Cover Image

Not for Sale



View/Hide Left Panel
Click for next page ( 48


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 47
47 30 Run 1 Torque @ 125C 25 Run 2 20 15 10 5 B C D E F G H I J K M N O Figure 26. Torque calculated at 125C from workability test results. Unfortunately, the poor repeatability and the low R2 for sev- oven until it reached 180C to start the workability test. The eral of the regressions indicate that this test may not be a systematic pattern that caused the second sample to be less dependable test to evaluate specific relationships between workable could have been due to a longer time in the oven for temperature and workability for binders. the second sample due to opening the oven to retrieve the first Another way to analyze the workability results was to com- sample and the workability bowl and paddle. The one binder pare the torque values at specific temperatures. Figure 26 that does not follow that trend was Binder O. This was the only through Figure 28 show a series of plots of torque values cal- case where the second sample was run on a different day from culated at three temperatures spanning 40C (72F). Although the first sample. In this review, it was discovered that the data the differences in torque values between replicates are evident, for the first sample for Binder G was incomplete: data was only relative to the range of torque values among the binders, the recorded from about 165C to 135C. Therefore results for replicate differences are in the order of 10% to 15%. The aver- this replicate were excluded in the analyses. age difference between replicates at 125C is 2.2 N-m, and the range in torque among the different binders is 20 N-m. At Compaction Tests 165C, the average difference between replicates is 1.3 N-m with an overall range of about 9 N-m. The mix compaction experiments were conducted to It also can be seen from these plots that the second run assess the effect of compaction temperature on the com- yielded higher torque results for most binders. The procedure pacted specimen density and resistance to compaction based was reviewed to determine whether the second sample was on the maximum shear ratio obtained from the specially aged longer and stiffened in the oven before testing. Each equipped Pine Instruments SGC, model AFG1A. In order to replicate sample was individually mixed and placed in the better differentiate the binder stiffness effects (i.e., binder ID 30 Run 1 Torque @ 145C 25 Run 2 20 15 10 5 B C D E F G H I J K M N O Figure 27. Torque calculated at 145C from workability test results.

OCR for page 47
48 30 Torque @ 165C Run 1 25 Run 2 20 15 10 5 B C D E F G H I J K M N O Figure 28. Torque calculated at 165C from workability test results. and temperatures), the specimens were compacted to only cally significant (P-value <0.05). Surprisingly however, the 25 gyrations. A summary of the main compaction experiment ANOVA results indicate that the interaction of the main is shown in Table 24. effects was not significant (P-value >0.05). This suggests Table 25 provides the Minitab ANOVA output for rela- that within each binder set, increasing the compaction tem- tive density, %Gmm, as the dependent variable. The main perature did not have a statistically significant impact on the effects (binder and compaction temperature) were statisti- mixture density. Table 24. Summary of Compaction Experiment A results. %Gmm Max. Shear Ratio Comp. Temp. C 110 130 150 170 110 130 150 170 Comp. Temp. F 230 266 302 338 230 266 302 338 M 85.5 -19.5 92.5 92.7 92.7 93.1 0.912 0.881 0.888 0.888 N 84.3 -25.5 91.7 92.0 92.3 92.9 0.782 0.788 0.744 0.793 G 82.5 -24.2 91.6 92.1 92.5 93.0 0.848 0.888 0.872 0.906 H 78.3 -26.1 92.4 92.8 92.6 93.0 0.860 0.821 0.861 0.871 C 75.1 -38.7 92.3 92.4 93.5 93.1 0.852 0.823 0.827 0.875 I 71.8 -29.2 92.4 92.1 92.5 92.7 0.812 0.732 0.854 0.786 B 69.3 -37.3 92.6 93.0 93.1 93.3 0.819 0.883 0.827 0.908 F 67.8 -21.3 92.7 92.9 92.5 93.5 0.863 0.837 0.979 0.961 O 65.6 -29.7 92.7 93.1 93.3 93.3 0.820 0.800 0.936 0.942 K 65.3 -13.0 92.5 93.3 93.3 93.6 0.802 0.832 1.000 0.983 J 64.3 -20.7 92.2 92.5 92.4 92.8 0.852 0.843 0.889 0.790 E 60.9 -33.1 92.8 92.7 93.3 93.6 0.832 0.872 0.881 0.862 D 60.3 -31.7 92.7 93.0 93.2 93.2 0.821 0.830 0.837 0.873 Table 25. ANOVA for relative density from Compaction Experiment A. Factor Type Levels Values Binder fixed 13 B,C,D,E,F,G,H,I,J,K,M,N,O Temperature fixed 4 230,266,302,338 Analysis of Variance for %Gmm, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Binder 12 9.80846 9.80846 0.81737 10.93 0.000 Temperature 3 8.11731 8.11731 2.70577 36.17 0.000 Binder*Temperature 36 3.57769 3.57769 0.09938 1.33 0.172 Error 52 3.89000 3.89000 0.07481 Total 103 25.39346 S = 0.273510 R-Sq = 84.68% R-Sq (adj) = 69.66%

