Reflective Cracking Phase III
Phase III Test
Full-scale tests began on June 3, 2014. After 6350 cycles, the overlay was completely separated and the test concluded on July 30, 2014.
The objective of Phase III Test was to identify an appropriate mitigation technique to retard thermally-induced reflective cracking and to evaluate its effectiveness under full-scale test conditions.
Phase III test pavement consisted of control and interlayer sections. The interlayer section was designed as 1-in.-thick strain relieving HMA plus a 4-in.-thick overlay using standard FAA P-401 (PG 64-22) material. It was believed that this overlay thickness was sufficient above the softer interlayer to prevent rutting, slippage cracks, and shoving in the overlay areas where aircraft accelerate, decelerate, or make sharp turns. On the control section, the entire 5-in. overlay consisted of the same FAA P-401 material. The strain relieving interlayer was constructed with a fine graded, polymer modified asphalt (PMA), asphalt-rich mixture. The fatigue, fracture, and viscoelastic performance of designed interlayer mix were verified through comprehensive laboratory tests.
| Binder (%)
| Maximum Specific Gravity
|Bulk Specific Gravity
| Air voids (%)
| VMA (%)
Prior to overlay construction, the milled concrete slab surface was thoroughly washed to remove all dirt and dust. To prevent interface slippage and secondary cracks, a thin tack coat of straight PG 64-22 asphalt was applied on the dry surface of two 12-in.-thick, 15- by 15-ft concrete slabs. To facilitate instrumentation, the overlay was built in multiple lifts (1.0, 1.5, and 2.5 in.). Between lifts, time was cautiously balanced to allow for the application of a tack coat, placement of instrumentation sensors, and an adequate mix temperature to achieve the desired density. Thermocouples were embedded at various depths and locations to monitor the overlay temperature. There were unexplained spikes and dips in the temperature ranges that might have contributed to the tearing appearance in the interlayer surface. However, after compaction, the interlayer was smooth and stable. The complete HMA overlay was 30- by 5-ft with a 2-ft gap in between.
During the overlay paving, H-type asphalt strain gages (EG) were placed at the mid-depth (2.5 in.) of the overlay. Phase II test suggested potential interference between the embedded sensors and crack propagation. It was speculated that such a negative impact would be more prominent if the H-type sensors were installed in the 1-in.-thick interlayer. Alternatively, Fiber-Bragg Grating optical strain sensors were chosen because of their slim profile. One installation lesson learned was to cover the embedded sensors with material close to the lift thickness (i.e., 1-in.) and then use the screed of the paver to strike off the excess HMA to the proper depth and grade. As a result, instrumentation damage could be reduced to a minimum. After the overlay construction, surface strain gages (SG) were installed at various locations on the surface. Note that all sensors were installed on the “best guess” crack path, which was directly above and perpendicular to the concrete joint.
RC Phase III Drawing
Temperature variations were approximated by a haversine waveform describing the relationship between the joint opening and cycle time:
where t is the time of interest, D is the amplitude of joint opening, T is the cycle time, and R is rest period, which was included at the end of each loading cycle to allow the HMA materials to relax. In Phase III testing, the joint opening (D) was set at 12 mils which corresponded to a large temperature drop (17oF) at the overlay bottom in the field. Each haversine loading cycle began with a loading time of 75 sec, once the actuators reached the maximum horizontal displacement (joint opening), a 75-sec unloading was executed and then followed by a rest period of 600 sec to allow the overlay to relax.
Strain responses were used as the primary tool to determine crack initiation. In addition, a visual examination of the test pavement was conducted multiple times a day to identify cracks off the “best guess” crack path and trace the extension of existing cracks.
For demonstration purposes, strain responses from one sensor at each depth were plotted, as shown in Figure 1. All strain sensors except SG11 recorded that the tensile strain continuously grew at a slow rate and then experienced a rapid increase. A surface inspection revealed that SG11 was 0.5 in. away from the crack (Figure 1a), but the other sensors were directly atop the crack (Figure 1b). The abrupt drop in the strain response from SG11 was most likely caused by the strain energy released during crack formation. The intact interlayer shifted the maximum tensile stresses from the overlay bottom to the interlayer bottom.
Figure 1. Crack initiation.
Because the control and interlayer sections had a different overlay thickness (5-in. vs. 4-in.), the relationship between normalized crack length and number of cycles is shown in Figure 2a. For both control and interlayer sections, the crack length developed gradually at the beginning, and its propagation rate became higher and higher with the crack growth. At the middle of the overlay, the crack propagation stage underwent a transition from quiescent to aggressive. It is also evident in Figure 2a that the interlayer section required constantly more load repetitions to penetrate the crack through the overlay. The strain relieving interlayer enhanced the reflective cracking resistance of the overlay. The effectiveness of strain relieving interlayer was more pronounced at an early stage of crack propagation and slowly diminished as the crack length increased.
Figures 2b and 2c show the overlay failure for the interlayer and control sections, respectively. On the interlayer section, the crack started above the intact interlayer. At first, the crack propagated upward, indicating fracture Mode I dominance. Once this vertical crack reached the 0.25-in. benchmark, it progressed in the horizontal direction, departing from the concrete joint until it reached the 1.0-in. benchmark. In the final stage, the vertical crack progressed gradually on its initial path, and the horizontal crack started to deviate at an angle toward the surface. After 3117 cycles, the diagonal crack made an appearance on the overlay surface, 3.5 in. offset from the joint. At the same time, the vertical crack barely reached the 0.5-in. benchmark above the intact interlayer. It can be concluded that the fracture on the interlayer section was a mixed mode. On the control section, the reflection crack initiated from the overlay bottom and progressed on its upright track. The horizontal tensile stresses seemed to be the principal driving force for the overlay fracture in this case.
(b) Interlayer section (c) Control section
Figure 2. Crack propagation.
Full-scale tests were conducted on HMA overlays with and without an interlayer. Side-by-side comparisons of overlay performance led to the following conclusions:
- The strain relieving interlayer enhanced the reflective cracking resistance of an HMA overlay. The effectiveness of strain relieving interlayer was more pronounced at an early stage of crack propagation and slowly diminished as the crack length increased.
- Inclusion of a 1-in.-thick interlayer between existing concrete slabs and the overlay extended overlay service life up to 15%. The intact interlayer had prevented spalling and moisture infiltration at the joint and therefore prolonged the structural integrity of the pavement.
- To realistically characterize the development of bottom-up reflection cracks, both mixed-mode fracture and crack channeling should be considered.
- Yin, H. and Ishee, C. (2015) "Evaluation of Strain Relieving Interlayer to Retard Thermally-Induced Reflective Cracking," Journal of the Association of Asphalt Paving Technologists, Volume 84.
- Yin, H. (2017) Reflective Cracking Phase III Comprehensive Report.