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.
53 C H A P T E R 4 The test methods and metrics described in Chapter 2 were employed in various field settings to evaluate their applica- bility to on-site locations, their usefulness in describing the binder foaming characteristics of various plant foaming units, and their ability to differentiate specimens prepared in the laboratory and the field. In addition, the proposed foamed mix design was validated with plant data. A. Initial Trial An initial plant field trial took place in Austin, Texas, on October 3, 2013. The template for field data acquisition shown in Appendix D was used to collect information about the plant characteristics. The plant type was a counterflow dryer drum (Figure 4-1) with a Maxam foaming unit installed in- line with an approximate 30-ft (9-m) run between the foamer and the inlet into the drum, an approximate 4- to 5-ft vertical rise, and extending into the drum about 8 ft. The pipeline was 4 in. in diameter and insulated. The day of the field test, con- ditions were fair and dry, temperatures were in the 70s, and a mild wind of 2 mph was blowing. The plant had a capac- ity of producing 300 tons of material each hour. The specific mix design being produced during the research teamâs visit is given in the following. The foaming unit was suspended approximately 10 ft above ground level. In order to sample the material from the foam- ing unit, a ¾-in. pipe extension approximately 6 ft long was inserted into the sampling port (Figure 4-2). The moisture level in the foamed binder was supposed to be 2.5%. A small water leak was noted on the foaming unit during its normal operation prior to sampling. The effect of this leak on the water content of the foam is not known, although measurements on the foamed binder indicated that foaming had occurred. Con- trol of the sampling valve required access with a ladder (Fig- ure 4-2) while the sample was collected below. In attempting to sample the foamed binder, it was found that the pipe extension caused the binder to cool and clog the line to the point that only a relatively thin stream of material flowed out of the pipe. After sampling in this manner, laser expansion and bubble size measurements were taken fol- lowing the procedure detailed in Appendix B. Two replicate measurements were done using the extension pipe. Later, a sample was taken from the sampling port with no pipe exten- sion. Although the foaming properties of this sample were more in line with what the team had seen in laboratory efforts (faster binder flow and better foaming), it was considered too dangerous to repeat sampling. Therefore, only one foaming measurement was done after the extension pipe was removed. A.1. Materials The binder used in this plant was a PG64-22 by NuStar Logistics, L.P. The optimum binder content was 5.1%. Lime- stone was the primary aggregate type used in the mix design. Other aggregates such as field sand and fractioned reclaimed asphalt pavement (RAP) were used. Details on the aggre- gate type, source, and gradation are presented in Table 4-1, Table 4-2, and Figure 4-3. A.2. Foaming Measurements A.2.1. Effect of Container Size A 1-gallon metal can was employed during the laboratory studies to conduct all foamed binder measurements. Because of concerns of potential splash and binder overflow when sampling in the field, a 5-gallon bucket was preferred for all field studies. Ideally, the expansion ratio of the foamed binder should be independent of the size of the container and the amount of binder dispensed since these variables are taken into account when calculating ER (i.e., determination of hfinal, using the density calculation method or the binder weightâ height calibration). The effect of container size was evaluated to verify the equivalency of the measurements using binders from two sources. Findings and Applications: Field Studies
54 Figure 4-1. Counterflow drum plant, Austin, Texas. Figure 4-2. Maxam Aquablack foam sampling port, Austin, Texas. Aggregate Type Bin No. 1 Bin No. 2 Bin No.3 Bin No.4 Bin No.5 Limestone Limestone Limestone Field Sand Fractioned RAP Aggregate Pit Marble Falls Marble Falls Marble Falls â â Table 4-1. Aggregate source, Austin, Texas. Sieve Analysis (Cum.% Passing) Aggregate Type % ¾â ½â â #4 #8 #30 #50 #200 Limestone 34.0 100 100 90.7 22.1 5.3 3.4 3.0 2.2 Limestone 11.0 100 100 100 69.1 7.1 2.1 1.7 1.1 Limestone 25.0 100 100 100 99.5 85.5 43.4 25.8 2.8 Field Sand 9.9 100 100 100 99.5 97.7 80.6 50.1 14.2 Fractioned RAP 20.1 100 100 94.0 68.1 49.3 31.5 23.2 7.1 Combined Gradation 100 100 100 95.6 63.5 43.5 26.5 17.3 4.4 3 8 Table 4-2. Aggregate gradation, Austin, Texas. Figure 4-3. Combined aggregate gradation, Austin, Texas. Note: Dashed purple line is the upper specification limit.
55 From the ERmax and k-value results for the first binder that are presented in Figure 4-4, it is apparent that the container size did have an influence in the measurements, with higher ER and k-value for the 1-gallon can measurements versus the measure- ments performed in the 5-gallon bucket. This difference was more pronounced at higher water contents (i.e., 5.5%). The k-value for the laboratory measurements done in the 5-gallon container with 1.5% water content was considerably higher than any of the other measurements. The ratios between the 1-gallon and 5-gallon average ERmax values measured in the laboratory were 1.7 and 2.1 for the 1.5% and 5.5% water contents, respec- tively. The difference between measurements using the 5-gallon container in the lab and in the plant was less pronounced. In this case, the largest difference in average ERmax values was for the 1.5% water content with a 1-gallon to 5-gallon ratio of 1.3. The second binder tested (i.e., binder R6 in Table 3-1) cor- roborated the observed differences in container size. ERmax and k-value are illustrated in Figure 4-5. As in the case of the first binder, the 1-gallon container had larger values than the 5-gallon container did, and the difference was more pronounced at higher water contents. In this case, the ratios between the 1-gallon and 5-gallon average ERmax values were 2.0 and 2.2 for the 1.0% and 3.0% water content, respectively. When combined with the values of the first binder tested, an average ratio between the 1-gallon and 5-gallon ERmax values of 2.0 was obtained. From these observations, it is apparent that the size of the container used to conduct the foaming measurements had an impact on the results. The rest of the field ERmax values collected in this study were corrected by a factor of 2.0 to make them comparable to the laboratory measurements. In addition, it is suggested to use the same size container when following the steps of the proposed mix design methodology or to compare the characteristics of different foamed binders. A.2.2. On-Site and Laboratory Measurements The average of the two replicate measurements with the extension pipe and the single measurement collected when the extension pipe was removed is presented along with the average of two replicate measurements performed in the laboratory using the Wirtgen foamer. Figure 4-6 presents the ERmax and k-value, while Figure 4-7 illustrates the FI obtained with the binder foam field and laboratory measurements. The corrected ERmax values correspond to the modified field measurements after applying the container size correction Figure 4-4. Comparison of ERmax and k-value for field and laboratory binder foam measurements using 1-gallon and 5-gallon containers. Figure 4-5. Comparison of ERmax and k-value for binder R6 using 1-gallon and 5-gallon containers. Figure 4-6. On-site and laboratory ERmax and k-value for the initial trial in Austin, Texas. Figure 4-7. ERmax and FI for on-site and laboratory binder foam measurements for the initial trial in Austin, Texas.
