Field Verification of Proposed Changes to the AASHTO R 30 Procedures for Laboratory Conditioning of Asphalt Mixtures (2019)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

14 Findings and Applications Chapter 3 presents the data collected from the eight field sites in which periodic post-construction coring was allowed. The analysis of long-term aging is presented in two ways. The first is the long-term, global model for comparing field aging with laboratory LTOA. An experiment was conducted to determine if another LTOA procedure could be used to predict stiffness farther into the future than the LTOA pro- tocols used in NCHRP Project 09-52. This information can also be used to model changes in stiffness with time for incor- poration into mechanistic-empirical pavement design. After the global model is presented and described, the analysis of the long-term impacts of the factors from the various test sites is given. This was done to show whether the factor differ- ences presented in NCHRP Project 09-52 converge, diverge, or remain the same from their initial values. Researchers also conducted limited CTs with field cores left over from NCHRP Projects 09-52 and 09-52A to understand the relationship between mixture stiffness and cracking resistance. This part of the study was done to determine whether further refine- ment to long-term aging models should include a fracture parameter. Global Models All Field Sites One goal of this project was creating a global model of MR stiffness values versus CDDs. The AV contents and MR stiffness values for the cores, for all eight field sites are listed by date in Appendix A. Figure 9 is a scatterplot of MR stiff- ness versus CDDs from all long-term aged field sites. This global model includes all the testing results from field cores at construction to the most recent core sampling, and it shows a trend of increasing stiffness with age. The stiffness in field cores changed primarily as a result of aging in the field, although changes in AVs may have also contributed to the increase. The overall trend was for stiffness values to approach a plateau after a long in-service time. This global model shows that there was a separation in stiffness between field sites with warmer climates (Florida, Texas I and II, New Mexico, and Indiana) and colder climates (Iowa, South Dakota, and Wyoming). Although this is not a hard and fast rule, the demarcation between warmer and colder climates seems to be about 40th parallel north. The field sites in the colder climate consistently exhibited lower MR stiffness values both at construction and post-construction when compared with those for the field sites in the warmer climate. The R2 value of 0.40 for the overall global model means that this trendline does not explain the variabilities of these field sites very well. The R2 value was improved by separat- ing the data into colder and warmer climates (Figure 10 and Figure 11). Cold and Warm Climate Field Sites As shown in Figure 10 and Figure 11, the field sites were divided into two categories: colder climate and warmer climate. The R2 improved to 0.80 for both climates (Figure 10 and Figure 11), which means the field sites were more alike in terms of stiffness when separated by climate. The colder climate MR stiffness values approximately doubled over 50,000 CDDs, which is about the same rate of change for 50,000 CDDs in the warmer climate. In fact, the exponents in the equations in Figure 10 and Figure 11 are approxi- mately 0.18, which means that both models had the same rate of change over time. The coefficients differ primarily because the initial aging of the materials was different in the two climates, and partially because of the grade of the virgin binder used. As was shown in Table 2, the binders used in the cold climates of Iowa, South Dakota, and Wyoming were PG 58-28, PG 58-34, and PG 64-28, respectively. In the warmer climates, PG 64-22 was used in Texas II and Indiana, PG 76-22 was used in New Mexico, and PG 70-22 was used in Texas I. Both Figure 10 and Figure 11, showing colder and C H A P T E R 3

