Validation of Laboratory Cracking Tests for Field Top-down Cracking Performance

Abstract

Top-down cracking (TDC) has been widely reported as a primary mode of distress in asphalt pavements. Currently, there is no consensus on a practical laboratory cracking test that can reliably predict an asphalt mixture’s resistance to TDC. Age hardening of asphalt binder has been identified as a critical factor that contributes to the development of TDC. Thus, laboratory specimens should be conditioned prior to testing to simulate the aging of pavements in-service. However, there is insufficient guidance on an appropriate aging protocol to simulate the critical field aging condition when TDC starts to develop. To address these needs regarding TDC, this study had two objectives: 1) identify the field aging condition when TDC starts to develop, then determine a laboratory aging protocol to condition asphalt mixtures for TDC evaluations; 2) evaluate the ability of laboratory cracking tests to identify mixtures resistant to TDC based on measured cracking performance in the real pavements using real loading conditions. In 2015, National Center for Asphalt Technology (NCAT) initiated an experiment on the NCAT Test Track (hereinafter referred to as the NCAT TDC experiment). Seven surface mixtures with a range of cracking susceptibilities were designed and constructed as a 1.5-inch surface layer on top of highly modified base and binder layers. After construction, truck traffic was applied to the test sections. Field cracking has been monitored weekly throughout the experiment. Six laboratory cracking tests were selected by the experiment sponsors: Energy Ratio (ER), the Texas Overlay Test (TX-OT), the NCAT Modified Overlay Test (NCAT-OT), the Semi-Circular Bend test (SCB), the Illinois Flexibility Index Test (I-FIT), and the indirect tensile asphalt cracking test (IDEAL-CT). A literature review was first conducted to determine the field aging condition for evaluating TDC. Data from a number of existing pavements showed that TDC typically initiated after approximately 70,000 cumulative degree-days (CDD). Both plant loose mixtures and field cores from five field projects in three states were used to determine an aging protocol that was representative of this critical CDD. The loose mixtures from each project were conditioned using four loose mixture aging protocols, including a 6-hour, 135°C protocol, a 12-hour, 135°C protocol, a 24-hour, 135°C protocol, and a 5-day, 95°C protocol. Asphalt binder extracted from conditioned loose mixtures and field cores were tested using a dynamic shear rheometer (DSR), a bending beam rheometer (BBR), and Fourier Transform Infrared Spectroscopy (FT-IR). Test results showed that the 5-day, 95°C protocol was most representative of 70,000 CDD of field aging, and an 8-hour, 135°C and 5-day, 95°C protocol was likely to achieve an equivalent aging level. Thus, the more practical 8-hour, 135°C protocol was selected for the remaining study to simulate 70,000 CDD of field aging. Two candidate critical aging (CA) protocols for loose mixture aging, 5 days at 95°C and 8 hours at 135°C, were further evaluated using four NCAT TDC experimental mixtures. For each mixture, the plant loose mixtures were conditioned using both protocols. Field cores were also collected annually after construction. Mixture and binder properties of the conditioned loose mixtures and field cores were tested and compared using I-FIT, IDEAL-CT, small-specimen Asphalt Mixture Performance Tester (AMPT) cyclic fatigue, linear amplitude (LAS), double-edge-notched tension (DENT), DSR, BBR, and FT-IR tests. Test results showed inconsistent trends when comparing two candidate CA protocols. Only I-FIT results indicated no significant difference between two protocols. Most of the asphalt binder and mixture properties indicated that two candidate CA protocols yielded a more severe aging level than 4 years field aging in Alabama. Seven NCAT TDC mixtures, including both plant-produced and laboratory-prepared mixtures, were conditioned and tested using the six selected cracking tests. Test results were utilized to conduct comparison, sensitivity, and correlation analyses. No significant differences were identified between plant mixture and laboratory mixture results after critical aging for all the mixture cracking test parameters. In addition, all the cracking test parameters of I-FIT, IDEAL-CT, and both OT tests were sensitive to aging for plant and laboratory mixtures. ER and Jc were not sensitive to aging for both plant mixtures and laboratory mixtures. Some cracking test parameters were found to be sensitive to air voids, recycled asphalt mixture, and modified binder for certain conditions, but not in every case considering combinations of plant- and lab-produces mixtures with short-term and critically aged conditioning. Based on mean value analysis, only NCAT-β was sensitive to all the influence factors for plant mixture at both aging conditions. The cracking test parameters of both OT tests were sensitive to all the influence factors for laboratory mixtures. Strong positive linear correlations exist among NCAT-Nf, TX-Nf, FI, and CTIndex results, and there are strong power relationships between the Nf and β parameters for both OT tests. The field cracking performance of seven test sections were utilized to evaluate the laboratory cracking test results. After approximately 15 million ESALs and 73,728 CDD, sections N1, N2, N5, and N8 exhibited a range of cracking severity and extent. Section N8 with 5% RAS had the highest cracking percentage of 70.7% of the lane area. No measurable cracks were found in sections S5, S6 or S13. The Pearson’s correlation coefficient (rp) results indicated that NCAT-β and TX-β generally have a very strong linear correlation with the field cracking performance for both plant mixtures and laboratory mixtures at both aging conditions. DCSEMin, TX-Nf, TX-β, NCAT-Nf, NCAT-β, FI, and CTIndex correctly identified mixture N8 as the most susceptible to TDC for both plant mixtures and laboratory mixtures at both aging conditions. ER did not identify mixture N8 as the most susceptible to TDC. No cracking test parameter discriminated uncracked sections from cracked sections for both plant mixture and laboratory mixture at both aging conditions. The test sections will remain in place for additional trafficking and environmental aging on the NCAT Test Track. This ongoing work will help further validate the aging protocols and identify the best test for determining the TDC susceptibility of asphalt mixtures for possible use in mix design and quality assurance.