56 factor of 2.0. The error bars in Figure 4-6 span ±1 standard deviation from the average ERmax value. From Figure 4-6, it is apparent that the ERmax of the field specimen without the extension pipe was only about 65% of the measurement obtained when the extension pipe was used. As mentioned before, the pipe extension likely caused the binder to cool and clog the line. Also, the ERmax for both field measure- ments (with and without the extension pipe) was significantly lower than the ERmax values obtained using the Wirtgen foamer, even after the container size correction factor was applied. As before, ERmax and k-value had a direct correlation with the water content used to foam the binder; the higher the water content, the higher ERmax and k-value (less stable foam). The FI presented in Figure 4-7 illustrates the area under the ER curve at 60 s after foaming. Only the laboratory FI values are presented since the differences in container size yielded smaller field FI values that were not comparable on the same scale. The FI values were similar for all water contents. The SAI is illustrated in Figure 4-8. Contrary to the FI trends that showed practically no difference between water contents, the SAI shows clear distinction between mixtures foamed with 1.0%, 2.0%, 2.5%, and 3.0% water content. The larger the water content, the smaller the SAI, which indicates that the binder is less stable after foaming, and fewer small- sized, long-lasting bubbles are formed. A.3. Mixture Workability Measurements The workability results for plant-mix laboratory-compacted (PMLC) and LMLC specimens fabricated at 1.0%, 2.0%, 2.5%, and 3.0% foaming water contents are shown in Figure 4-9. Each bar in Figure 4-9 represents the average value of three replicates, and the error bars represent ±1 standard deviation from the average value. As shown in Figure 4-9, a significantly lower maximum shear stress (tmax) value was shown for the PMLC specimen as com- pared to LMLC specimens at different foaming water contents, which indicated that the PMLC specimen had a better workabil- ity than the LMLC counterparts. The comparison in workabil- ity of LMLC specimens with different foaming water contents illustrated that the LMLC specimen at 1.0% foaming water content had better workability characteristics (indicated by a lower tmax value) as compared to those foamed at higher water contents (i.e., 2.0%, 2.5%, and 3.0%). Therefore, 1.0% was the optimum foaming water content for this certain mixture. A.4. Performance Evaluation To evaluate the performance of foamed mixtures at the optimum foaming water content, a new set of LMLC speci- mens were fabricated at 1.0% foaming water content using the Wirtgen foamer. In addition, another set of HMA LMLC specimens was also produced. Both the foamed mixture and the control HMA were mixed at 305°F (152°C) and then short-term aged for 2 hours at 275°F (135°C) prior to compaction. Then, the compacted specimens were tested for MR, IDT strength test, and HWTT. Testing parameters, including MR stiffness, wet IDT strength, and TSR at 77°F (25°C), and HWTT LCSN, LCST, and DevpSN at 122°F (50°C) and 104°F (40°C), were used to evaluate mixture stiffness, rutting resistance, and moisture susceptibility and to compare the performance of the foamed mixture versus the control HMA. A.4.1. Resilient Modulus Test The MR stiffness results for the foamed mixture and the control HMA are shown in Figure 4-10. Each bar in Fig- ure 4-10 represents the average MR stiffness value of three replicates, and the error bars represent ±1 standard deviation from the average value. As illustrated in Figure 4-10, equivalent MR stiffness at 77°F (25°C) was achieved by the control HMA and the foamed mixture at 1.0% foaming water content. Thus, in this case the inclusion of water from the foaming process did not reduce the mixture stiffness. Figure 4-8. FI and SAI for on-site and laboratory binder foam measurements for the initial trial in Austin, Texas. Figure 4-9. Workability results for Austin mixtures.
57 A.4.2. Indirect Tensile Strength Test Three specimens were tested in a dry condition and three specimens after moisture conditioning following the proce- dure outlined by AASHTO T 283. The wet IDT strength and TSR value at 77°F (25°C) were considered as indicators of mixture moisture susceptibility. The IDT strength test results for the foamed mixture and the control HMA are shown in Figure 4-11. The solid bar and pattern-filled bar represent the average dry and wet IDT strength, respectively. The error bars represent ±1 standard deviation from the average value. In addi- tion, the TSR values are shown on top of the IDT strength bars. As illustrated in Figure 4-11, equivalent dry IDT strength at 77°F (25°C) was achieved by the control HMA and the foamed mixture at 1.0% foaming water content. However, a significantly higher wet IDT strength and, subsequently, a higher TSR value, were shown by the foamed mixture as compared to the control HMA, which indicated that the foamed mixture produced at the optimum foaming water content had lower moisture sus- ceptibility than the control HMA in the IDT strength test. A.4.3. Hamburg Wheel Tracking Test The HWTT was performed at 122°F (50°C) following AASHTO T 324, and test parameters, including LCSN, LCST, and DevpSN, were used to evaluate moisture susceptibility and rutting resistance. HWTT results for the control HMA and the foamed mixture at 1.0% foaming water content are shown in Figure 4-12. As illustrated, the mixtures did not pass the failure criteria of 20,000 load cycles with less than 0.5-in. (12.5-mm) rut depth. Figure 4-13 shows the LCSN, LCST, and DevpSN of the foamed Figure 4-10. MR test results for Austin mixtures. Figure 4-11. IDT strength test results for Austin mixtures. Figure 4-12. HWTT rut depth versus load cycles at 122î¸F (50î¸C) for Austin mixtures. (a) (b) Figure 4-13. HWTT results at 122î¸F (50î¸C) for Austin mixtures; (a) LCSN and LCST, and (b) Dî« vp SN.
