture susceptibility refers to the deterioration of the adhesive bond between aggregate and binder in Hot Mix Asphalt (HMA) due to moisture, often resulting in a phenomenon known as stripping This issue can also arise from the loss of cohesive resistance in the binder film that coats the aggregate Moisture exposure weakens the binder matrix, reducing the stability and load-carrying capacity of the HMA mixture The consequences of stripping can lead to various pavement distresses, including rutting, shoving, and fatigue cracking.
FIGURE 2.1 Cohesive and Adhesive Failures (after [21])
The stripping process in asphalt mixtures is explained by six primary mechanisms: detachment, displacement, spontaneous emulsification, film rupture, pore pressure, and hydraulic scouring Detachment involves the separation of the binder film from the aggregate by a thin water layer, while displacement refers to the binder film being removed by water Spontaneous emulsification occurs when adhesion between the binder and aggregate is lost due to water being suspended within the binder, leading to cohesion failure without visible loss of coating Film rupture happens when sharp aggregate points break the binder film, often due to construction loads or freeze-thaw cycles, allowing moisture to reach the binder-aggregate interface The pore pressure theory suggests that increased water pressure from trapped moisture can exceed the adhesive strength between the binder and aggregate, facilitating stripping, particularly in mixtures with larger pore diameters Lastly, hydraulic scouring is caused by vehicle movement, which pushes water into the asphalt mixture and can lead to raveling.
Moisture damage poses a significant challenge across the country, leading to millions in reconstruction and maintenance expenses since the introduction of the Superpave specification Stripping typically begins at the base of the Hot Mix Asphalt (HMA) layer, often remaining undetected until the issue becomes severe Consequently, it is crucial to manage the factors that heighten the risk of stripping to prevent costly repairs.
Inadequate compaction is a primary cause of stripping in hot mix asphalt (HMA) pavements, making the control of air voids crucial Ideally, HMA mixtures should contain 4-5 percent air voids, which prevents moisture infiltration due to the lack of interconnected voids Newly constructed pavements typically specify a maximum in-place air voids content of 8 percent, with the expectation that traffic will compact the mixture to the design air voids level over two to three years However, if the initial air voids are not properly managed and exceed this limit, the pavement can become permeable, allowing water to infiltrate and lead to stripping.
Stripping in pavement can be caused by several factors, including dust or clay coatings on aggregates, insufficient drying of aggregates, the presence of weak and friable aggregates, and poor drainage When aggregates are coated with dust or clay, the bond between the binder and aggregate weakens, allowing moisture to penetrate through formed channels Inadequate drying of aggregates prior to mixing with hot mix asphalt (HMA) can lead to poor adhesion, raising the risk of stripping Additionally, weak aggregates may deteriorate under traffic, exposing surfaces that can absorb moisture Lastly, ineffective drainage can allow water to infiltrate the pavement, further increasing the likelihood of moisture damage.
Since 1945, multiple methods have been established to assess moisture susceptibility in Hot Mix Asphalt (HMA) mixtures, including the Immersion-Compression test, Lottman test, Root/Tunnicliff test, Marshall Immersion test, Resilient Modulus, and the Double Punch Method For detailed information on these tests, refer to reference (22) Additionally, wheel tracking devices, initially created to evaluate rutting in HMA mixtures, can also effectively predict moisture susceptibility when tested in water.
The work reported herein uses the AASHTO T283, “Standard Method of Test for Resistance of Compacted Asphalt Mixtures to Moisture-Induced
Damage,” as well as the PurWheel to determine the stripping potential in HMA mixtures AASHTO T283 is the conventional method used by the Indiana
The Indiana Department of Transportation (INDOT) utilizes the indirect tensile strength ratio (TSR) of both conditioned and unconditioned samples to predict moisture damage in materials Developed by Purdue University, the PurWheel is a laboratory tracking test that assesses specimens submerged in water Detailed testing procedures are outlined in Chapter 5 of the report.
AASHTO T283 is the most commonly used test method for predicting stripping potential, although its accuracy in forecasting in-service moisture damage has faced criticism Kandhal identified AASHTO T283 as the best method for predicting moisture damage, while Choubane et al examined the impact of varying saturation levels on test outcomes and proposed modifications to enhance the assessment of stripping potential They recommended establishing a minimum wet tensile strength of 410 kPa and a minimum tensile strength ratio (TSR), noting a strong correlation between saturation levels and TSR results Based on observed moisture damage in the field, they advised using a saturation level exceeding 90 percent and incorporating a freeze-thaw cycle in testing.
Pan and White (25) found that AASHTO T283 may not accurately represent the stripping potential of Hot Mix Asphalt (HMA) mixtures, as larger 150-mm diameter samples exhibited higher Tensile Strength Ratio (TSR) values than smaller 100-mm diameter specimens, indicating they were less impacted by conditioning procedures They suggested a minimum conditioned tensile strength of 600 kPa Furthermore, Mahoney and Stephens (26) reported that numerous HMA samples tested in Connecticut failed to meet the specified TSR limit.
