|FHWA > HfL > Technology Partnerships > Final Report Phase 1 > Asphalt Binder Cracking Device to reduce Low Temperature Asphalt Pavement Cracking|
Asphalt Binder Cracking Device to reduce Low Temperature Asphalt Pavement Cracking
Task 1. Refinement of Test Procedures
The objective of Task 1 was to refine test procedures and equipment to improve the precision of the ABCD test and shorten the test time. Improving the precision of the test was investigated by modifying the silicone sample mold and trimming process. The effect of isothermal conditioning prior to testing was also reviewed in regard to test precision. The shortening of the test time was investigated by increasing the test cooling rate, improving pouring methods and eliminating or simplifying the trimming process.
Through many trials of cooling rates for binders with warm cracking temperatures, it was determined that the optimum cooling profile for the ABCD chamber consists of cooling from 20°C to 0°C in 30 minutes followed by cooling from 0°C to –60°C at the desired cooling rate.
1A. Cooling Rate
The effect of cooling rate on ABCD cracking temperature was investigated. In investigations prior to Phase 1, only the 10C/hr rate was used with few exceptions. For the Phase 1 study, four unmodified asphalts (EM1 PAV, EM2 PAV, ABM–1 unaged, and AAM–1 unaged) and two SBS modified binders (OD1 unaged and OD3 unaged) were tested with five different cooling rates (1, 3, 10, 20, and 40°C/hour). The results are shown in Table 5 and Figure 1. For each sample, quadruplicate samples were tested at each cooling rate except two triplicate sample tests (EM2 @ 20°C/hr and AAM–1 @1.0°C/hr).
The two polymer modified binders show trends similar to those seen in the unmodified binders. There is a clear trend between the ABCD cracking temperature and the cooling rate. As the rate of cooling increases, the cracking temperature of the asphalt specimens increases (specimen cracks at warmer temperatures). This agrees with the viscoelastic nature of asphalt binders. At a higher cooling rate, the rate of thermal stress accumulation is faster than the rate of stress relaxation, leading to rapid stress development and early/warm fractures. Except AAM–1 unaged, the rate increase from 1°C/hr to 3°C/hr resulted in a significant increase in cracking temperature (4.5°C on average). However, subsequent rate increases greater than 3°C/hr did not affect cracking temperature as significantly as at rates less than 3°C/hr. For all cooling rates, the rank of the thermal cracking resistance of the four asphalt binders remained the same. The standard deviations of four ABCD measurements were less then 1.00°C for all cooling rates except the highest cooling rate, 40°C/hr. The results of the cooling rate tests suggest that the use of a 20°C/hour cooling rate provides the same repeatability of ABCD cracking temperatures as a 10°C/hour cooling rate for both unmodified and polymer modified binders.
The relationship between ABCD cracking temperature and cooling rate can be summarized by a multiple linear regression as given below.
Tcr = –38.88 + 0.74*(Cooling Rate) + (Binder Constant) (r2 = 0.94)
The measured ABCD cracking temperatures and the predicted values using the above predictive equations are plotted in Figure 2. As cooling rate increases by 1°C/hour (faster cooling), the ABCD cracking temperature increases (warms) by 0.074 °C. Similar results will be obtained for log–transformed cooling rate data.
More asphalt binders were tested at the 20°C/hour cooling rate and compared with the results of 10°C/hour cooling rate test as summarized in Table 6. Differences are less than 0.5°C except AAC–1 and AAM–1 asphalt binders (both binders are known to be waxy).
1B. Physical Hardening
During SHRP, it was found that the phenomenon known as physical hardening (or physical aging) in polymers also existed in asphalt binders. When an asphalt binder is kept at low temperatures for an extended period of time, its modulus increases with time. Unaged and RTFO/PAV aged SHRP core asphalt binders were used to determine the effects of physical hardening on the ABCD cracking temperature. The environmental chamber was programmed to hold the chamber air temperature at –15°C for 24 hours before cooling it at a rate of 10°C/hour. During this 24 hour isothermal conditioning, the sample temperatures were between –14.1°C and –14.8°C depending on the location of the specimen in the chamber and remained constant with a temperature fluctuation less then 0.15°C. The summary of the physical hardening effect on ABCD tests is presented in Table 7.
A statistical analysis was performed to determine the effects of three independent variables (asphalt source, aging, and physical hardening) on the ABCD cracking temperature. A three factor analysis of variance (ANOVA) is given in Table 8. The interaction among binder type, aging status and isothermal conditioning (Aspalt*Aging*PH) is very significant (p–value < 0.01). This means that the effects of physical hardening on ABCD cracking temperature is statistically significant and differs for aging status and binder type. As shown in Figure 3, for unaged binders, the 24 hours of isothermal conditioning at –15°C caused the ABCD cracking temperatures to rise significantly. However, the effects of physical hardening are quite different for aged binders. For RTFO/PAV aged binders, the 24 hour conditioning at –15°C significantly lowered cracking temperature. At this time, there is no good explanation for these different trends for unaged and aged binders. In general, the repeatability of the ABCD cracking temperature became poorer when subjected to the long isothermal conditioning. On average for all binders, the standard deviations for the ABCD test with and without the 24 hour conditioning at –15°C were 1.1°C and 0.9°C, respectively.
NM : Not Measured
a R Squared = 0.972 (Adjusted R Squared = 0.964)
1C. Refinement of Test Procedures
To help increase the ease and consistency of ABCD sample preparation, the use of turntables was incorporated into the procedure. Sixteen additional aluminum turntables were made for the ruggedness test program. For sample pouring, silicone molds were placed on 4 in. x 4 in. x 1/8 in. aluminum turntables. A tin cup holding the heated asphalt was placed on a pouring platform. As a sample was poured, the mold on the turntable was slowly turned. These turntables helped greatly in preventing overfilling and spillage due to misalignment of the pouring stream and the ¼ in. annular gap between the ABCD ring and the mold. The aluminum turntable also served as a rigid support for the flexible silicone mold. After pouring and trimming, the sample (consisting of ABCD ring, binder, mold, and turntable) was moved to the environmental chamber. The flat smooth surface of the turntable on which the sample rested reduced the risk of deforming the sample shape by accidental bending of the mold. Overall, it was felt that use of turntables enabled cleaner pouring of samples, thus easier and more consistent trimming. The cleaner trimming further enabled easier cleaning of molds and ABCD rings following the tests. The use of a turntable was added to the ruggedness test program as a variable to be tested.
In comparison with the results of EZ Asphalt Technology laboratory tests, ABCD test results of two other participating laboratories showed larger variability. From the follow–up investigation, ABCD sample preparation procedures were modified. After pouring asphalt binder in the molds, they were allowed to cool to room temperature for one hour (instead of cooling at 0°C for 20 minutes). Then, the rings were rotated between 5 and 30 degrees and back prior to trimming. These two changes made a significant improvement in the repeatability of the ABCD test. The details are further discussed in the sections for "Ruggedness Test" and "Pilot Interlaboratory Study". These modifications to the test procedures are reflected in the procedures presented in Appendices B and C.