15 y = 74.191x0.218 R² = 0.3968 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) New Mexico Iowa South Dakota TX I Wyoming Indiana TX II Florida Power (All Points) Figure 9. Scatterplot of all long-term test section data. warmer climates, respectively, illustrate that the MR stiffness values accelerated faster during the first 10,000 CDDs after construction. MR Stiffness Ratio The MR stiffness ratio is the ratio of the MR stiffness value of the post-construction field cores to the MR stiffness value of the field cores at construction. The purpose of analyzing the data this way is to show how field aging affects mixture stiff- ness when the MR stiffness values are normalized. Figure 12 shows the MR stiffness ratio versus CDDs at all field sites. The data points are the normalized stiffness values at the cor- responding CDDs for all the field samples. The projected line was developed based on the power equation (Equation 3-1) from the field core data. As shown in Figure 12 through Figure 14, both the colder climate and warmer climate field sites appeared to have a decelerating rate for the MR stiffness ratio after a long in-service time. The R2 values for these rela- tionships are similar. The primary reason for using the MR stiffness ratio is to provide a frame of reference for the LTOA protocols investigated. However, the relationships for MR stiffness values in Figure 9 through Figure 11 seem to have more value as models: Property Ratio (3-1)= α × βCDD 1 exp (3-2)= + α × − β          γ CDD Where: CDD = cumulative degree-days for cores. α, β, and γ = fitting coefficients. Figure 15 shows an MR stiffness ratio correlation of field aging and LTOA protocols. The field core data, in terms of MR stiffness ratio, are the diamonds, the line is the best fit through those points, and the dots with error bars are the laboratory sample MR stiffness ratios placed on the fitted line. The error bars show the expected standard deviation of the MR stiffness ratio and CDD for each predicted point. The number of CDDs modeled by the laboratory sample MR stiffness ratios is directly below the dots. The LTOA proto- col using a compacted specimen aged for 2 weeks in a 140°F (60°C) oven corresponded to the least number of CDDs (9,000). The next highest number of CDDs (23,000) simu- lated in the laboratory was for the LTOA protocol in which a compacted specimen was aged for 5 days at 185°F (85°C) before compaction. Both of those LTOA protocols were used in NCHRP Project 09-52. In the current project, another protocol was used, in which some South Dakota loose mix sampled from the plant at the time of construction was aged for 5 days at 185°F (85°C) prior to compaction. It was thought that air circulating through the mix would accelerate oxidation. This is shown by the third dot on the line, which lies beyond the CDD for the latest cores. This last protocol resulted in a simulation of 134,000 CDDs. Note that the MR stiffness ratio results of the loose mix aging protocol cor- responded to one field project only; different simulations of CDDs could be obtained as additional results become avail- able. For the LTOA loose mix CDD, the MR stiffness ratio was calculated using the measured MR stiffness values of the South Dakota LTOA PMLC over the MR stiffness values of the South Dakota STOA LMLC. The South Dakota LMLC was aged at 275°F (135°C) for HMA or 240°F (116°C) for WMA for 2 hours after compaction in NCHRP Project 09-52. The error associated with the model increased with time. This

16 y = 147.44x0.1795 R² = 0.8006 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) New Mexico TX I TX II Florida Indiana Power (All Points) Figure 11. Long-term trend for field sites in warmer climates. y = 54.873x0.1787 R² = 0.8011 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) Iowa Wyoming South Dakota Power (All Points) Figure 10. Long-term trend for field sites in colder climates.

17 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 0 30000 60000 90000 120000 M R S tif fn es s R at io Cumulative Degree-Days (°F-days) Predicted MeasuredR2 = 0.51 Figure 12. MR stiffness ratio versus CDD for all field sites. 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 30000 60000 90000 120000 M R S ti ff ne ss R at io Cumulative Degree-Days (°F-days) Predicted MeasuredR 2 = 0.46 Figure 13. MR stiffness ratio versus CDD for colder climates. 0.5 1.0 1.5 2.0 2.5 3.0 3.5 0 30000 60000 90000 120000 M R S ti ff ne ss R at io Cumulative Degree-Days (°F-days) Predicted MeasuredR2 = 0.59 Figure 14. MR stiffness ratio versus CDD for warmer climates.