AASHTO T283, yet most of the mixtures had no stripping problems in the field
The PurWheel is a laboratory wheel tracking test created at Purdue University
At the time of its introduction, three distinct laboratory wheel tracking devices were in use: the French Rutting Tester from the Laboratory Central de Ponts et Chausses (LCPC), the Georgia Loaded Wheel Tester (GLWT), and the Hamburg Steel Wheel Tracking Device (HSWT) The PurWheel, which is an enhanced version of the HSWT, incorporates improvements aimed at more accurately simulating real-world conditions.
PurWheel is designed to conduct tests under both dry and wet conditions, utilizing either steel or rubber wheels It effectively incorporates wheel wander and employs a larger sample size to reduce boundary effects Furthermore, the rut depth can be accurately measured along the entire length of a specimen using movable transducers These enhancements enable PurWheel to simulate conditions related to rutting and stripping, such as high moisture, elevated temperatures, and dynamic wheel loads.
Pavement deformation includes two key elements: compactive deformation occurring in the wheel path and plastic deformation, which refers to the elevation in the Hot Mix Asphalt (HMA) mixture that can happen both within and beyond the wheel paths.
PurWheel accurately measures compactive deformation with each wheel pass, allowing for the assessment of total deformation—both compactive and plastic—at the conclusion of the test When testing PurWheel specimens with water present, additional rutting damage may arise from hydraulic scouring and excessive pore pressure from moving wheel loads The risk of moisture damage can be forecasted by graphing compactive deformation against the number of wheel passes, as demonstrated by Izzo and Tahmoressi.
The plot described in Figure 2.2 consists of three distinct sections: the creep slope, the stripping slope, and the stripping inflection point The creep slope is associated with rutting caused by plastic flow, while the stripping slope pertains to rutting primarily resulting from moisture damage The stripping inflection point marks the intersection of the creep and stripping slopes, indicating the number of passes at which these two behaviors converge.
FIGURE 2.2 Wheel Tracking Results (after [25])
Kandhal (23) indicated that stripping of fine aggregate is more critical and
Pan and White (25) found that fine aggregate plays a crucial role in the adhesion loss of asphalt binder films They highlighted that the size of the aggregate and the gradation of the mixture significantly influence rutting resistance Additionally, the inclusion of crushed sand may mitigate moisture damage Their research also indicated a connection between fine aggregate angularity and/or gradation and the instability of the mixture.
2.2.4 Comparison of In-Service and Laboratory Results
Research has shown the effectiveness of laboratory wheel tracking to predict stripping potential in HMA mixtures Pan and White (25) concluded that the
Long-Term Performance and Durability
The main issues affecting HMA pavement include permanent deformation, fatigue cracking, and thermal cracking, all of which significantly impact its durability and long-term performance.
Permanent deformation, fatigue, and thermal cracking in Hot Mix Asphalt (HMA) are influenced by environmental conditions, loading factors, and the mixture composition Engineers tasked with HMA mixture design must assess the specific external conditions at the pavement site and carefully choose materials and their combinations to ensure optimal pavement performance.
Permanent deformation, commonly known as rutting, occurs in pavement layers due to traffic loads and can be categorized into two stages: compactive deformation and plastic deformation Compactive deformation results in a deformed surface that is lower than the original pavement surface, while plastic deformation leads to a deformed surface that exceeds the height of the initial surface.
Zhou and Scullion identify three distinct stages of rutting in hot mix asphalt (HMA) pavements The first stage, known as the primary or pre-failure zone, is characterized by rapid accumulation of permanent deformation due to increased total plastic strain as HMA mixtures undergo hardening with repeated loading This phenomenon is explained by the physical processes of motion and dislocation, leading to micro-flow The second stage, referred to as the steady stage, sees a constant rate of permanent deformation, where micro-cracking contributes to further dislocation development.
In the tertiary or failure zone, the formation and growth of micro-cracks lead to an increase in the rate of permanent deformation As these micro-cracks propagate and coalesce, they create macro-cracks, which further accelerate work-softening This dynamic process results in a significant rise in the permanent deformation rate, highlighting the critical transition from work-hardening to work-softening in the material's behavior.
The rutting performance of Hot Mix Asphalt (HMA) mixtures is influenced by the properties of aggregates and binders, as well as their interactions within the mixture Research by Tarefder et al evaluated Superpave mixtures using the Asphalt Pavement Analyzer (APA) to identify key factors affecting rutting Their findings indicated that the type of binder, specimen type, test temperature, and their interactions were the most critical factors, while moisture, wheel load, binder content, and hose pressure were deemed less significant.
FIGURE 2.3 Stages of Permanent Deformation (after [1])
Pavement rutting occurs when the subgrade soil and HMA layers are overstressed, leading to densification or shear failure Key factors contributing to rutting include inadequate pavement structure, poor drainage, insufficient initial compaction, and improper HMA mixture design, such as high binder content, excessive filler, or an overabundance of rounded particles in the aggregates.