58 mixture versus the control HMA at the test temperature of 122°F (50°C). As shown in Figure 4-13(a), slightly higher LCSN and LCST values were shown for the control HMA as compared to the foamed mixture, indicating better moisture resistance. The LCSN values for both mixtures were less than 2,000 load cycles, and therefore, both mixtures experienced early stripping during the test at 122°F (50°C). Test results presented in Figure 4-13(b) illustrated that the control HMA had a DevpSN approximately 1.5 me/cycle (equivalent to rut depth of 0.09 mm per 1,000 load cycles) higher than the foamed mixture did. Thus, better rutting resistance was exhibited by the foamed mixture at 1.0% foaming water content than the control HMA in the HWTT at 122°F (50°C). Laboratory experience with the HWTT shows that for an asphalt mixture exhibiting early stripping (i.e., LCSN less than 2,000 load cycles), the determination of the viscoplastic deformation and the rutting resistance parameter DevpSN is likely to be biased due to the limited duration of the creep phase. Therefore, to better evaluate the rutting resistance of the foamed mixture and the control HMA in the HWTT, the test was performed at a lower temperature [i.e., 102°F (40°C)], with HWTT results in terms of rut depth versus load cycle shown in Figure 4-14. As illustrated, no stripping occurred for either mixture during the entire test (LCSN equaled 20,000 load cycles), and both mixtures passed the failure criteria of 20,000 load cycles with less than 0.5-in. (12.5-mm) rut depth. The DevpSN results at 102°F (40°C) for the foamed mixture and the control HMA are shown in Figure 4-15. As illustrated, both mixtures had insignificant viscoplastic deformation (indicated by DevpSN values lower than 10-6 me/cycle) due to the wheel load, and more specifically, a higher DevpSN value was shown for the control HMA as compared to the foamed mixture at 1.0% foaming water content. Thus, the foamed mixture at 1.0% foaming water content had better rutting resistance at 40°C in the HWTT than the control HMA, despite the inclusion of additional moisture involved in the foaming process. B. Comparison of Plant Foaming Units Two plants in the Cincinnati, Ohio, area and operated by the same material supplier were visited by the research team on November 5 and 6, 2013, although originally three plants had been planned. The temperatures were falling at this point in time, and operations for this contractor were ending. Since the source of binder for the two plants was the same, this was seen as an opportunity to test two different plant foamers using the same binder. This trip was facilitated by the con- tractorâs laboratory, which possessed SGCs with shear stress measuring capabilities. Thus, the research team was able to test PMLCs shortly after they were mixed. B.1. Description B.1.1. Plant 1 This plant, located in Cleves, Ohio, was a counterflow plant with a continuous pugmill or coater box (Figure 4-16) to pro- vide mixing. The plant had a capacity of 300 tons per hour Figure 4-14. HWTT rut depth versus load cycles at 104î¸F (40î¸C). Figure 4-15. HWTT Dî«vpSN results at 104î¸F (40î¸C) for Austin mixtures.
59 with a Gencor foamer (Figure 4-17), and the details of the mix being produced during the teamâs visit are given in the following. The foaming water content used during production was 1.5%. The foaming unit was located about 12 ft from the binder inlet to the coater box with a 3-ft drop and 5 ft to the sampling port. The binder line was 4 in. in diameter. Tempera- tures for that day were around 50°F, the skies were overcast, winds were calm, and rain was intermittent. The sampling port for the foaming unit was about 3 ft away from the foaming unit itself, and although it was acces- sible from a platform, there was low overhead clearance (Figure 4-18). Because the sampling port had not been used before, it was necessary to heat the valve in order to turn it (Figure 4-18). Once the valve was loosened sufficiently to open it, the flow of the binder was restricted (Figure 4-19), possibly due to material being caked in the valve. Once the sample was taken, it needed to be handed down and to the testing area, which required the sampler handing the material to another member of the staff that carried it down a ladder to the testing area (Figure 4-20). B.1.2. Plant 2 The second plant tested in the Cincinnati area was a Terex counterflow plant (Figure 4-21) located in Fairborn, Ohio, with a 270-ton/hour capacity and a Terex foamer (Figure 4-22). The foaming water content used during pro- duction was 2.2%. The binder line was 4 in. in diameter. The distance from the foaming unit to the drum inlet was 10 ft, and the sampling port was located on the foaming unit. The Figure 4-16. Coater box, Plant 1, Cincinnati, Ohio. Figure 4-17. Gencor foamer, Plant 1, Cincinnati, Ohio. Figure 4-18. Location of foam sampling port and heating to loosen valve, Plant 1, Cincinnati, Ohio. Figure 4-19. Foamed binder flow from sampling port, Plant 1, Cincinnati, Ohio.
60 line from the sampling port was 1 in. in diameter and 10 in. long (Figure 4-23). B.2. Materials B.2.1. Plant 1 The binder used in Plant 1 was a Marathon PG64-22, although the contractor later specified that the tank at the ter- minal had a blend of several sources. The binder content used in the mix design produced that day was 5.6%. Local gravel #8 was the primary aggregate used. Other aggregates, such as local natural sand and fractioned RAP, were also used. Details on the aggregate type, source, and gradation are presented in Table 4-3, Table 4-4, and Figure 4-24. Figure 4-20. Overall picture of foam sampling area relative to testing area. Figure 4-21. Terex counterflow drum, Plant 2, Fairborn, Ohio. Figure 4-22. Terex foamer, Plant 2, Fairborn, Ohio. Figure 4-23. Foam sampling port, Plant 2, Fairborn, Ohio. B.2.2. Plant 2 The same binder as used in Plant 1 was employed at this location. The binder content used in the mixture produced that day was 3.9%. Local stone, gravel, natural sand, and 40% crushed RAP were used in the mixture. Details on the aggre- gate type, source, and gradation are presented in Table 4-5, Table 4-6, and Figure 4-25.
