Click for next page ( 29

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 28
28 4000 3500 Coarse-Graded Fine-Graded Lab Permeability, E-5 cm/sec Coarse-graded 3000 y = 2.279e0.4225x R2 = 0.6942 2500 2000 1500 1000 Fine-graded y = 0.0309e0.7457x 500 R2 = 0.8602 0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 In-place Air Voids (Corelok), % Figure 35. Plot of permeability versus in-place air voids for 9.5-mm NMAS mixes. lower void levels the coarse-graded mixes are more permeable is significant (p-value = 0.000). Based on the trend line, per- than fine-graded mixes. meability is very low at air void contents less than 8 percent. The relationships for the coarse-graded and fine-graded At air void contents above 8 percent, the permeability begins 12.5-mm NMAS mixes are shown in Figure 36. For these to increase rapidly with a small increase in in-place air void projects there was no significant difference between fine and content. At 8 percent air voids, the fine-graded 19.0-mm coarse graded mixes. The relationships between in-place air NMAS mix has a permeability value of 16 10-5 cm/sec. voids and permeability for both gradation types were reason- able and significant with an R2 of 0.61 for coarse-graded mixes (p-value = 0.000) and 0.58 for fine-graded mixes (p-value 4.10 CONTROLLED LABORATORY = 0.000). As shown by the best-fitted lines, the permeability EXPERIMENT TO EVALUATE METHODS values for both gradation types were basically the same at a OF MEASURING THE BULK SPECIFIC GRAVITY OF COMPACTED HMA given air void content. The permeability at 8.0 percent air voids for coarse-graded and fine-graded mixes was approximately 4.10.1 Introduction and Problem Statement 30 10-5 cm/sec. Figure 37 illustrates the relationship between in-place A major concern of the HMA industry is the proper mea- air voids and permeability for fine-graded 19.0-mm NMAS surement of bulk specific gravity (Gmb) for compacted samples. mixes. The R2 value for this figure is 0.59 and the relationship This issue has become a bigger problem with the increased 1000 12.5 mm Coarse-Graded 900 12.5 mm Fine-Graded 800 Lab Permeability, E-5 cm/sec 700 600 500 400 300 200 100 0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 22.0 24.0 In-place Air Voids (Corelok), % Figure 36. Plot of permeability versus in-place air voids for 12.5-mm NMAS Mixes.

OCR for page 28
29 140 120 Lab Permeability, E-5 cm/sec 100 80 0.7362x y = 0.0437e 2 R = 0.5923 60 40 20 0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 In-place Air Voids (Corelok), % Figure 37. Plot of permeability versus in-place air voids for 9.5-mm NMAS mixes. use of coarse gradations. Bulk specific gravity measurements gradations passing close to or above the Superpave defined are the basis for volumetric calculations used during HMA maximum density line (i.e., fine-graded). However, since the mix design, field control, and construction acceptance. Dur- adoption of the Superpave mix design system and the increased ing mix design, volumetric properties such as air voids, voids use of SMA, mixes are being designed with coarser gradations in mineral aggregates, voids filled with asphalt, and percent than in the past. theoretical maximum density at a certain number of gyrations The potential problem in measuring the Gmb of mixes like are used to evaluate the acceptability of mixes. All of these coarse-graded Superpave and SMA using the SSD method properties are based upon Gmb. comes from the internal air void structure within these mix In most states, acceptance of HMA construction by the types. These types of mixes tend to have larger internal air owner is typically based upon percent compaction (density voids than the finer conventional mixes, at similar overall based upon Gmb and theoretical maximum density). Whether air void contents. Mixes with coarser gradations have a much nondestructive (e.g., nuclear gauges) or destructive (e.g., cores) higher percentage of large aggregate particles. At a certain tests are used as the basis of acceptance, Gmb measurements overall air void volume, which is mix specific, the large are equally important. When nondestructive devices are uti- internal air voids of the coarse mixes can become inter- lized, each device first has to be calibrated to the Gmb of cores. connected. During Gmb testing with the SSD method, water If the Gmb measurements of the cores are inaccurate in this can quickly infiltrate into the sample through these intercon- calibration step, then the nondestructive device will provide nected voids. However, after removing the sample from the inaccurate data. Additionally, pay factors for construction, water bath to obtain the saturated-surface dry condition the whether reductions or bonuses, are generally based upon water can also drain from the sample quickly. This drain- percent compaction. Thus, errors in Gmb measurements can ing of the water from the sample is what causes errors when potentially affect both the agency and producer. using the SSD method. For many years, the measurement of Gmb for compacted Because of the potential errors noted with the saturated HMA has been accomplished by the water displacement surface-dry test method of determining the bulk specific grav- method using saturated-surface dry (SSD) samples. This ity of compacted HMA, the primary objective of this task was method consists of first weighing a dry sample in air, then to compare AASHTO T166 with other methods of measur- obtaining a submerged mass after the sample has been placed ing bulk specific gravity to determine under what conditions in a water bath for a specified time interval. Upon removal AASHTO T166 is accurate. from the water bath, the SSD mass is determined after The plan for this part of the study was to evaluate two sep- patting the sample dry using a damp towel. Procedures for arate sample types: laboratory compacted and field compacted. this test method are outlined in AASHTO T166 (ASTM Laboratory compacted mixtures having various aggregate D2726). types, nominal maximum aggregate sizes, gradation shapes, The SSD method has proven to be adequate for conven- and air void levels were prepared. Each of the prepared tionally designed mixes, such as those designed according to samples was tested to determine bulk specific gravity by four the Marshall and Hveem methods, that generally utilized fine- different test methods: water displacement, vacuum-sealing, graded aggregates. Historically, mixes were designed to have gamma ray, and dimensional.