HMA rutting is influenced by the properties of the binder and aggregates, along with their interactions To mitigate rutting in the HMA layer, it is recommended to increase the voids in the mineral aggregate (VMA), set minimum and maximum air voids contents, limit natural sand usage, ensure a minimum percentage of crushed coarse and fine aggregates, utilize a stiffer binder, and adopt coarser mixture gradations.
Low air voids in mixtures can lead to instability when they reach around 2 percent, while higher air voids increase the risk of rutting due to reduced stiffness and moisture damage Excessive air voids elevate the permeability of Hot Mix Asphalt (HMA), heightening the risk of water and air infiltration, which can cause binder brittleness from oxidation and stripping of the binder from aggregates Additionally, high air voids decrease the mixture's stiffness, making it more susceptible to permanent deformation.
The implementation of Superpave introduced various specifications to aid in material selection for optimal pavement performance A key feature of these specifications was the establishment of a "restricted zone" that the blended aggregate gradation of the mixture could not exceed This restricted zone aims to create a robust stone skeleton, enhancing the pavement's performance and durability Research has categorized gradation into three groups concerning the restricted zone: above it (fine gradation), below it (coarse gradation), and within it.
Research has shown that while coarser gradations are often believed to enhance rut resistance in Hot Mix Asphalt (HMA) mixtures, several studies indicate that finer-graded mixtures may actually exhibit lower rut potential Cross et al found that finer gradations demonstrated significantly greater resistance to shear deformation compared to coarser ones Similarly, Tarefder et al concluded that finer-graded mixtures are less prone to rutting Chowdhury et al corroborated these findings in their study of permanent deformation in Superpave mixtures, which also indicated a preference for finer gradations Furthermore, the WestTrack project revealed that coarser-graded mixtures performed poorly, showing increased rutting and fatigue cracking Additional tests by Haddock et al confirmed that fine-graded mixtures outperformed coarse-graded ones in both wet and dry conditions during PurWheel assessments.
The shape and surface texture of aggregates significantly influence rutting resistance in mixtures Crushed manufactured sand enhances rut resistance compared to natural sand, while angular particles improve mechanical stability through better interlock and internal friction Additionally, a rough aggregate surface texture strengthens the mechanical bond and increases voids in mineral aggregate (VMA), leading to improved stability and reduced permanent deformation in the mixture.
2.3.1.4 Comparison of In-Service and Laboratory Results
Pellinen and Witczak (31) found that the unconfined dynamic modulus in the linear viscoelastic range is the most effective measure for predicting in-service rutting when used alongside the Superpave volumetric mixture design method The dynamic modulus (│E*│) of hot mix asphalt (HMA) is inversely related to rutting, indicating that less stiff mixtures are more susceptible to deformation Traditionally, the Superpave Shear Tester (SST) was employed to assess the permanent deformation of HMA mixtures by measuring shear modulus (G*) and shear phase angle However, the current mechanistic-empirical pavement design method now favors the dynamic modulus test over the SST for evaluating the rutting potential of HMA mixtures.
The dynamic modulus and SST tests yield directly related results, indicating that G* decreases at elevated temperatures and lower frequencies, as found by Chowdhury et al in their study on permanent deformation in Superpave mixtures This reduction suggests an increased rutting potential in pavements during hot weather and at slower vehicle speeds, a conclusion that Clyne et al also confirmed through their dynamic modulus testing.
The fatigue resistance of a Hot Mix Asphalt (HMA) mixture refers to its capacity to endure repeated bending without breaking When subjected to traffic loads, vehicles induce bending in the pavement The fatigue properties of an HMA mixture are quantified by the number of load repetitions it can withstand before failure occurs.
EXPERIMENTAL METHODS
Experimental Design
This study aimed to assess how initial density affects the performance of Hot Mix Asphalt (HMA) mixtures and to quantify the relationships between initial HMA density, permeability, and moisture-induced damage Four HMA mixtures were analyzed, with variations in nominal maximum aggregate size (NMAS) of 9.5 mm and 19.0 mm, gradation types (coarse and fine), and densities ranging from 90% to 96% of Gmm After finalizing the mixture designs, test specimens were prepared and evaluated to determine the performance of each mixture.
The research was completed using three groups of test methods; permeability and porosity testing, moisture susceptibility testing, and performance and durability testing
Permeability testing was conducted utilizing the falling head permeability method and the Corelok tests, with equipment and procedures developed by the FDOT The Corelok method assessed the porosity of the mixture samples, while both permeability and porosity tests were performed on specimens prepared in the SGC Additionally, moisture sensitivity testing was carried out using the AASHTO T283 and PurWheel test methods.
The performance and durability of Hot Mix Asphalt (HMA) were evaluated using dynamic modulus and beam fatigue tests on mixture samples in both unconditioned and conditioned states The project utilized two conditioning methods: moisture and air For moisture conditioning, samples were partially saturated and immersed in a 60°C water bath for 24 hours before testing, while air conditioning involved placing specimens in an 85°C forced draft oven for five days Moisture conditioning is known to promote moisture damage, allowing the quantification of its impact on performance and durability through the dynamic modulus and beam fatigue tests.