61 Aggregate Source Bin No. 1 Bin No. 2 Bin No.3 Gravel #8 Natural Sand RAP Aggregate Pit Martin Marietta/Fairfield, OH Martin Marietta/E-Town, OH Barrett/Cleves Table 4-3. Aggregate source, Plant 1, Cincinnati, Ohio. Table 4-4. Aggregate gradation, Plant 1, Cincinnati, Ohio. Sieve Analysis (Cum.% Passing) Combined Gradation ½â #4 #8 #16 #30 #50 #100 #200 100 97.0 65.1 40.8 30.1 20.3 10.1 6.3 4.6 â 3 8 Figure 4-24. Combined aggregate gradation, Plant 1, Cincinnati, Ohio. Note: Dashed purple line is the upper specification limit. Table 4-5. Aggregate source, Plant 2, Fairborn, Ohio. Aggregate Source Bin No. 1 Bin No. 2 Bin No. 3 Bin No.4 Stone Gravel Natural Sand Crushed RAP Aggregate Pit Barrett Paving/Miami River Stone Quarry Barrett Paving/Fairborn Barrett Paving/ Fairborn Barrett/ Fairborn Table 4-6. Aggregate gradation, Plant 2, Fairborn, Ohio. Sieve Analysis (Cum.% Passing) Combined Gradation 1½â 1⠾⠽â #4 #8 #16 #30 #50 #100 #200 100 86 70 60 52 40 31 22 15 9 5 4.2 â 3 8
62 B.3. Foaming Measurements Average ERmax and k-value of three replicate measure- ments performed on-site and two replicate measurements performed in the laboratory using the Wirtgen foamer are presented in Figure 4-26. The corrected ERmax values corre- spond to the modified field measurements after applying the container size correction factor of 2.0. The error bars in Figure 4-26 span ±1 standard deviation from the average ERmax value. From the figure, it is apparent that ERmax and k-value for both field measurements were lower than the laboratory measurements, even after the container size correction factor was applied. Within the field measurements, the Terex foamed binder with 2.2% water content had a slightly higher ERmax and a lower k-value than the Gencor foamed binder with 1.5% water content. The laboratory test results performed with the Wirtgen foamer showed that the binder was not sensitive to increasing water content, showing almost no change in ERmax with added water. By contrast, the laboratory results obtained using the binder from the initial field trial showed increasing ERmax with increasing water content (Figure 4-26). Figure 4-27 illustrates the laboratory FI of the same mea- surements. Only the laboratory FI values are presented since the differences in container size yielded smaller field FI val- ues that were not comparable on the same scale. Opposite to the almost constant ERmax value observed for the laboratory measurements, the FI at 60 s did show differences, with lower water contents showing larger FI values. Binder foamed in the laboratory with 2.0%, 2.2%, and 3.0% water content had very similar FI values. The SAI results are illustrated in Figure 4-28. As in the case of the FI at 60 s, clear differences were observed, with the binder foamed at lower water contents having higher SAI val- ues. This indicates that the bubbles at lower water contents Figure 4-26. ERmax and k-value for the Ohio on-site and laboratory binder foam measurements. Figure 4-27. ERmax and FI for the Ohio laboratory binder foam measurements. Figure 4-25. Combined aggregate gradation, Plant 2, Fairborn, Ohio. Note: Dashed purple line is the upper specification limit.
63 were smaller and lasted longer. The SAI values decreased pro- gressively as the water content increased. B.4. Mixture Workability and Coatability Measurements The workability and coatability results for foamed PMLC and LMLC specimens fabricated at 1.0%, 2.0%, and 3.0% foaming water contents and HMA LMLC specimens are shown in Figure 4-29 and Figure 4-30, respectively. Each bar represents the average maximum shear stress value of three replicates, and the error bars in Figure 4-29 represent ±1 stan- dard deviation from the average value. As illustrated in Figure 4-29, a significantly lower maxi- mum shear stress value was shown for the foamed PMLC specimens as compared to foamed LMLC specimens at different foaming water contents, which indicated that the foamed PMLC specimens had a better workability than their LMLC counterparts. The comparison in workability of LMLC specimens illustrated that an equivalent maximum shear stress value was achieved among foamed mixtures at different foam- ing water contents and the control HMA. The comparison in coatability results of HMA versus foamed LMLC specimens at different foaming water contents shown in Figure 4-30 illustrates that all foamed specimens had higher CI values as compared to their HMA counterpart. In addition, foamed specimens at 1.0% foaming water content had the best CI as compared to those specimens prepared with higher water contents (i.e., 2.0% and 3.0%). Therefore, according to the workability and coatability results presented in Figure 4-29 and Figure 4-30, the opti- mum foaming water content for this specific mixture was 1.0% and was able to produce LMLC specimens with the best workability and coatability characteristics. B.5. Performance Evaluation To evaluate the performance of foamed mixtures at the optimum foaming water content, a new set of LMLC speci- mens was fabricated at 1.0% foaming water content using the Wirtgen foamer. In addition, another set of HMA specimens was also produced. Both the foamed mixture and the control HMA were mixed at 300°F (149°C) and then short-term aged for 2 hours at 275°F (135°C) prior to compaction. Then, the compacted specimens were tested for MR, IDT strength, and HWTT. Testing parameters included MR stiffness, wet IDT strength, and TSR at 77°F (25°C). HWTT LCSN, LCST, and DevpSN at 122°F (50°C) were used to evaluate mixture stiffness, rutting resistance, and moisture susceptibility and to compare the performance of foamed mixtures at the optimum foaming water content versus the control HMA. B.5.1. Resilient Modulus Test The MR stiffness results for the control HMA and the foamed mixture at the optimum foaming water content (i.e., 1.0%) are shown in Figure 4-31. Each bar presents the average MR stiff- ness of three replicates, and the error bars represent ±1 standard deviation from the average value. As illustrated in Figure 4-31, equivalent MR stiffness at 77°F (25°C) was achieved by the control HMA and the foamed Figure 4-28. FI and SAI for the Ohio laboratory binder foam measurements. Figure 4-29. Workability results for the Ohio Plant 1 mixtures. Figure 4-30. Coatability results for the Ohio Plant 1 mixtures.