18 finding is sensible because the uncertainty of predictions tends to increase further in the future. The data in Table 3 suggest that a more realistic approach to LTOA could be aging loose mix at 185°F (85°C) for 5 days before compacting. For the sites in this study, that equates to 7 to 10 years for warmer climates and 12 to 14 years for cooler climates. Factor Analysis Although the factors in the experiment design were included to study their effects on short-term aging, the data collected from NCHRP Projects 09-52 and 09-52A allowed for the study of long-term aging effects as well. The long-term aging effects on MR stiffness by factor are described in this section in two ways. The first is to plot the paired levels of the factors against each other to see where they lie in relation to the line of equality. The second is to plot the levels with time to see the trends and whether or not the differences are really significant in relation to the overall changes with time. Aggregate Type Both the Florida and the Iowa field sites had aggregate type (absorption) as a factor (Figure 16). Both the field cores Field Site Equivalent Time Aging Protocols NCHRP Project 09-52/9-52A Compacted Specimens 2 weeks at 60°C NCHRP Project 09-52 Compacted Specimens 5 days at 85°C NCHRP Project 09-52A Compacted Specimens 5 days at 85°C NCHRP Project 09-52A Loose Mix 5 days at 85°C CDD 9,000 16,000 23,000 134,000 Months Years Months Years Months Years Months Years Texas I 6 0.5 11 0.9 15 1.3 90 7.5 New Mexico 8 0.7 13 1.1 18 1.5 102 8.5 Wyoming 12 1.0 20 1.7 29 2.4 161 12.5 South Dakota 14 1.2 22 1.8 30 2.5 159 13.4 Iowa 11 0.9 20 1.7 28 2.3 163 13.3 Florida 6 0.5 11 0.9 15 1.3 89 7.4 Indiana 11 0.9 19 1.6 26 2.2 144 12.0 Texas II 6 0.5 11 0.9 16 1.3 94 7.8 Table 3. Summary of equivalent time of laboratory LTOA specimens. Figure 15. MR stiffness ratio correlation between field aging and laboratory LTOA protocols.

19 with high-absorptive aggregates and low-absorptive aggre- gates at the Iowa field site had similar MR stiffness values, so the points were close to the line of equality. The mixture with a low-absorption aggregate at the Florida field site exhibited higher MR stiffness values, which is most likely a function of the RAP content in the mixture. The deflection of the line of observation from the line of equality is about 8%, as shown by the equation in Figure 16, which is greater than the amount of deflection to be considered insignificant (<5%). Figure 17 shows the change in MR stiffness with CDD and time for the Iowa field site, in which the two lines are, for all practical purposes, identical. NCHRP Report 815 describes how a thicker effective binder film thickness (FTbe) result- ing from incorporating higher binder contents in lower- absorption aggregate mixtures could reduce the stiffness. Figure 18 shows that the granite and limestone mixtures at the Florida field site had a relatively large separation at con- struction and in the early in-service time of the project and seem to converge later in the pavement’s life. Asphalt Source The Texas II field site is the only one that had two different asphalt sources as one of its factors. Table 4 presents the results of continuous PG grading on the original asphalt samples. Both sources successfully met the requirements of a PG 64-22 binder, with the minor difference between the two being the intermediate temperature grade, which was 22.5 for Binder A and 19.5 for Binder V. The high- and low-temperature PG grades measured from materials at construction showed negligible differences between the two sources, as did the carbonyl area (Table 5). The asphalt grading of the extracted LMLC binder was essentially the same as the loose mix sample and the cores taken at construction. y = 1.079x 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 Lo w A bs or pt io n M R S tif fn es s ( ks i) High-Absorption MR Stiffness (ksi) Florida Iowa Linear (All Points) Figure 16. Aggregate type MR stiffness global. High Abs Low Abs Power (High Abs) Power (Low Abs) 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) 10 Months 40 Months Figure 17. Aggregate type at the Iowa field site.

20 Figure 19 shows the plot of MR stiffness values for mixtures with Binder A versus those for mixtures with Binder V. The mixture with Binder A has much higher MR stiffness values, which means the factor of asphalt source is significant based on MR stiffness testing. Figure 20 shows the long-term trends for the Texas II field site. Because NCHRP Project 09-52 initially showed no difference between the asphalt plant types but found that asphalt source was significant, only the two DMPs with Binder A and Binder V were sampled for the 51-month comparison. The trend of stiffening with time for the mixtures shown in Figure 20 is interesting in that the two mixtures showed a difference at the time of construction, and the difference increased with time. By the time of sampling and testing of the materials at 51 months, the mixture with Binder A was approximately 20% stiffer. The PG grading of the two 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) Granite Limestone Power (Granite) Power (Limestone) 9 Months 44 Months 51 Months15 Months Figure 18. Aggregate type at the Florida field site. Binder Source Specified PG Grade High PG Grade (DSR Unaged Binders) High PG Grade (DSR RTFO) Intermediate PG Grade (DSR on RTFO + PAV) Low PG Grade (BBR on RTFO + PAV) Continuous PG Grade Binder A 64-22 65.7 64.4 22.5 −24.4 64.4–24.4 (22.5) Binder V 64-22 67.6 64.2 19.5 −24.0 64.2–24.0 (19.5) Note: DSR = Dynamic shear rheometer; RTFO = Rolling thin film oven; PAV = Pressure aging vessel; and BBR= Bending beam rheometer. Table 4. Comparison of original true binder grading at Texas II. Mixture Type Specimen Type High PG Grade Low PG Grade FTIR CA Asphalt Content, % BMP—Binder A Plant Loose Mix 68.5 −22.7 0.75 6.36 Construction Core 67.9 −23.0 0.70 6.36 LMLC STOA 67.9 −23.5 0.78 6.20 BMP—Binder V Plant Loose Mix 65.9 −24.3 0.69 6.37 Construction Core 65.3 −24.7 0.67 6.37 LMLC STOA 65.7 −23.8 0.71 6.30 DMP—Binder A Plant Loose Mix 66.4 −24.0 0.68 5.91 Construction Core 66.0 −25.3 0.65 5.91 LMLC STOA 66.9 −25.1 0.78 6.20 DMP—Binder V Plant Loose Mix 67.5 −23.5 0.70 5.90 Construction Core 66.9 −23.8 0.71 5.90 LMLC STOA 65.7 −23.6 0.71 6.30 Note: FTIR = Fourier-transform infrared spectroscopy. Table 5. Comparison of extracted and recovered binder grades from construction at Texas II and mixture asphalt contents.