OCR for page 47
49 Main Effects Plot for %Gmm Data Means Binder Temperature 93.2 93.0 92.8 Mean 92.6 92.4 92.2 B C D E F G H I J K M N O 230 266 302 338 Modified binders are shown with a triangle symbol Figure 29. Main effects plot for %Gmm from Compaction Experiment A. Figure 29 is a plot of the main effects that shows as com- ilarly graded Binder N. Other comparisons associated with paction temperature increased, the relative density increased Binders H, C, and J were unexpected. Compacted density for the relatively low compactive effort used in this experi- results for modified Binders H (78.8-26.1) and C (75.1-38.7) ment. The mean densities (averaged over all compaction compared more favorably with most of the unmodified binders temperatures) for each of the unmodified binders (D, E, F, J, than with the other modified binders. On the other hand, K, and O) plot above the grand mean except for Binder J. This unmodified Binder J (64.3-20.7) compared favorably with all provides a level of confidence that binder stiffness or consis- of the modified binders and was statistically different than the tency has some effect on compacted density. two unmodified binders with the closest true grades, Binders O Figure 30 summarizes a Tukey's pair-wise comparison analy- and K. As with the coating test results and the workability test sis of just the binder effect. This figure shows a cross matrix of results, there are several compaction test results that seem the binders shown in rows and columns. Black cells identify inconsistent with the binder grades. Unfortunately there does binder pairs that had %Gmm results that were not statistically not appear to be a pattern with the inconsistencies attributed different. For example, the average density for Binder M (85.5- to certain binders. 19.5) was statistically different only with Binder N. This indi- Linear regressions were established between %Gmm and cates that the warm asphalt additive Sasobit used in Binder M compaction temperature for each binder. The regressions are improved the compactability of the mix compared to the sim- shown in Table 26. Given that the equiviscous compaction M N G H C I B F O K J E D M 85.5-19.5 N 84.3-22.5 G 82.5-24.2 H 78.3-26.1 C 75.1-38.7 I 71.8-29.2 B 69.3-37.3 F 67.8-21.3 O 65.6-29.7 K 65.3-13.0 J 64.3-20.7 E 60.9-33.1 D 60.3-31.7 Figure 30. Matrix chart showing binder pairs that had statistically different compacted densities.