64 mixture at 1.0% foaming water content. Thus, the inclusion of water from the foaming process had no significant effect on the mixture stiffness. B.5.2. Indirect Tensile Strength Test Three specimens were tested in a dry condition and three specimens after moisture conditioning following the pro- cedure outlined by AASHTO T 283. The wet IDT strength and TSR value at 77°F (25°C) were considered as indica- tors of mixture moisture susceptibility. The IDT strength test results for the foamed mixture and the control HMA are shown in Figure 4-32. The solid bar and pattern-filled bar represent the average dry and wet IDT strength, respectively. The error bars represent ±1 standard deviation from the average value. In addition, the TSR values are shown above the IDT strength bars. As illustrated in Figure 4-32, a slightly higher dry IDT strength at 77°F (25°C) was observed for the control HMA as compared to the foamed mixture at 1.0% foaming water content. However, an equivalent wet IDT strength and, sub- sequently, a higher TSR value were shown by the foamed mixture as compared to the control HMA. Therefore, IDT strength test results indicated that the foamed mixture fab- ricated at the optimum foaming water content had slightly better moisture susceptibility than the control HMA. B.5.3. Hamburg Wheel Tracking Test The HWTT was performed at 122°F (50°C) following AASHTO T 324, and test parameters, including LCSN, LCST, and DevpSN, were used to evaluate mixture moisture suscep- tibility and rutting resistance. HWTT results in terms of rut depth versus load cycle for the control HMA and the foamed mixture at 1.0% foaming water content are shown in Figure 4-33. Both mixtures passed the failure criteria of 20,000 load cycles with less than the 0.5-in. (12.5-mm) rut depth of TxDOT spec- ifications. Figure 4-34 presents the comparison in LCSN, LCST, and DevpSN results of the foamed mixture versus the control HMA. As illustrated in Figure 4-34(a), higher LCSN and LCST values were shown for the control HMA as compared to the foamed mixture, indicating better moisture resistance in the HWTT. Test results presented in Figure 4-34(b) illustrate that the foamed mixture at the optimum foaming water content had a DevpSN value approximately 0.5 me/cycle [equivalent to rut depth of 0.0012 in. (0.03 mm) per 1,000 load cycles] higher than the control HMA did. Thus, worse moisture susceptibility and rutting resistance were exhibited by the foamed mixture at the optimum foaming water content than the control HMA in the HWTT. C. Field Validation of Proposed Mix Design The field validation of the proposed foamed mix design presented in Section 2.C was done at a plant located in Huntsville, Texas. This plant consisted of an Astec Double Figure 4-31. MR test results for the Ohio Plant 1 mixtures. Figure 4-32. IDT strength test results for the Ohio Plant 1 mixtures. Figure 4-33. HWTT rut depth versus load cycles at 122î¸F (50î¸C) for the Ohio Plant 1 mixtures.
65 Barrel with a Stansteel Accushear foaming unit. The foaming unit was located on its own platform (Figure 4-35) with an approximately 15-ft run to the inlet on the drum with a 3-ft drop. The plant was capable of producing 250 tons of mix per hour; the mix design details are given later. The foam sam- pling port on this plant was located about 3 ft downstream from the foaming unit with a ¾-in. sampling port. Although Figure 4-34. HWTT results at 122î¸F (50î¸C) for the Ohio Plant 1 mixtures; (a) LCSN and LCST, and (b) Dî« vp SN. (a) (b) Figure 4-35. Foaming unit on platform, Huntsville, Texas. Figure 4-36. Heating valve prior to sampling foam, Huntsville, Texas. Figure 4-37. Free-flowing foamed binder sample, Huntsville, Texas. the valve needed to be heated prior to sampling the foamed binder (Figure 4-36), there were no problems taking and test- ing samples from this plant (Figure 4-37). C.1. Experimental Plan Figure 4-38 presents the experimental design followed in this portion of the study. During the first plant visit, on November 29, 2013, the plant was producing foamed mixtures with 5.5% water content at approximately 300°F (149°C). Binder foam- ing characteristics produced by the plant foaming unit were measured on-site using a laser device and a digital camera on a side platform. Foamed loose mix produced at the plant was sampled from the trucks after being loaded from the silo and then transported back to the laboratory (approximately 60 miles away from the plant) for fabricating PMLC specimens and evaluated for their workability. Raw materials, including virgin asphalt binder, aggregates, and RAP, were also sampled during the visit. In the laboratory, asphalt was foamed using a Wirtgen foamer at the following water contents: 0.7% (which was the
66 minimum water content the equipment was able to output), 1.0%, 1.5%, 2.0%, 3.0%, and 5.5%. Evaluation of binder foam- ing characteristics and foamed mixture workability was per- formed to determine an optimum foaming water content, referring to the specific water content at which the laboratory- foamed mixture had the best workability. Afterward, foamed LMLC specimens were fabricated at the optimum foaming water content for performance evaluation and comparison to the performance of the HMA. A second visit to the plant was made on December 3, 2013, adjusting the water content on-site to match the optimum value obtained via laboratory measurements. Binder foam- ing characteristics produced by the plant foaming unit at the adjusted water content were measured, and samples of the foamed loose mix were collected for workability and perfor- mance evaluation. The foamed plant loose mix acquired during produc- tion was used to fabricate PMLC specimens by reheat- ing in an oven at 275°F (135°C). The temperature of the loose mix was monitored using a digital thermometer every 15 minutes after being reheated for 1 hour. Once the tem- perature achieved 275°F (135°C), the foamed loose mix was compacted in the SGC. LMLC specimens were produced at a mixing temperature of 300°F (149°C) and then short-term aged for 2 hours at 275°F (135°C) prior to compaction. For mixture performance