A study was performed to determine the influence of material properties on the thermal cracking performance of hot mix asphalt (HMA), and to determine the ability to predict thermal cracking from pavements of known field performance. The testing device used to measure the HMA properties was the thermal-stress, restrained-specimen test (TSRST), and the device used to measure the binder properties was the bending beam rheometer (BBR). The laboratory study was conducted to determine the variability of test results as an influence of 1) asphalt cement stiffness, 2) asphalt cement quantity, 3) mixes with various aggregate qualities, 4) aging, and 5) the presence of hydrated lime. The influence of the asphalt cement stiffness was the single largest factor that controlled the test results.

The work detailed in this report represents a continuation of the research performed in phase one of this national pooled fund study. A number of significant contributions were made in phase two of this comprehensive research effort. Two fracture testing methods are proposed and specifications are developed for selecting mixtures based on fracture energy criteria. A draft SCB specification, that received approval by the ETG and has been taken to AASHTO committee of materials, is included in the report. In addition, alternative methods are proposed to obtain mixture creep compliance needed to calculate thermal stresses. Dilatometric measurements performed on asphalt mixtures are used to more accurately predict thermal stresses, and physical hardening effects are evaluated and an improved model is proposed to take these effects into account. In addition, two methods for obtaining asphalt binder fracture properties are summarized and discussed. A new thermal cracking model, called "ILLI-TC," is developed and validated. This model represents a significant step forward in accurately quantifying the cracking mechanism in pavements, compared to the existing TCMODEL. A comprehensive evaluation of the cyclic behavior of asphalt mixtures is presented, that may hold the key to developing cracking resistant mixtures under multiple cycles of temperature.

Introduction and Research Approach -- Findings -- Interpretation, Appraisal, and Applications -- Conclusions and Recommendations -- References -- Appendixes.

The current means of evaluating the low temperature cracking resistance of HMA relies on extensive test methods that require assumptions about material behaviors and the use of complicated loading equipment. The purpose of this study was to develop and validate a simple test method to directly measure the cracking resistance of hot mix asphalt under field-like conditions. A ring shape asphalt concrete cracking device (ACCD) was developed. ACCD utilizes the low thermal expansion coefficient of Invar steel to induce tensile stresses in a HMA sample as temperature is lowered. The results of the tests of the notched ring shaped specimens compacted around an ACCD Invar ring showed good repeatability with less than 1.0°C (1.8°F) standard deviation in cracking temperature. A laboratory validation indicated that ACCD results of five mixes correlate well with thermal stress restrained specimen test (TSRST) results with the coefficient of determination , r2 = 0.86. To prepare a sample and complete TSRST measurement, it takes minimum 2-3 days. For ACCD, two samples can be easily prepared and tested in a single day with a small test set-up. The capacity of ACCD can be increased easily with minimal cost to accommodate a larger number of samples. Among factors affecting the low temperature performance of HMA, the coefficient of thermal expansion (CTE) of aggregate has been overlooked for years. A composite model of HMA is proposed to describe the low temperature cracking phenomenon. Due to the orthotropic and composite nature of asphalt pavement contraction during cooling, the effects of aggregate CTE is amplified up to 18 times for a typical HMA. Of 14 Ohio aggregates studied, the maximum and the minimum CTEs are 11.4 and 4.0 x 10-6/°C, respectively. During cooling, the contraction of Ohio aggregate with high CTE can double the thermal strain of asphalt binders in the asphalt mix and may cause asphalt pavement thermal cracking at warmer temperature.

This report describes the thermal stress restrained specimen test (TSRST), which was selected to evaluate the low-temperature cracking resistance of asphalt concrete mixtures. The TSRST system includes a load frame, step-motor-driven load ram, data acquisition hardware and software, temperature controller, and specimen alignment stand. An experiment design that considered a range of mixture and test condition variables was developed to evaluate the suitability of TSRST for characterizing low-temperature cracking resistance of asphalt concrete mixtures. Four asphalts and two aggregates were selected for the experiment. The mixture variables included asphalt type, aggregate type, and air voids content; the test condition variables included specimen size, stress relaxation, aging, and cooling rate.

This report examines the feasibility of using the thermal stress restrained specimen test to evaluate low temperature cracking in asphalt pavement mixes. Data were collected from laboratory and field evaluations. Various mixing, aging, and compaction methods were used to prepare test samples with materials obtained from two Wyoming Department of Transportation (WYDOT) highway projects. Field data were obtained from two recently built test sections and compared with laboratory test results. Pavement condition surveys quantified low temperature cracking of both test sections after one winter. Temperature data for the project sites also were collected. Pavement condition and temperature data were compared to results from the thermal stress restrained specimen test. The thermal stress restrained specimen test was effective in testing asphalt pavement mixes. However, test results indicated that lab prepared samples did not closely simulate field samples. Also, comparisons of lab results with field conditions were performed. However, it is recommended that a more comprehensive analysis be performed after test sections have been in service for a few years.

Creep compliance function is one of the fundamental properties of viscoelastic materials. In asphalt research, creep compliance is used in the prediction of low-temperature cracking, which is a prevalent distress in asphalt pavements in northern parts of the United States and Canada. The thermal cracking prediction model included in the current version of the Mechanistic-Empirical Pavement Design Guide requires the creep compliance of hot mix asphalt (HMA) materials as a primary input. The current AASHTO procedure for determining creep compliance of HMA is based on the indirect tension (IDT) test. The IDT test is performed on relatively thick cylindrical specimens, which makes this test unsuitable for field cores from thin layers and construction lifts. In addition, thick specimens do not allow for investigation of the gradual aging that occurs in asphalt layers due to oxidation and volatilization. This paper outlines the procedure that uses the bending beam rheometer (BBR) as an alternative to the IDT test to determine the low-temperature creep compliance of HMA using thin beam specimens. Twenty different HMA mixtures are tested at three temperature levels using both the IDT and the BBR instruments. The analysis of test results is presented in two parts. First, the variability in creep compliance values obtained from the IDT and the BBR tests is discussed. Next, the artificial neural networks (ANNs) are trained to (1) predict IDT results from BBR measurements using design parameters of HMA mixture and testing temperature as the model inputs and (2) backcalculate HMA creep compliance from the binder creep compliance and vice versa. It is concluded that testing HMA beams can be used instead of IDT specimens for the low-temperature characterization of HMA mixtures, and the ANN can be successfully used for the backcalculation of the creep compliance measured by both IDT and BBR devices.