|FHWA > HfL > Technology Partnerships > ABCD Final Report > Bending Beam Rheometer Critical Temperature Interlaboratory Study|
Asphalt Binder Cracking Device to Reduce Low-Temperature Asphalt Pavement Cracking: Final Report
Bending Beam Rheometer Critical Temperature Interlaboratory Study
The same three asphalt binders used in ABCD ILS were also used for the BBR Critical Temperature ILS following ASTM C802. For the BBR ILS, one run consisted of two specimens tested at each of two test temperatures (total of four BBR specimens). The averages of two measurements at each of two temperatures were used to find the BBR critical temperature. The BBR critical temperature is the higher temperature of two at which creep stiffness (S) equals 300 MPa (43.51 ksi) and at which m-value equals 0.300. For each binder, the critical temperature measurements were repeated three times (total of 12 BBR specimens per binder). Typically, BBR critical temperature is determined by performing BBR at two temperatures where resulting creep stiffness and m-values bracket the specification limit values (300 MPa [43.51 ksi] stiffness and 0.300 m- value). However, as shown in table 7, none of the S and m-values bracketed the current AASHTO specification limit values. Determination of the BBR critical temperature for 300 MPa (43.51 ksi) stiffness and 0.300 m-value using these data would require an extrapolation process and would result in a larger variability than the interpolation process typically used in the current grading practice. To make determination of BBR critical temperature an interpolation process, arbitrary limit values bracketed by the results at two adjacent test temperatures were chosen for each binder. For BBR critical temperature for stiffness, 100 MPa (14.5 ksi), 100 MPa, and 200 (29.01 ksi) MPa were chosen for the low-, medium-, and high-stiffness binders, respectively. With these limit values, the determination of BBR critical temperature became an interpolating operation. For BBR critical temperature for m-value, 0.500, 0.400, and 0.400 were chosen for the low-, medium-, and high-stiffness binders, respectively, for the same reason. As mentioned earlier, for limit values of 300 MPa (43.51 ksi) creep stiffness and 0.300 m-value, determination of BBR critical temperature required an extrapolating operation. The results of the BBR ILS, including the averages and standard deviations of extrapolated and interpolated critical temperatures, are summarized in tables 7 and 8.
The coefficients of variation (CVs) of the BBR creep stiffness and slope (m-value) are summarized in table 9 and compared with the values given in AASHTO T313. For the BBR ILS data, CVs of the creep stiffness and slope for single- operator measurement were similar to the values given in AASHTO T313. For multilaboratory measurement, CVs for the low- and medium-stiffness binders were significantly smaller than the values given in AASHTO T313. The very high CVs for the high-stiffness binder were probably due to the unusually high stiffness, which made the BBR sample preparation difficult. The high-stiffness binder was not a typical paving-grade asphalt binder, and none of the ILS laboratories would have had sufficient experience testing such a binder with the BBR test. Overall, it can be concluded that the ILS BBR data were valid and could be used to determine the precision estimates of the BBR critical temperature in comparison with the precision estimates for the ABCD cracking temperature.
A statistical summary of BBR critical temperature ILS results is presented in table 10, and the precision estimates of the test are given in table 11. While the BBR critical temperatures at 300 MPa (43.51 ksi) creep stiffness and 0.300 m-value were much warmer than the ABCD cracking temperatures, the relative ranks of the binders from both tests were the same. For the low- and medium-stiffness binders, when interpolated, the within-lab and the between-lab variances were very low (s2 = 0.06-0.23 °C2). The variances for the high-stiffness binder were significantly larger (s2 = 0.45 and 1.38 °C2), probably due to the difficulty experienced in trimming the stiff samples.
As expected, a better precision was resulted for an interpolating condition where the specification limit values were bracketed by two test values obtained from BBR tests performed at two adjacent grading temperatures. When a single operator runs the BBR critical temperature determination twice on the same sample, the acceptable range of two test results should be less than 1.9 ºC (3.4 ºF) and 1.2 ºC (2.2 ºF) for extrapolating and interpolating conditions, respectively. When two different operators at two different laboratories perform the BBR critical temperature determination on the same sample, the acceptable range of two test results should be less than 3.5 ºC (6.3 ºF) and 2.1 ºC (3.8 ºF) for extrapolating and interpolating conditions, respectively.
The precision of ABCD cracking temperature was somewhat less than that of the BBR critical temperature. However, it should be pointed out that the real point of thermal cracking is determined by comparing thermal stress and the strength of asphalt binder. Thermal cracking occurs when thermal stress exceeds strength. The BBR critical temperature does not consider strength. If the variability of strength determination were added to BBR variability for real cracking temperature determination, the level of precision would be similar to that of the ABCD. Furthermore, most of the participating laboratories had only about a week of experience with the ABCD test procedure before starting the ILS ABCD tests while they had years of experience with the BBR. As laboratories gain more experience with ABCD testing, the precision of ABCD test results will improve.
a no replicate; b outlier; -- c data were not reported; NA = not available
a no replicate; -- b data were not reported; NA = not available
a Average excluding PG 88+6 data
aThese values represent the 1S and D2S limits described in ASTM Practice C 670.
PDF files can be viewed with the Acrobat® Reader®