evaluation, PMLC and LMLC speci- mens of both plant-produced and laboratory-produced foamed mixtures were tested for MR, IDT strength, and by the HWTT to evaluate mixture stiffness, rutting resistance, and moisture susceptibility, and to compare the performance versus the control HMA. C.2. Materials The binder used was a Valero PG64-22. The optimum binder content per mix design was 4.5%. Limestone was the primary aggregate used in the mixture, and sandstone and fractioned RAP were also used. Details on the aggregate type, source, and gradation are presented in Table 4-7, Table 4-8, and Figure 4-39. C.3. Initial Measurements at Plant C.3.1. Foaming The ERmax and k-value of the on-site binder foaming measurements performed are illustrated in Figure 4-40. The bars and red squares represent the average ERmax and k-value of three replicate measurements, respectively. The error bars span ±1 standard deviation from the average ERmax value. The initial plant measurement with a water content of about 5.5% yielded an ERmax of about 12 and a k-value of 0.04. The repeatability of the measurements seemed adequate. The FI at 60 s is illustrated in Figure 4-41. The FI for the initial plant measurement had a steady increase with time. C.3.2. Workability The workability results for plant-produced HMA and laboratory-foamed mixture with 5.5% water content are shown in Figure 4-42. Each bar in Figure 4-42 represents the average value of three replicates, and the error bars represent Figure 4-38. Experimental plan for the field validation of the foamed mix design.
67 Aggregate Source Bin No. 1 Bin No. 2 Bin No.3 Bin No.4 Bin No.5 Sandstone Limestone Limestone Austin White Lime Fractioned RAP Aggregate Pit Marble Falls Marble Falls Marble Falls â â Table 4-7. Aggregate source, Huntsville, Texas. Table 4-8. Aggregate gradation, Huntsville, Texas. Sieve Analysis (Cum.% Passing) Aggregate Source % 1â ¾â #4 #8 #30 #50 #200 Sandstone 28.0 100 99.0 12.0 6.0 4.3 3.9 3.7 2.5 Limestone 24.1 100 100 91.0 22.0 4.2 2.4 2.2 1.5 Limestone 27.0 100 100 100 99.0 80.0 39.0 25.0 4.2 Austin White Lime 1.0 100 100 100 100 100 100 100 100 Fractioned RAP 19.9 100 100 94.0 68.1 49.3 31.5 23.2 7.1 Combined Gradation 100 100 100 95.5 78.0 62.5 40.0 17.6 3.0 â 3 8 Figure 4-40. ERmax and k-value for the Huntsville on-site binder foam measurements. Figure 4-41. ERmax and FI for the Huntsville on-site binder foam measurements. Figure 4-39. Combined aggregate gradation, Huntsville, Texas. Note: Dashed purple line is the upper specification limit.
68 ±1 standard deviation from the average value. As illustrated, a lower tmax value was exhibited for the control HMA versus the laboratory-foamed mixture with 5.5% water content, indi- cating that plant-produced HMA had a better workability than the foamed mixture. C.4. Laboratory Mix Design The proposed mix design method illustrated in Figure 2-9 was applied to the materials collected at the Huntsville plant. Once the foaming ability of the binder was corroborated, the traditional HMA mix design used during production was replicated in the laboratory to estimate the optimum water content via workability and coatability measures. C.4.1. Foaming Measurements The binder foaming measurements were performed in the laboratory using the Wirtgen foamer. The water content was varied from 0.7% (which was the minimum water level the equipment was able to output) to 3.0%. The following water contents were tested: 0.7%, 1.0%, 1.5%, 2.0%, 3.0%, and 5.5%. The ERmax and k-value results are illustrated in Figure 4-43. As before, the bars represent the average ERmax of two replicates, the dots correspond to the average k-value, and the error bars span ±1 standard deviation from the average value. A direct correlation was observed between the amount of water added to the foaming process and the ERmax value. The foam was relatively stable, with a low k-value for almost all samples, except for the one with 5.5% water content, which showed a very large value. The FI results are presented in Figure 4-44. Consistent with the ERmax and k-value results, the FI at 60 s also showed the lowest values for the sample foamed with 5.5% water content. The samples with lower water contents, from 0.7% through 2.0%, showed higher and equivalent FI values at 60 s. Therefore, as more water was used in the foaming pro- cess, the binder foaming achieved a higher volume expansion but also a faster foam collapse rate, which ultimately led to lower stability. However, an opposite trend was observed for plant binder foaming measurements, where foamed binder at 1.5% water content exhibited lower stability, as indicated by a lower FI value than that at a higher water content of 5.5% (Fig- ure 4-41). In addition, the comparison between plant binder foaming versus laboratory binder foaming showed that plant binder foaming had a significantly higher FI value at 60 s than the laboratory-foamed sample at the same water contents. The SAI values obtained from the analysis of the digital images acquired during the foaming process are illustrated in Figure 4-45. As previously mentioned, this parameter takes into consideration the total number of foaming bubbles and the bubble size distribution. As time after foaming increases, the decrease in ER and the increase in surface area of the foaming bubbles (due to the transition from larger bubbles to smaller bubbles) are competing mechanisms. Foamed binder with larger SAI values (throughout the dynamic foaming process) is expected to have better aggregate coating ability. For the results shown in Figure 4-45, SAI could not be calcu- lated for 3.0% and 5.5% water content due to rapid bubble collapse, which also corresponds to lower values of FI. As will be explained in more detail in the next section, the maximum shear stress also follows a decreasing trend up Figure 4-42. Workability results for the Huntsville plantâproduced HMA and laboratory-foamed mixture with 5.5% water content. Figure 4-43. ERmax and k-value for the Huntsville laboratory binder foam measurements. Figure 4-44. Foamability index versus foaming time for the Huntsville laboratory binder foam measurements.