OCR for page 47
50 Table 26. Compaction experiment regressions. Compaction Regression Equation ID True Grade R2 Temperature for (T is Temperature C) 92.9%Gmm, F (C) M 85.5 -19.5 %Gmm = 0.0087T + 91.527 0.75 317 (159) N 84.3 -25.5 %Gmm = 0.0193T + 89.510 0.85 349 (176) G 82.5 -24.2 %Gmm = 0.0228T + 89.085 0.78 334 (168) H 78.3 -26.1 %Gmm = 0.0085T + 91.495 0.47 331 (166) C 75.1 -38.7 %Gmm = 0.0182T + 90.260 0.59 293 (145) I 71.8 -29.2 %Gmm = 0.0066T + 91.496 0.30 417 (214) B 69.3 -37.3 %Gmm = 0.0119T + 91.319 0.65 272 (133) F 67.8 -21.3 %Gmm = 0.0098T + 91.508 0.27 289 (143) O 65.6 -29.7 %Gmm = 0.0098T + 91.718 0.44 284 (140) K 65.3 -13.0 %Gmm = 0.0151T + 91.060 0.54 257 (125) J 64.3 -20.7 %Gmm = 0.0098T + 91.097 0.57 354 (179) E 60.9 -33.1 %Gmm = 0.0140T + 91.119 0.71 262 (128) D 60.3 -31.7 %Gmm = 0.0085T + 91.852 0.51 264 (129) temperatures for unmodified binders are considered reason- process. According to Pine Instrument Company (52), maxi- able and satisfactory, they were used to establish a reference mum shear ratio is a parameter that indicates the resistance of density. The relative density for the mixtures with the five the mixture to compaction as measured with the specially unmodified binders at their respective equiviscous compaction equipped AFG1 Pine SGC. The ANOVA results with shear temperatures ranged from 92.4 to 93.1%, with an average den- ratio as the dependent variable are shown in Table 27. This sity of 92.9 %Gmm. Using the linear regressions equations for analysis also shows that the main factors, binder and com- each binder in Table 26, the compaction temperature to achieve paction temperature, as well as their interaction are statisti- 92.9 %Gmm was predicted for each binder. These results are cally significant. However, the main effects and interactions also summarized in Table 26. These results are generally con- plots, shown in Figure 31 and Figure 32, do not follow any rea- sistent with the binder grades with a few exceptions. The pre- sonable trends for individual binders such as decreasing shear dicted compaction temperature for Binder I is extremely high ratio with increasing compaction temperatures, or higher as a result of the low slope term in the regression. The R2 of the shear ratios for mixtures with modified binders compared linear regression for this binder was not as good as for most with unmodified binders. For this reason, the maximum shear other binders, and the predicted temperature is well outside of ratio does not appear to be a useful indicator of compactability. the experimental range of the data. The predicted compaction Compaction Experiment B was a 1/2 factorial experiment temperature for Binder J is also very high, considering that it is designed to evaluate the factors binder, temperature, aggregate, an unmodified binder. and gradation. Since binder and temperature effects were ana- Another output of the main compaction experiment was lyzed in Compaction Experiment A, the factors of primary the maximum shear ratio developed during the compaction interest in this analysis were gradation (coarse versus fine) and Table 27. ANOVA for maximum compaction shear ratio from Compaction Experiment A. Factor Type Levels Values Binder fixed 13 B,C,D,E,F,G,H,I,J,K,M,N,O Temperature fixed 4 230,266,302,338 Analysis of Variance for Shear Ratio, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Binder 12 0.138555 0.131887 0.010991 10.83 0.000 Temperature 3 0.046824 0.047087 0.015696 15.46 0.000 Binder*Temperature 36 0.132823 0.132823 0.003690 3.63 0.000 Error 50 0.050751 0.050751 0.001015 Total 101 0.368952 S = 0.0318592 R-Sq = 86.24% R-Sq(adj) = 72.21%

OCR for page 47
51 Main Effects Plot for %Gmm Data Means Binder Temperature 93.2 93.0 92.8 Mean 92.6 92.4 92.2 B C D E F G H I J K M N O 230 266 302 338 Modified binders are shown with a triangle symbol Figure 31. Main effects plot for maximum shear ratio from Compaction Experiment A. aggregate type (low absorption and angular granite compared and the two-way interactions. These results show that grada- with a high absorption, more rounded gravel). In the experi- tion has the largest affect on relative density, followed by aggre- mental design, binder and temperature had four levels, and gate type. Recall that each of these mixtures was, in effect, aggregate and gradation had two levels. Binder and tempera- normalized by designing them with a single binder to 96% ture were reduced to two 2-level factors for the design. This Gmm at 75 gyrations. These experimental results indicate resulted in a 26 design, which was reduced to a 1/2 factorial, that a mixture's aggregate components (gradation, particle yielding 32 observations. Since the experiment was repli- shapes, texture, absorption, etc.) have a greater affect on com- cated, a total of 63 observations were made. The ANOVA paction behavior than the binder characteristics. This finding table for this experiment is shown in Table 28. Due to the 1/2 fac- is consistent with other research at NCAT (18). torial design, some of the interactions of the factors were sac- Although it has little significance to the primary objective of rificed. The ANOVA table shows the effects of the main effects this project, data from the compaction experiments also made Interaction Plot for Shear Ratio Data Means 1.00 Binder B C 0.95 D E F 0.90 G H I Mean 0.85 J K M 0.80 N O 0.75 Modified binders shown with open triangles and dashed lines 0.70 230 266 302 338 Temperature Figure 32. Interaction plot for binder and temperature on maximum shear ratio.