Materials
To simplify the project and maintain manageable variables, only one binder type was utilized, specifically PG 64-22, chosen for its extensive use across Indiana.
This project utilized natural sand as fine aggregate and crushed limestone as coarse aggregate Comprehensive testing was conducted on the coarse aggregate to evaluate gradation, angularity, and the presence of flat and elongated particles Similarly, the fine aggregate underwent tests for gradation, angularity, and sand equivalent value Detailed test results can be found in Appendix A.
Mixture Designs…
Laboratory mixture designs for Mixtures 2, 3, and 4 were completed in the
The Mixture 1 design at Purdue University's Bituminous Laboratory was successfully executed by a local HMA contractor and implemented on US-52 in Indiana All four mixture designs adhered to the Superpave mixture design specifications, specifically tailored for a traffic level of 2,300,000 Equivalent Single Axle Loads (ESAL), aligning with ESAL Category 2 as per INDOT specifications.
The Superpave mixture design method was employed to identify the optimal binder content for mixtures, targeting 4 percent air voids in samples compacted with the Superpave Gyratory Compactor (SGC) To establish the maximum theoretical specific gravity (Gmm) of each mixture, the AASHTO T209 standard was utilized, alongside AASHTO T166 for measuring bulk specific gravity.
Gravity of Compacted Bituminous Mixtures Using Saturated Surface-Dry
The bulk specific gravity (Gmb) of each sample was determined using "specimens," while the air voids content for all specimens was calculated based on the results of Gmm and Gmb A summary of the mixture designs can be found in Table 3.2, with detailed information for each mixture provided in Appendix B.
Gradation Coarse Fine Coarse Fine
Analysis Procedures
The Analysis of Variance (ANOVA) and Tukey multiple comparison procedure were used to evaluate the test results Both were performed using the SAS statistical software
Statistical analyses were conducted using the ANOVA method, a tool for examining the relationship between a response variable and one or more explanatory variables This analysis aims to assess the statistical relationship between the mean response and the levels of the predictor variables, without assuming any prior relationships The primary goal is to determine if there are significant differences in the response across various levels of the predictor variables.
In the ANOVA output, the F-test corresponds to the goodness of the fit for the relationship analyzed The F-value is defined as:
MSR = Mean squares of regression; and
In the ANOVA process, the p-value represents the likelihood of rejecting the hypothesis that factor levels have no effect on the dependent variable A p-value lower than 0.05 suggests that variations in factor levels significantly impact the response of the dependent variable.
The ANOVA process evaluates each factor individually, whereas the multiple comparison procedure enables simultaneous comparison of all variables This approach offers insights into statistical significance across all levels for each factor.
PERMEABILITY
Falling Head Permeability
The hydraulic gradient, defined as the head loss per unit length, is a crucial factor in assessing permeability According to Darcy's principle, the flow of water is directly proportional to both the hydraulic gradient and the cross-sectional area of the sample, represented by the equation kiA.
Q = flow rate (cm 3 /s); k = coefficient of permeability (or simply permeability) (cm/s); i = hydraulic gradient (cm/cm); and
A = total cross sectional area (cm 2 )
The equation assumes a homogeneous material, with steady state, laminar, one- dimensional flow conditions, and that the fluid is incompressible and the material completely saturated
The Florida Department of Transportation (FDOT) has developed a falling head device to assess the permeability of Hot Mix Asphalt (HMA) mixtures, addressing issues related to permeability and stripping in pavements This method involves measuring the time it takes for a sample to lose a head of water, which is then used to calculate permeability Darcy’s equation is applied in this evaluation process to ensure accurate results.
The coefficient of permeability (k), measured in cm/s, can be determined using the formula k = (aL / (h1 - h2)) * (1/t), where 'a' represents the area of the stand pipe in cm², 'h1' and 'h2' denote the water head at the start and end of the test in cm, and 't' is the duration in seconds over which the head decreases.
L = length of the sample (cm); and
A = cross-sectional area of the sample (cm 2 )
The FDOT falling head approach was chosen for its simplicity and quick testing times to determine the permeability of the samples in this project, utilizing the Virginia Test Method–120 as the standard testing method.
Measurement of Permeability of Bituminous Paving Mixtures Using a Flexible Wall Permeameter.”
Samples were prepared using the SGC and compacted to the heights as specified in the test method For 9.5-mm NMAS mixtures the required height is
The standard specifies that the height of samples should be 38 ± 2 mm for a 12.5-mm NMAS mixture and 50 ± 2 mm for a 19.0-mm NMAS mixture To maintain accurate permeability values, sawing is not permitted, as it can seal the external pores of the sawn surface Instead, additional top plates were added over the bottom plate in the Superpave Gyratory Compactor (SGC) mold to create specimens without the need for sawing.