21 asphalts at the time of construction showed very little differ- ence between them. The only indication was that the inter- mediate PG grade for Binder V was about 3°C lower than that for Binder A. Recycled Materials Figure 21 shows the MR stiffness global plot for the factor of recycled materials for the Texas I and New Mexico field sites. Mixtures with RAP/RAS had a greater stiffness at con- struction compared with the corresponding virgin mixtures. The slope magnitude of the trend line was 0.31, which means the presence of recycled materials was significant. Figure 22 shows the average MR stiffness results at the Texas I field site for the factor of recycled materials. The mixture with RAP/RAS had a higher MR stiffness value at construction and afterward. There was a recently discovered error in the testing protocol at the 14-month test period in which the RAP/RAS mixtures were handled incorrectly. There are no results on the RAP/RAS mixture at 22 months after construction because of a communication problem with the sampling crew. In Figure 23, the RAP mixtures at the New Mexico field site maintained higher MR stiffness values compared with the virgin mixture. However, the New Mexico virgin mixture did show a greater variability at the 51- and 60-month sampling periods, which may have been as a result of a misidentification of the test sections. WMA Technologies Seven out of the eight field sites had mixtures produced with WMA technologies. On the whole, WMA mixtures were softer than their HMA counterparts, although differences in the production temperatures of 30°F (-1°C) did not show y = 0.8161x 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 Bi nd er V M R S tif fn es s ( ks i) Binder A MR Stiffness (ksi) Texas II Linear (All Points) Figure 19. Asphalt source MR stiffness global. 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) Binder A Binder V Power (Binder A) Power (Binder V) 42 Months 51 Months Figure 20. Asphalt source at the Texas II field site.

22 y = 1.3152x 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 R A P/ R A S M R S tif fn es s ( ks i) No RAP/RAS MR Stiffness (ksi) New Mexico TX I Linear (All Points) Figure 21. Recycled material MR stiffness global. 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) RAP/RAS No RAP/RAS Power (RAP/RAS) Power (No RAP/RAS) 0 Month 8 Months 22 Months 60 Months 14 Months Figure 22. Recycled materials at the Texas I field site. 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) No RAP RAP Power (No RAP) Power (RAP) 10 Months 51 Months 60 Months Figure 23. Recycled materials at the New Mexico field site.