69 to a point where the maximum shear stress increases again (Figure 4-46). These parallel observations suggest that either workability or binder foamed characteristics could be used during mix design to obtain the optimum water content. C.4.2. Mixture Workability Measurements The workability results for LMLC specimens produced at 0.7%, 1.0%, 1.5%, 2.0%, 3.0%, and 5.5% foaming water contents are shown in Figure 4-46. Each bar in Figure 4-46 represents the average value of three replicates, and the error bars represent ±1 standard deviation from the average value. As illustrated, tmax values of laboratory-produced foamed mixtures decreased as foaming water contents increased from 0.7% to 1.5%, while the opposite trend was shown for foam- ing water contents higher than 1.5%. Therefore, 1.5% was selected as the optimum foaming water content that was able to produce foamed mixture with the best workability charac- teristic (the lowest tmax value). C.5. Recommended Plant Adjustments During a second visit to the Huntsville plant, the foaming water content was modified to 1.5% to match the optimum obtained via the laboratory measurements. On-site binder foaming measurements and loose mix were collected during the second visit to the plant to compare against the values collected during the first visit. C.5.1. Foaming Measurements The on-site foaming measurements are shown in Fig- ure 4-40 and Figure 4-41. In contrast to the first visit, ERmax and k-value were lower and the FI at 60 s smaller than when 5.5% water content was used. This observation is opposite to what has been observed in the laboratory measurements, where lower water contents seem to have better FI. C.5.2. Mixture Workability Measurements The workability results for plant-produced foamed mix- ture at adjusted foaming water content (1.5%) were com- pared to the results of plant-produced HMA and foamed mixtures at 5.5% foaming water content and are shown in Figure 4-47. Each bar in Figure 4-47 represents the average value of three replicates, and the error bars represent ±1 stan- dard deviation from the average value. Plant-produced foamed mixture at 1.5% foaming water content had a better workability characteristic (indicated by a lower tmax value) as compared to both the control HMA and the foamed mixture with 5.5% water content. Therefore, the optimum foaming water content of 1.5%, which was deter- mined by laboratory foaming, was verified by plant foaming. C.5.3. Performance Evaluation To fabricate PMLC specimens for performance evaluation, plant loose mix acquired during production was taken out of buckets and reheated in an oven at 275°F (135°C). The temperature of the loose mix was monitored using a digi- tal thermometer every 15 minutes after being reheated in Figure 4-45. Surface area index for the Huntsville laboratory binder foam measurements. Figure 4-46. Workability results for the Huntsville laboratory-produced foamed mixtures. Figure 4-47. Workability results for plant-produced HMA and foamed mixtures at 1.5% and 5.5% foaming water contents.
70 the oven for 1 hour. Once the temperature achieved 275°F (135°C), plant loose mix specimens were compacted using an SGC. Then, the compacted specimens were tested for MR, IDT strength, and by the HWTT to evaluate mixture stiffness, rutting resistance, and moisture susceptibility. C.5.3.1. Resilient Modulus Test. The MR stiffness results for plant-produced HMA and foamed mixtures at 1.5% foaming water content are shown in Figure 4-48. Each bar in Figure 4-48 represents the average MR stiffness value at 77°F (25°C) of three replicates, and the error bars represent ±1 standard deviation from the average value. A slightly lower MR stiffness was shown by foamed mixture at 1.5% foaming water content as compared to the HMA. The larger stiffness for HMA was possibly due to additional aging expe- rienced by the mixture in the plant versus the laboratory setting. C.5.3.2. Indirect Tensile Strength Test. Three PMLC specimens were tested in a dry condition and three specimens after moisture conditioning following the procedure outlined by AASHTO T 283. The wet IDT strength and TSR value at 77°F (25°C) were considered as indicators of mixture mois- ture susceptibility. The IDT strength test results for plant- produced foamed mixture and control HMA are shown in Figure 4-49. The solid bar and pattern-filled bar represent the average dry and wet IDT strength, respectively. The error bars represent ±1 standard deviation from the average value. In addition, the TSR values are shown on top of the IDT strength bars. Equivalent dry and wet IDT strength at 77°F (25°C) and, consequently, an equal TSR value were achieved by HMA and foamed mixture at 1.5% foaming water content. Therefore, plant-produced HMA and foamed mixture at the optimum foaming water content had equivalent moisture susceptibility in the IDT strength test. C.5.3.3. Hamburg Wheel Tracking Test. The HWTT was performed at 122°F (50°C) following AASHTO T 324, and test parameters including LCSN, LCST, and DevpSN were used to evaluate mixture moisture susceptibility and rutting resis- tance. HWTT results in terms of rut depth versus load cycle for the control HMA and foamed mixture at 1.5% foaming water content are shown in Figure 4-50. Both mixtures passed the failure criteria of 20,000 load cycles with less than 0.5-in. (12.5-mm) rut depth of TxDOT specifications. Figure 4-51 presents the comparison of plant- produced foamed mixture versus control HMA in LCSN, LCST, and DevpSN. As shown in Figure 4-51(a), the foamed mixture had a significantly higher LCSN value than the control HMA, indicating better moisture susceptibility before stripping occurred. A LCST value of 45,627 was obtained for the plant- produced HMA. However, the determination of LCST for the foamed mixture was not available since no stripping occurred within the mixture during the test (indicated by a LCSN value of 20,000). Test results shown in Figure 4-51(b) illustrate Figure 4-48. MR test results for the Huntsville plant-produced HMA versus foamed mixtures at 1.5% foaming water content. Figure 4-49. IDT strength test results for the Huntsville plant-produced HMA versus foamed mixtures at 1.5% foaming water content. Figure 4-50. HWTT rut depth versus load-cycle results for the Huntsville plant-produced HMA versus foamed mixtures at 1.5% foaming water content.