After compaction, samples were cooled to room temperature for 24 hours and then measured before undergoing vacuum saturation at a residual pressure of 90 ± 2 mm of Hg for 15 ± 2 minutes Following this process, each sample rested underwater for five minutes, and the side surfaces were coated with petroleum jelly to ensure a proper seal between the membrane and the sample.
After conditioning, a sample was placed in a testing device under a confining pressure of 96 ± 7 kPa to prevent water from flowing into the lateral area The test involved measuring the time it took for water to flow from the upper to the lower marks, with any time exceeding 10 minutes indicating that the lower mark represented the water's position at that moment Permeability values were then calculated using Equation 4.2 This process was repeated for each sample until the last three permeability values varied by less than 10 percent, with the final permeability reported as the average of these three measurements.
The permeability results, illustrated in Table 4.1 and Figure 4.2, reveal a logarithmic relationship between air voids content and permeability This indicates significant variations in permeability across the four mixtures, likely attributed to differences in the size and interconnectivity of the air voids present.
Air Voids (%) P e rm eab il it y x10 -5 (c m /s )
Figures 4.3 to 4.6 illustrate the permeability of various mixtures plotted against their air voids content Mixtures 1, 2, and 4 exhibit low permeability, measuring less than 100×10^-5 cm/s when air voids are below 8 percent, with no permeability observed at around 6 percent air voids In contrast, Mixture 3 demonstrates significantly higher permeability, exceeding twice that of the other mixtures.
(200×10 -5 cm/s) at 8 percent air voids content This quickly increases to approximately 850×10 -5 cm/s as the air voids content reaches 10 percent
Mixture 3 is a coarse-graded, 19.0-mm NMAS mixture The permeability data appears to indicate that this mixture has relativity large, interconnected air voids
Air Voids (%) P e rm eab il it y x10 -5 (c m/ s )
Air Voids (%) P e rm eab il it y x10 -5 (c m /s )
Air Voids (%) P e rm eab il it y x10 -5 (c m /s )
Air Voids (%) P e rm eab il it y x10 -5 (c m /s )
In this study, it is established that an air void content of 8 percent is considered a critical threshold, as higher air void levels lead to increased permeability in mixtures, potentially resulting in significant performance issues.
The air voids content for 8 percent is measured at 52x10^-5, 38x10^-5, 200x10^-5, and 62x10^-5 cm/s for Mixtures 1 through 4 However, the 8 percent air voids rule is not appropriate for Mixture 3, as it leads to excessively high permeability in the pavement.
The Florida Department of Transportation (FDOT) identifies a critical permeability value of 125x10^-5 cm/s for the falling head permeameter method, which is essential for ensuring optimal performance of hot mix asphalt (HMA) mixtures To achieve this standard, the study's four mixtures require initial air void contents of 9.0%, 9.3%, 7.4%, and 9.1%, translating to initial densities of 91.0%, 90.7%, 92.6%, and 90.9%, respectively In contrast, Westerman suggests a lower critical permeability value of 10x10^-5 cm/s, which would necessitate air void contents of 6.3%, 6.8%, 5.0%, and 5.7% for the same mixtures, resulting in higher initial in-place densities of 93.7%, 93.2%, 95.0%, and 94.3%.
The permeability data suggest that the initial in-place density necessary for a hot mix asphalt (HMA) mixture to achieve optimal pavement performance is influenced by both the mixture type and the selected critical permeability level For a critical permeability of 125x10^-5 cm/s, Mixture 3 must be compacted to an initial density approximately 2% higher than the other mixtures Conversely, when the critical permeability is reduced to 10x10^-5 cm/s, Mixture 3 requires an average initial density that is 1.3% higher As the critical permeability level decreases, the required initial densities for all mixtures increase, highlighting the significance of carefully selecting the appropriate critical permeability level.
CoreLok
The air void content of each sample was assessed using the AASHTO T166 method and CoreLok equipment CoreLok serves as an alternative technique for measuring the bulk specific gravity of compacted hot mix asphalt (HMA) samples and for calculating their porosity.
Cooley et al (13) compared the Gmb results from the CoreLok and AASHTO T166 methods, finding significant differences in outcomes for coarse-graded mixtures, which varied by mixture type and air voids content They concluded that CoreLok provides a more accurate measurement of air voids, particularly at higher contents, as it does not overestimate Gmb values like AASHTO T166 The discrepancies arise from the greater interconnectivity of air voids in coarse-graded samples; when using AASHTO T166, water can drain from the sample after removal from the water bath, leading to a lower surface-saturated dry (SSD) mass and an inflated Gmb value that underrepresents actual air voids content.
To determine the Gmb values using the AASHTO T166 method, first weigh a dry sample and record its mass in grams as Mass A Next, measure the submerged mass (C) and the SSD mass (B) With these measurements, utilize Equations 4.3 and 4.4 to calculate the Gmb and the amount of water absorbed by the sample.