23 an effect on stiffness of HMA. It may be that the presence of WMA technology, whether foaming or additives, created the difference between HMA and WMA. For example, it is possible, with foaming or additives, that the film thickness of asphalt may have had an impact on the behavior of the mixtures. The New Mexico mixture stiffnesses all lie above the line of equality, which could be indicative that the pres- ence of RAP had a larger impact on their stiffness. The slope magnitude was 0.90, which indicates that the factor of WMA technologies had a significant effect on the MR stiffness values (see Figure 24). As shown in Figure 25, the HMA mixtures at the Texas I field site had higher MR stiffness values at construction com- pared with the WMA mixtures, although the HMA and WMA mixture stiffness values seem to converge at 60 months. Although the New Mexico WMA cores show a higher MR stiff- ness at all sampling intervals, the difference seems to diminish slightly with time (see Figure 26). For the New Mexico field site results, the presence of a stiff RAP in the WMA mixtures and the virgin mixture being HMA may have influenced the trend. The Wyoming (Figure 27), South Dakota (Figure 28), and Iowa (Figure 29) field sites all produced results that did not show much, if any, difference between HMA and WMA. It is believed that this was because of the slow aging of the materials in these cold climates. HMA mixtures at the Florida field site (Figure 30) main- tained a higher MR stiffness than the WMA mixtures over the 51-month time period. Figure 31 shows the trend for WMA and HMA at the Indiana field site. The HMA MR stiffness y = 0.8962x 0 200 400 600 800 1000 1200 1400 1600 1800 0 200 400 600 800 1000 1200 1400 1600 1800 W M A M R S tif fn es s ( ks i) HMA MR Stiffness (ksi) New Mexico TX I Iowa Wyoming Florida South Dakota Indiana Linear (All Points) Figure 24. WMA technologies MR stiffness global (all locations). 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) HMA WMA Power (HMA) Power (WMA) 8 Months 22 Months 60 Months14 Months Figure 25. WMA technologies at the Texas I field site.

24 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss (k si ) Cumulative Degree-Days (°F-days) HMA WMA Power (HMA) Power (WMA) 10 Months 51 Months 60 Months Figure 26. WMA technologies at the New Mexico field site. 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss (k si ) Cumulative Degree-Days (°F-days) HMA WMA Power (HMA) Power (WMA) 10 Months 59 Months51 Months Figure 27. WMA technologies at the Wyoming field site. 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) HMA WMA Power (HMA) Power (WMA) 14 Months 83 Months30 Months Figure 28. WMA technologies at the South Dakota field site.

25 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss (k si ) Cumulative Degree-Days (°F-days) HMA WMA Power (HMA) Power (WMA) 11 Months 85 Months28 Months Figure 29. WMA technologies at the Iowa field site. 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss (k si ) Cumulative Degree-Days (°F-days) HMA WMA Power (HMA) Power (WMA) 9 Months 44 Months 51 Months15 Months Figure 30. WMA technologies at the Florida field site. 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss ( ks i) Cumulative Degree-Days (°F-days) HMA WMA Power (HMA) Power (WMA) 9 Months 47 Months Figure 31. HMA versus WMA technologies at the Indiana field site.

26 y = 0.9897x 0 100 200 300 400 500 600 0 100 200 300 400 500 600 H ig h T em pe ra tu re M R S ti ff ne ss ( ks i) Control Temperature MR Stiffness (ksi) Iowa Wyoming Linear (All Points) Figure 32. Production temperatures MR stiffness global. 0 200 400 600 800 1000 1200 1400 1600 0 20000 40000 60000 80000 100000 M R S ti ff ne ss (k si ) Cumulative Degree-Days (°F-days) Low Temperature High Temperature Power (Low Temperature) Power (High Temperature) 10 Months 59 Months 51 Months Figure 33. Production temperatures at the Wyoming field site. was the lower of the two at the last sampling time, which caused the HMA curve to be lower than the WMA curve. A lower AV content that could contribute to reduced aging in the HMA (shown in Appendix A) may be one of the causes. Production Temperature Production temperature was one of the two factors iden- tified in NCHRP Project 09-52 as having no effect on aging. The Iowa and Wyoming field sites were the locations that compared normally heated asphalt mix against mix that was produced at a temperature 30°F (17°C) hotter. Although researchers did not intend to monitor this variable in NCHRP Project 09-52A, the samples were taken and deliv- ered, so the testing was completed, and the results are discussed herein. Figure 32 shows the MR stiffness global plot for the differ- ence in production temperatures at the Iowa and Wyoming field sites. Both field sites show very similar MR stiffness values for the stiffness of mixtures produced at the different temperatures. All the data points are very close to the line of equality, with a slope magnitude of 0.99, which indicates that the factor of production temperature does not have a significant effect on the values of MR stiffness. Figure 33 for Wyoming and Figure 34 for Iowa show that the relative position of the stiffness values did not change with time. Summary of Findings This project found that the long-term aging of asphalt mix- tures is time and temperature dependent, and that aging can be modeled using MR stiffness and CDD. The relationships