71 that plant-produced foamed mixture at 1.5% foaming water content had a better rutting resistance than the HMA counterpart, as indicated by a lower DevpSN value. Thus, better moisture susceptibility and rutting resistance were exhibited by the plant-produced foamed mixture at 1.5% foaming water content than by the HMA counterpart in the HWTT. C.6. Performance Evaluation To evaluate the performance of laboratory-produced foamed mixtures at the optimum foaming water content, a new set of LMLC specimens were fabricated at 1.5% foaming water content using the Wirtgen foamer. In addition, a compan- ion set of HMA LMLC specimens were produced as well. Both the foamed mixtures and control HMA were mixed at 300°F (149°C) and then short-term aged for 2 hours at 275°F (135°C) prior to compaction. Then, the compacted specimens were tested for MR, IDT strength, and by the HWTT. Testing parameters, including MR stiffness, wet IDT strength, TSR at 77°F (25°C), and HWTT LCSN, LCST, and DevpSN at 122°F (50°C), were used to evaluate mixture stiff- ness, rutting resistance, and moisture susceptibility and to compare the performance of foamed mixtures versus the control HMA. C.6.1. Resilient Modulus Test The MR stiffness results for plant- and laboratory-produced foamed mixtures at the optimum foaming water content of 1.5% and the control HMA are shown in Figure 4-52. Each bar in Figure 4-52 represents the average MR stiffness value at 77°F (25°C) of three replicates, and the error bars represent ±1 standard deviation from the average value. As illustrated, a slightly lower MR stiffness was shown by plant-produced foamed mixtures as compared to plant-produced HMA. Equivalent stiffness was exhibited for a laboratory-produced foamed mixture versus the HMA counterpart. Therefore, the inclusion of water as part of the foaming process did not reduce the mixture stiffness in this case. C.6.2. Indirect Tensile Strength Test The IDT strength test results for plant- and laboratory- produced foamed mixtures and the control HMA are shown in Figure 4-53. The solid bar and pattern-filled bar represent the average dry and wet IDT strength, respectively, after moisture conditioning according to AASHTO T 283 with partial vac- uum saturation, one freezeâthaw cycle, and soaking in warm water of three replicates. The error bars represent ±1 standard deviation from the average value. In addition, the TSR values are shown at the top of each pair of IDT strength bars. The IDT strength test results for plant-produced mixtures indicated that equivalent dry and wet IDT strength at 77°F (25°C) and, consequently, an equal TSR value of approxi- mately 100% were achieved by both HMA and the foamed mixture at 1.5% foaming water content. For laboratory- produced mixtures, equivalent dry IDT strength at 77°F Figure 4-51. HWTT results for the Huntsville plant-produced HMA versus foamed mixtures at 1.5% foaming water content; (a) LCSN and LCST, and (b) Dî« vp SN. (a) (b) Figure 4-52. MR test results for the Huntsville plant- and laboratory-produced HMA versus foamed mixtures at 1.5% foaming water content.
72 (25°C) was achieved between the foamed mixture at 1.5% foaming water content and the control HMA. However, a sig- nificantly higher wet IDT strength and, subsequently, a higher TSR value were shown for the foamed mixture as compared to its HMA counterpart, which indicated that laboratory- produced foamed mixtures produced at the optimum foam- ing water contents had better moisture susceptibility than the control HMA in the IDT strength test. C.6.3. Hamburg Wheel Tracking Test HWTT results in terms of rut depth versus load cycle for laboratory-produced foamed mixture at 1.5% foaming water content and the control HMA are shown in Figure 4-54. As illustrated, both laboratory-produced mixtures passed the failure criteria of 20,000 load cycles with less than 0.5-in. (12.5-mm) rut depth. Additionally, the shape of the HWTT curve indicated no sign of stripping occurring during the test for either mixture. Figure 4-53. IDT strength test results for the Huntsville plant- and laboratory-produced HMA versus foamed mixtures at 1.5% foaming water content. Figure 4-55 presents the LCSN, LCST, and DevpSN results of plant- and laboratory-produced foamed mixtures at 1.5% foaming water content versus the control HMA. As shown in Figure 4-55(a), a LCSN value of 20,000 load cycles was observed for a plant-produced foamed mixture, a laboratory- produced foamed mixture, and laboratory-produced HMA, indicating that no stripping occurred during the test for any mixtures. A significantly lower LCSN value was shown for plant-produced HMA as compared to the other three mixtures, indicating a higher moisture susceptibility in the HWTT. This is consistent with findings from Abbas et al. (2013) that show greater rutting susceptibility of the plant HMA than the laboratory-produced mixtures. A LCST value of 45,627 load cycles was obtained for the plant-produced HMA, while the determination of LCST for the other three mixtures was not available since no stripping occurred within the mixture during the test. Test results in Figure 4-55(b) show that for both plant- produced and laboratory-produced mixtures, a better rutting Figure 4-54. HWTT rut depth results for the Huntsville HMA versus foamed mixtures at 1.5% foaming water content; (a) plant produced and (b) laboratory produced. (a) (b)
73 resistance, as indicated by lower DevpSN values, was exhibited by foamed mixtures at the optimum foaming water content as compared to their HMA counterparts. Thus, better rutting resistance was exhibited by the mixtures produced with 1.5% foaming water content versus their HMA counterparts in the HWTT. Additionally, a better performance in the HWTT was observed for laboratory-produced mixtures than for those produced in the plant. Therefore, HWTT results presented in Figure 4-55 indicate that equivalent or better performance in terms of moisture susceptibility and rutting resistance can be achieved by foamed mixtures than by the control HMA when produced at the optimum foaming water content. Figure 4-55. HWTT results for the Huntsville laboratory-produced HMA versus foamed mixtures at 1.5% foaming water content; (a) LCSN and LCST, and (b) Dî« vp SN. (a) (b)