According to AASHTO T166, if the water absorbed exceeds two percent, the AASHTO T275 method, “Bulk Specific Gravity of Compacted Bituminous
Mixtures Using Paraffin Coated Specimens,” must be used instead of T166 in order to obtain the Gmb
The CoreLok device serves as an alternative to AASHTO T166 for determining the density of specimens In this method, the sample is first vacuum sealed in a plastic bag after measuring its dry mass Once the air is evacuated and the bag is sealed, it is submerged in water, allowing for the recording of the combined mass of the sealed sample and bag Afterward, the bag is opened underwater, and the submerged mass is recorded again The specific gravity (Gmb) and porosity of the sample are then calculated using established equations.
A = mass of dry specimen in air (g);
B = mass of dry sealed specimen (g);
E = mass of sealed specimen under water (g); and
FT = apparent specific gravity of plastic bag at 25C
Porosity is calculated using the formula ρ = (ρ1 - ρ2) / ρ1 × 100%, where p represents porosity in percentage, ρ1 is the vacuum-sealed density of the specimen in grams per cubic centimeter (g/cm³), and ρ2 is the apparent or maximum density, which includes the volume of inaccessible air voids This apparent density is measured as the density of the vacuum-sealed specimen after it has been submerged in water.
Table 4.2 displays the results from the AASHTO T166 and CoreLok tests, while Figure 4.7 illustrates the comparison of air void contents derived from both methods The findings indicate that the air void contents measured using the AASHTO T166 method are generally higher than those obtained through the CoreLok technique.
Comparison in air voids from CoreLok and
FIGURE 4.7 Comparison of CoreLok and AASHTO T166
Figures 4.8 and 4.9 illustrate the correlation between porosity and air voids content as measured by the CoreLok method and AASHTO T166, respectively The data reveals significant insights regarding Mixtures 1, highlighting the relationship between these two critical parameters in asphalt mixtures.
3, and 4 have an approximate one-to-one correspondence between air voids content and porosity However, Mixture 2 (9.5-mm NMAS, fine gradation) has a lower porosity values at every air voids content
TABLE 4.2 Bulk Specific Gravity, Porosity and Absorption Results
FIGURE 4.8 CoreLok Porosity and Air Voids
FIGURE 4.9 CoreLok Porosity and AASHTO T166 Air Voids
Statistical Analysis of Results
To assess the permeability of the four HMA mixtures, a falling head permeameter was employed, testing each mixture at four distinct air voids contents Statistical analysis was conducted using SAS software, with a 95% significance level, to identify the impact of main experimental factors and potential interaction effects An initial ANOVA test was performed to determine the significance of these factors and explore any interaction effects that may have occurred.
Here is a rewritten paragraph that captures the essential information in a coherent and SEO-friendly manner:"To investigate the differences in factor levels for each of the three primary factors - gradation, NMAS, and air voids content - the Tukey’s Studentized Range procedure was employed Prior to conducting statistical analyses, air voids content and permeability were transformed into natural logarithm values to ensure accuracy Gradation was categorized as either coarse-graded (0) or fine-graded (1), while permeability values were measured in units of 10^-5 cm/s and air voids content was represented as voids in the total mixture (VTM)."
The ANOVA results are shown in Table 4.3 and the Tukey’s Studentized
Range results in Table 4.4 The ANOVA results indicate that NMAS, gradation, air voids content, and their interactions are significant in predicting HMA mixture permeability Air voids content (density) appears to have the most influence followed by NMAS and finally gradation
(Dependent variable: permeability) F-value = 19.18 Pr > F: F NMAS 1 101899.43 101899.43 14.54 0.0008 Gradation 1 48667.92 48667.92 6.94 0.0145 NMAS×Gradation 1 33656.73 33656.73 4.80 0.0384 VTM 3 733030.54 244343.51 34.86 < 0.0001 NMA×SVTM 3 208880.48 69626.83 9.93 0.0002 Gradation×VTM 3 109655.51 36551.84 5.21 0.0065 NMAS×Gradation×VTM 3 124286.93 41428.98 5.91 0.0036
From Table 4.4, one observes that 19.0-mm coarse-graded mixture at 10 percent air voids content had the highest permeability and the fine-graded mixture with a 9.5-mm NMAS and 4 percent air voids content was the least permeable Also, on average, the permeability of fine-graded mixtures is 55 percent less than that of the coarse-graded mixtures The 9.5-mm NMAS mixtures are 73 percent less permeable than the 19.0-mm mixtures Specimens at 10 percent air voids had the highest permeability As the air voids contents go to 8, 6, and 4 percent, the permeability is reduced 78, 96, and 100 percent, respectively However, the differences in permeability at 8, 6, and 4 percent are not statistically different This suggests that these HMA mixtures may be impervious to moisture at air voids contents lower that 8 percent, but at higher air voids can became significantly permeable
In an additional attempt to investigate the relationships among the project factors, a regression analysis was completed The resulting equation is:
NMAS = Nominal Maximum Aggregate Size (mm);
VTM = Voids in the Total Mixture; and
Gradation = 0 for coarse-graded, 1 for fine-graded
The equation has an excellent goodness of fit (adjusted R 2 of 0.93) The air voids content appears to have the most effect on permeability
The relationship between air voids content and permeability was further analyzed using regression techniques, with air voids content as the sole independent factor This analysis yielded a predictive model for HMA mixture permeability, expressed as VTM k = b * e^(a * VTM), where a and b are regression constants, and VTM represents air voids content The regression constants and goodness of fit estimates for each mixture are presented in Table 4.5, with R^2 values indicating that air voids content alone can reasonably predict HMA mixture permeability.