27 between stiffness and temperature-time are best modeled for warm and cold climates separately using a power curve. To compare laboratory LTOA protocols to field aging, a property ratio (MR stiffness ratio) to normalize the values was deter- mined and plotted against CDDs (Figures 13 and 14). The MR stiffness ratios from the LTOA mixtures were then plotted along the curve to determine the corresponding CDDs (Figure 15). An additional aging protocol was used during this study with the mixtures from the South Dakota field site. The data in Table 3 suggest that a more realistic approach to LTOA could be aging loose mix at 185°F (85°C) for 5 days before compacting. For the sites in this study, that equates to 7 to 10 years for warmer climates and 12 to 14 years for cooler climates. Table 6 summarizes the slope magnitudes of the paired factors in Figure 16, Figure 19, Figure 21, Figure 24, and Figure 32 and indicates which factors were significant. The results from NCHRP Project 09-52 are also included. While NCHRP Project 09-52 included LMLC and PMPC speci- mens and a limited number of field cores, NCHRP Project 09-52A tested an extended number of field cores, which makes the data more useful for long-term study. When the slope magnitude of the divergent line with respect to the line of equality was greater than 1.05 or less than 0.95, a fac- tor was deemed significant. Thus, it can be concluded that the factors of aggregate absorption, asphalt source, recycled material content, and WMA technologies have significant effects on asphalt mixture stiffness. Production temperature was found to have no significant effect on mixture stiffness. These findings are consistent with those from NCHRP Proj- ect 09-52. IDEAL-CT Results Table 7 provides the CT index results. The field cores were chosen based on availability and representability of a range in CDD. A higher value of the CT index indicates better cracking resistance. Figure 35 and Figure 36 show the plots between CT index versus CDD and MR stiffness, respectively. CT index tends to decrease as the CDD and MR stiffness increase. Because it was a limited experiment, there were not enough data points to reasonably infer a model. The IDEAL-CT is introduced here to be another characteristic that will sup- port, along with stiffness, the performance of field cores. In Figure 37, only the CT index of the HMA mixtures (no WMA) at the field sites was plotted versus CDD. The trend is more linear than that shown in Figure 35 and Figure 36 because of the absence of the WMA data. 0 200 400 600 800 1000 1200 1400 1600 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 M R S ti ff ne ss (k si ) Cumulative Degree-Days (°F-days) Control Temperature High Temperature Power (Control Temperature) Power (High Temperature) 11 Months 85 Months28 Months Figure 34. Production temperatures at the Iowa field site. Factor Slope Magnitude Significant Effect Significant Effect in NCHRP Project 09-52 Aggregate Types 1.08 Yes Yes Asphalt Sources 0.82 Yes N/A Recycled Materials 1.32 Yes Yes WMA Technologies 0.90 Yes Yes Production Temperatures 0.99 No No Table 6. Summary of the effects of factors on mixture aging characteristics.

28 Field Site Months after Placement CDD AV, % MR stiffness (ksi) CT Index Texas I 60 89,777 7.7 1,433 3 New Mexico 22 29,158 4.5 1,080 27 Wyoming 59 49,490 3.6 370 52 South Dakota 57 47,785 6.4 350 37 Iowa 10 6,174 N/A 231 249 Florida 44 66,638 5.7 1,073 41 51 78,565 5.8 1,128 30 Indiana 47 44,065 8.0 1,086 76 Table 7. Summary of CT index at all field sites. 1 10 100 1000 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 C T I nd ex Cumulative Degree-Days (°F-days) New Mexico TX I Iowa Wyoming Florida South Dakota Indiana Power (All Points) Figure 35. CT index versus CDD global plot (all locations). 0 50 100 150 200 250 300 0 200 400 600 800 1000 1200 1400 1600 C T I nd ex MR Stiffness (ksi) New Mexico TX I Iowa Wyoming Florida South Dakota Indiana Linear (All Points) Figure 36. CT index versus MR stiffness global plot (all locations).

29 Figure 37. HMA CT index versus CDD global plot (all locations). 0 20 40 60 80 100 120 140 160 180 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 H M A C T I nd ex Cumulative Degree-Days (°F-days) New Mexico TX I Iowa Wyoming Florida South Dakota Indiana Linear (All Points)