OCR for page 47
52 Table 28. ANOVA for Compaction Experiment B. Factor Type Levels Values Aggregate fixed 2 Granite,Gravel Grad. Fixed 2 Coarse,Fine Binder fixed 4 B,E,G,M Temperature fixed 4 230,266,302,338 Analysis of Variance for %Gmm, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Agg. 1 15.8006 15.8006 15.8006 238.50 0.0000 Grad. 1 40.6406 40.6406 40.6406 613.44 0.0000 Binder 3 16.7987 16.7987 5.5996 84.59 0.0000 Temp. 3 5.4162 5.4162 1.8054 27.27 0.0000 Agg.*Grad. 1 11.5600 11.5600 11.5600 174.49 0.0000 Agg.*Binder 2 1.5781 1.5780 0.7890 11.92 0.0001 Agg.*Temp. 2 0.3931 0.3931 0.1966 2.97 0.0658 Grad.*Binder 2 0.8406 0.8406 0.4203 6.35 0.0048 Grad.*Temp 2 0.2356 0.2356 0.1178 1.78 0.1852 Temp.*Binder 8 2.2199 2.2199 0.2775 4.19 0.0016 AB1B2=GT1T2 1 0.0400 0.0400 0.0400 0.60 0.4428 AT1T2=B1B2G 1 0.3600 0.3600 0.3600 5.44 0.0262 AB1T1=B2GT2 1 0.0303 0.0306 0.0306 0.46 0.5016 AB1T2=B2GT1 1 0.8100 0.8100 0.8100 12.23 0.0014 AB2T1=B1GT2 1 0.0400 0.0400 0.0400 0.60 0.4428 AB2T2+B1GT1 1 0.0156 0.0156 0.0156 0.24 0.6308 Error 32 2.1200 2.1200 0.0662 Total 63 98.9000 it possible to assess the effect of RAP on the compactability of content, have a statistically significant effect on the com- HMA in an SGC. This limited analysis examined the effects of pactability of the mixtures. The interaction plot from this three factors on the density (% of Gmm) of the compacted analysis is shown in Figure 33. This diagram shows that den- specimens at 25 gyrations. RAP content was the variable of pri- sities increased as compaction temperatures increased and mary interest and was evaluated at two levels, 0 and 15%. Both also that the densities for the samples with the softer-grade mixtures used a fine-graded blend of granite aggregates. Binder B were higher than the samples with the stiffer Binder M Two binders were also included in the analysis: B (64-34) at equivalent temperatures. The right side of the interaction and M (82-16). The third variable was compaction temper- plot shows that the mixtures with 15% RAP had higher den- ature, which was tested at four levels. sities than the virgin mixtures. These results do not support The ANOVA on this data set is shown in Table 29. These the hypothesis that aged binder from the RAP makes the mix results indicate that each of the main factors, including RAP stiffer and harder to compact. The interaction of RAP and Table 29. ANOVA for compaction %Gmm including RAP. Factor Type Levels Values Binder fixed 2 B,M Temp fixed 4 110,130,150,170 RAP fixed 2 0,15 Analysis of Variance for %Gmm, using Adjusted SS for Tests Source DF Seq SS Adj SS Adj MS F P Binder 1 2.20500 2.20500 2.20500 47.68 0.000 Temp 3 5.44500 5.44500 1.81500 39.24 0.000 RAP 1 3.00125 3.00125 3.00125 64.89 0.000 Binder*Temp 3 0.65500 0.65500 0.21833 4.72 0.015 Binder*RAP 1 0.66125 0.66125 0.66125 14.30 0.002 Temp*RAP 3 1.21375 1.21375 0.40458 8.75 0.001 Binder*Temp*RAP 3 0.29375 0.29375 0.09792 2.12 0.138 Error 16 0.74000 0.74000 0.04625 Total 31 14.21500