As discussed in Chapter 2, field tests of permeability have shown to correlate well to laboratory permeability test results This suggests that it might be possible to control permeability during HMA pavement construction Mixture permeability can be determined in the laboratory after the mixture design is completed and with an acceptable field permeability value selected, specification limits could be established for initial, in-place permeability Field permeability testing could be completed on the compacted pavement for permeability quality control purposes Cores taken from the completed pavement for density analyses could be tested for permeability in the laboratory and used for quality assurance
Given the relationship between air voids content (density) and HMA mixture permeability, it may be possible, given further study, to use field permeability testing to control in-place HMA pavement density This would be significant in that destructive testing would no longer need to be accomplished in order to establish in-place density
The porosity data was further analyzed using ANOVA and Tukey procedures, with results summarized in Tables 4.7 and 4.8 Notably, ANOVA results identified gradation and air voids content as significant variables influencing porosity This finding was corroborated by the Tukey grouping, which revealed distinct differences between coarse and fine gradations, as well as varying air voids contents In contrast, the nominal maximum aggregate size (NMAS) did not exhibit significant differences Furthermore, Mixtures 1, 3, and 4 demonstrated similar porosity values at identical air voids content levels, as evident in Figures 4.8 and 4.9.
GLM Procedure (dependent variable is rut depth)
Source DF Type III SS Mean
Square F value Pr > F Gradation 1 3.8678 3.8678 5.38 0.0310 VTM 3 166.8741 55.6247 77.39 < 0.0001 Gradation×VTM 3 0.9039 0.301 0.42 0.7412 NMAS 1 0.3804 0.3804 0.53 0.4753 Gradation×NMAS 1 1.1226 0.1226 1.56 0.2258 NMAS×VTM 3 5.9297 1.9766 2.75 0.0697 Gradation×NMAS×VTM 3 0.2687 0.0896 0.12 0.9445
Tukey Group Mean Porosity (%) No of Observations Gradation
B 5.6944 18 Fine Tukey Group Mean Porosity (%) No of Observations NMAS (mm)
A 5.9706 17 9.5 Tukey Group Mean Porosity (%) No of Observations VTM (%)
Equation 4.9 shows the regression results for porosity as a function of air voids content measured in the CoreLok
A predictive equation for porosity (p) has been established, expressed as VTM p= + − (Eq 4.8), with a high R-squared value of 0.95, indicating a strong correlation This result suggests that porosity can be effectively predicted using air voids content, similar to permeability However, the findings also reveal that gradation and the interaction between nominal maximum aggregate size (NMAS) and gradation have a significant impact on porosity, implying that porosity encompasses more information than previously thought.
HMA mixture’s permeability than does air voids content Porosity not only accounts for air voids content, but may also account for the size and interconnectedness of the air voids It may be that porosity is more important to the performance of an HMA mixture than air voids content, but the results from this project are not extensive enough to make any conclusions Additional work on porosity needs to be accomplished.
MOISTURE SUSCEPTIBILITY
AASHTO T283…
The AASHTO T283 test method is used to evaluate the moisture susceptibility of HMA mixtures and was established in 1985 based on the Modified Lottman Test, which is a combination of the Lottman and the Root-Tunnicliff tests The T283 test consists of producing six specimens having air voids contents between 6 and
8 percent This high air voids content helps to accelerate moisture damage in the cores Two groups of three specimens are used The first group, without any type of conditioning, is the control group The second group is moisture saturated to
70 to 80 percent by applying vacuum saturation These specimens are then further conditioned by placing them in a water bath at 60C for 24 hours After conditioning, the Indirect Tensile Strength (ITS) test is performed at 25C with a loading rate of 50 mm/minute The indirect tensile strength of each sample is determined and the average values for the conditioned and control groups are calculated The ratio of the average conditioned and unconditioned values is calculated and multiplied by 100 to determine the Tensile Strength Ratio (TSR) For most user agencies, a minimum acceptable TSR value for an HMA mixture is
80 percent The ITS equipment is shown in Figure 5.1.