OCR for page 28
30 For the field compacted samples, cores obtained during Statistically, results of the DMRT comparisons show that the field validation portion of this study were subjected to the all methods produced statistically different air void contents. same four bulk specific gravity test methods. Because cores However, vacuum-sealing and gamma ray bulk specific grav- have a different surface texture than laboratory compacted ity methods provided similar results given a difference of samples, it was necessary to evaluate them also. Testing also 0.24 percent air voids. On average, the dimensional method conducted on core samples included laboratory permeabil- resulted in the highest air void contents, followed by the ity tests and effective air void content using the vacuum- vacuum-sealing and gamma ray methods, respectively. Air sealing device. void contents determined from AASHTO T166 resulted in the lowest air void contents. None of the alternative meth- ods provided similar results to AASHTO T166. The results for SMA mixtures are provided in Figure 39. As 4.10.2 Field Compacted Samples with the Superpave mixes, the vacuum-sealing and gamma ray methods resulted in similar air void contents. The dimen- Each of the cores obtained during the Task 5 field valida- sional method again resulted in the highest air voids and tion were tested to determine bulk specific gravity using the the AASHTO T166 method resulted in the lowest air voids. same four tests as the laboratory experiment: water displace- Analysis of both the Superpave and SMA data indicated that ment, vacuum sealing, gamma ray, and dimensional analysis. the four methods of measuring bulk specific gravity signif- Because of the differences in surface texture between labora- icantly affected resulting air voids. For both mix types, the tory compacted samples (surface texture around entire sam- vacuum-sealing and gamma ray methods provided similar air ple) and field compacted samples (surface texture only on top voids; however, the dimensional method provided significantly of sample because of core bit and sawing), the experiment higher air voids and AASHTO T166 provided significantly was also extended to core samples. lower air void contents. Because of the differences in resulting air voids for the Theoretically, the dimensional method should provide the four methods of measuring bulk specific gravity, a Duncan's highest measured air void content, as this method includes multiple range test (DMRT) was conducted to determine both the internal air voids and the surface texture of the sam- which methods, if any, provided similar results. This analy- ple. Therefore, the results in Figures 38 and 39 pass the test sis method provides a ranking comparison between the dif- of reasonableness for the vacuum-sealing, gamma ray, and ferent methods. The range of sample means for a given set of AASHTO T166 methods as all three provided air void content data (method) can be compared to a critical valued based on lower than the dimensional method. the percentiles of the sampling distribution. The critical value Because it was assumed that the T-166 method would be is based on the number of means being compared (four, rep- accurate at low water absorption levels, it was decided to test resenting the different methods) and number of degrees of the mixes with low absorption, less than 0.5 percent, to see freedom at a given level of significance (0.05 for this analy- which mixes provided results similar to the T-166 method. sis). Results of the DMRT analysis for the Superpave mixes The results are provided in Figure 40. This figure shows that are illustrated in Figure 38. the vacuum-sealing and AASHTO T166 methods provided 11 Letters represent results of Duncan's A Multiple Range Test for air voids 10 9.39 resulting from the bulk specific gravity methods. Methods with the same letter 9 ranking are not significantly different. Average Air Void Content, % B C 8 7.50 7.26 D 7 6.22 6 5 4 3 2 1 0 Dimensional Vacuum-Sealing Gamma Ray AASHTO T166 Test Method Figure 38. Average air voids and DMRT results for Superpave mixes.

OCR for page 28
31 12 Letters represent results of Duncan's A Multiple Range Test for air voids 10.11 resulting from the bulk specific gravity methods. Methods with the same letter 10 ranking are not significantly different. Average Air Void Content, % B B 8 7.24 7.09 C 6 4.97 4 2 0 Dimensional Vacuum-Sealing Gamma Ray AASHTO T166 Test Method Figure 39. Average air voids and DMRT results for SMA mixes. similar results and that both were significantly different than to the gamma ray method may make this method a viable the dimensional and gamma ray methods. The dimensional option in the future. method provided the highest air void content, as expected. The AASHTO T166 method is accurate for low water absorp- tion mixes and at these low void levels provide similar den- 4.10.3 Analysis of Field Compacted Samples sity values to that of the vacuum seal method. These results suggest that the vacuum-sealing method provides an accurate Included within this portion of the study were the cores density for low voids, which indicates that it also provides an obtained during the Task 5 field validation experiment. Only accurate density at higher void levels because the plastic seal the vacuum-sealing and AASHTO T166 test methods were will clearly prevent water from being absorbed into the mix- analyzed, as they were shown most accurate during the labo- ture. Figures 38 and 39 suggest that the gamma ray method ratory phase of this experiment. Figure 41 illustrates the rela- does an overall adequate job of estimating bulk specific grav- tionship between air voids determined from the two methods ity; however, Figure 40 suggests that it is not as accurate as for all field cores obtained from the 20 field projects during AASHTO T166 or the vacuum-sealing methods. Refinements Task 5. This figure illustrates that when air void content is less 7.0 Letters represent results of Duncan's Multiple Range Test for air voids 6.0 resulting from the bulk specific gravity methods. Methods with the same letter A B ranking are not significantly different. 4.9 5.0 Air Void Content, % 4.3 C C 4.0 4.0 3.8 3.0 2.0 1.0 0.0 Dimensional Gamma Ray Vacuum-Sealing AASHTO T166 Test Method Figure 40. Comparison of test methods, mixes with low water absorption level.