After mixing and short-term aging the HMA mixtures for two hours at 145C, the samples were compacted in the SGC The specimens are 150 mm in diameter and 95 mm in height The day after compaction (24±3 hours), the samples were tested according to AASHTO T166 to determine their bulk specific gravities
(Gmb) The six samples were separated into two groups of three so that the average air voids content of the two groups was approximately equal
FIGURE 5.1 Indirect Tensile Strength Equipment
After compaction and determination of the Gmb, the unconditioned samples were placed in a plastic bag and submerged in a 25C water bath for two hours after which time they were tested in the ITS apparatus The conditioned samples were first vacuum-saturated at a pressure of 250 to 660 mm of Hg until reaching a saturation level of 70 to 80 percent The samples were then placed in a 60±1C water bath for 24±1 hours, followed by a 25±0.5C water bath for 2 hours At the end of the two hour period the conditioned samples were tested
The indirect tensile breaking apparatus uses two steel loading strips to apply a load along the diameter of the specimen Once the sample has broken, the tensile strength is calculated by the equation: tD
P = maximum load (kN); t = specimen thickness (average of three measurements) (cm); and
D = specimen diameter (average of three measurements) (cm)
The TSR is calculated according to the equation:
S1 = average tensile strength of three conditioned samples (kPa); and
S2 = average tensile strength of three unconditioned samples (kPa)
INDOT specifications require HMA mixtures to have a minimum TSR of 80 percent
For Mixture 1, the mixture actually placed in the field, the mixture design formula indicated a TSR of 97 percent Mixtures 2, 3, and 4 were prepared and tested in the laboratory for this project The results are presented in Table 5.1 According to INDOT specifications, none of the four mixtures is considered susceptible to moisture
Figure 5.2 shows the four mixtures after conditioning and testing As can be seen, some of the asphalt binder has been stripped from the aggregates
However, these samples do not appear to be moisture damaged beyond what might be considered normal for the AASHTO T283 test
TABLE 5.1 Moisture Susceptibility Test Results
PurWheel…
The Purdue laboratory wheel tracking device (PurWheel) was developed to recreate the conditions associated with rutting and stripping It simulates field conditions of high moisture, high temperature, and traffic, all of which can
AAHSTO T283 MIXTURE 4 CONDITIONED SAMPLE contribute to rutting and stripping The load is applied with a pneumatic tire typically inflated to produce a contact pressure of 690 kPa Pan and White’s report (25) gives a detailed description of the PurWheel parameters of testing In this project, the PurWheel was used to evaluate the stripping potential of the four mixtures at each of the four air voids contents in hot, moist conditions
After mixing and short-term-oven aging laboratory prepared HMA mixture at 135C for four hours, the PurWheel samples were compacted at 145C in a linear compactor to achieve the target air voids of 4, 6, 8, and 10 percent The linear compactor (Figure 5.3) includes a rectangular mold attached to an air cylinder, a set of steel plates, a loading frame with a steel roller and hydraulic ram to apply a compaction force (hydraulic pressure supplied by an electric powered hydraulic pump) Once the sample is compacted and cooled, it is cut into halves After overnight drying, the length and width dimensions are taken at three different points and the thickness measured at eight different points With the average dimensions the volume is calculated and using this in conjunction with the theoretical maximum density of the mixture, the air voids content is determined
Typical specimen dimensions are 290 mm wide and 310 mm long The thickness varies with the NMAS of the mixtures For 9.5-mm NMAS mixtures a
50 mm thick test specimen is used For 19.0-mm NMAS mixtures, a 63.5 mm thick specimen is used
Here is a rewritten paragraph that condenses the original text into key sentences, optimized for SEO:"The PurWheel apparatus was used to evaluate the moisture susceptibility of slab specimens, with a contact pressure of 690 kPa and a wheel speed of 33±2 cm/s Prior to testing, samples were conditioned in 50C water for 20 minutes and then submerged during the test The computer recorded wheel passes, deformation, and elapsed time, with the test ending automatically at 20 mm rut depth or 20,000 wheel passes The testing process involved securing the specimens in the mold with plaster-of-paris, allowing them to dry for six hours before commencing the test."
FIGURE 5.5 PurWheel Samples after Testing
The PurWheel data are shown in Table 5.2 The rut depth is the compactive deformation, or downward deformation measured in the sample during the test Total deformation is measured at the end of the test and includes both the compactive and upward lift deformation
Figures 5.6 through 5.9 show the PurWheel results for each mixture In these plots, it is seen that none of the samples had a stripping inflection point as defined earlier (Chapter 2) It is therefore concluded that none of the four mixtures is susceptible to moisture damage This agrees with the AASHTO T283 results However, since none of the four mixtures has a stripping inflection point at any of the four air voids contents, no analyses can be completed to determine if the moisture susceptibility is dependent upon the air voids content It may be possible that the air voids contents of these particular mixtures must be higher than 10 percent before the mixtures begin to become more susceptible to moisture
(1) Sample was damaged and discarded
FIGURE 5.6 Mixture 1 (9.5-mm NMAS, Coarse-graded) PurWheel Results
FIGURE 5.7 Mixture 2 (9.5-mm NMAS, Fine-graded) PurWheel Results
FIGURE 5.8 Mixture 3 (19.0-mmm NMAS, Coarse-graded) PurWheel Results
FIGURE 5.9 Mixture 4 (19.0-mmm NMAS, Fine-graded) PurWheel Results