Validation of a Performance-Based Mix Design Method for Porous Friction Courses (2021)

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.

15 C H A P T E R 4 Results and Analysis This chapter includes results of the implementation effort in which the research team worked with GDOT, SCDOT and ALDOT in three construction projects. Raw materials were sampled for mix design verification, and samples of plant-produced mixtures were taken from each project for testing in the NCAT laboratory. Results of the implementation effort are presented in the following sections. GDOT Project The PFC resurfacing project in Georgia is on SR-520/US-82 in Worth County from MP 9.4 through MP 11.02. As agreed by GDOT and the contractor for the project, the project team sampled both raw materials, including aggregates, fibers and asphalt binder, and plant-produced mixtures for testing in this implementation study. The raw materials were used to prepare laboratory-mixed, laboratory-compacted (LMLC) specimens based on the PFC job mix formula (JMF) provided by GDOT while the plant mix samples were reheated to prepare plant-mixed, laboratory-compacted specimens. Table 4-1 summarizes the test results for both LMLC and PMLC specimens as well as the respective thresholds proposed in the NCHRP 01-55 performance-based PFC mix design procedure. Their gradations are compared in Table 4- 2. The following observations can be drawn from the test results.  The results of LMLC specimens prepared based on the GDOT mix design met all of the proposed criteria, as shown in Table 4-1.  The Cantabro loss of the PMLC specimens did not meet the proposed threshold, but the other parameters of the PMLC specimens met the proposed criteria. As shown in Table 4-1, the PMLC mixture had an average Cantabro loss of 21.5 percent versus the proposed 20 percent maximum from NCHRP Report 877. The proposed Cantabro loss threshold is an important criterion as it separates the good performing from the poor performing mixtures evaluated in NCHRP Project 01-55.  After further laboratory investigation, it was found that the PMLC sample had an asphalt content of 5.4 percent as compared to 6.2 percent shown in the approved PFC mix design (Table 4-1). This difference may have resulted in the higher Cantabro loss of the PMLC mixture. The binder content of the PMLC mixture was verified by two different methods (i.e., ignition oven and solvent extraction). The binder content results obtained from the two methods were similar (i.e., 5.43 percent for ignition oven and 5.41 percent for solvent extraction).  The PFC mix design gradation met the GDOT specification requirements, except for the fines (passing No. 200), as shown in Table 4-2. The design gradation may have been proportioned with a lower fines content to account for aggregate breakdown during production. Thus, the PMLC gradation met all of the GDOT gradation requirements and was within allowable tolerances from the design gradation. The gradation of LMLC specimens was batched in the laboratory following the GDOT-approved JMF.

16 Table 4-1. Test Results of LMLC and PMLC Specimens for Georgia Project. Test Conducted LMLC PMLC NCHRP 01-55 Criteria Binder Content (%) 6.2 (JMF) 5.4 (Ignition)1 None2 Draindown (%) 0.00 0.00 Max 0.3 Air Voids (Corelok) (%) 18.8 19.5 15 – 20 Air Voids (Dimensional) (%) 21.4 21.3 17 – 22 Gmm 2.470 2.498 None Cantabro Loss (%) 11.5 21.5 Max 20 Conditioned ITS (psi) 53.4 76.8 Min 50 Unconditioned ITS (psi) 58.0 90.6 None TSR 0.92 0.85 Min 0.70 Note: 1The average of two solvent-extraction binder contents was 5.41 percent. 2While there is no threshold proposed in NCHRP Project 01-55 for binder content, SHAs typically require a minimum of 6 percent binder content for their PFC mixtures. Table 4-2. Aggregate Gradations of LMLC and PMLC Specimens for Georgia Project. Sieve Size Percent Passing GDOT Specification 01-55 Proposed 12.5-mm Gradation Std. Metric JMF LMLC (Batch) PMLC (Ignition) 1" 25.0 mm 100 100 100 100 100 3/4" 19.0 mm 100 100 100 100 100 1/2" 12.5 mm 97 95 95 85 – 100 80 – 100 3/8" 9.5 mm 71 70 68 55 – 75 35 – 60 #4 4.75 mm 17 14 18 15 – 28 10 – 25 #8 2.36 mm 6 7 6 5 – 15 5 – 10 #16 1.18 mm 4 5 5 #30 0.600 mm 3 4 4 #50 0.300 mm 3 3 4 #100 0.150 mm 2 2 4 #200 0.075 mm 2.1 1.9 3.1 3 – 5 2 – 8 SCDOT Project The PFC construction project in South Carolina is on I-26 near Spartanburg from MP 23 through MP 44. Both raw materials and plant-produced mixtures were sampled. LMLC and PMLC specimen test results are summarized in Table 4-3 and compared with the respective thresholds proposed in the NCHRP 01-55 performance-based PFC mix design procedure. Table 4-4 compares the gradations of LMLC and PMLC specimens with those of the JMF and SCDOT specifications. Based on the test results, the following observations can be offered.

17  The LMLC specimens met all the proposed requirements, as shown in Table 4-3. The PMLC specimens barely failed the minimum air void threshold, but all of the other parameters met the proposed criteria. The plant-produced mixture had an average air void content of 14.9 percent by Corelok and 17.0 percent by dimensional analysis. Since the average air void content measured by the Corelok method marginally missed the proposed threshold by only 0.1 percent, the plant-produced mixture can be considered to practically meet all the proposed criteria. The LMLC mixture showed a lower Cantabro loss and a higher TSR than the PMLC mixture, which may be attributed to the 0.3 percent difference in binder content between the two mixtures.  The LMLC gradation was slightly different from the JMF but was close to the PMLC gradation. The gradations for JMF and LMLC mixture passed the SCDOT gradation requriments, as shown in Table 4- 4. The gradation of the PMLC mixture also passed the SCDOT gradation requriments, except for the fines content that was slightly higher than the SCDOT upper limit. Table 4-3. Test Results of LMLC and PMLC Specimens for South Carolina Project. Test Conducted LMLC PMLC NCHRP 01-55 Criteria Binder Content (%) 6.0 (JMF) 5.7 (Ignition) None1 Draindown 0.07 0.01 Max 0.3 Air Voids (Corelok) (%) 15.0 14.9 15 – 20 Air Voids (Dimensional) (%) 17.3 17.0 17 – 22 Gmm 2.448 2.448 None Cantabro Loss (%) 9.6 14.3 Max 20 Conditioned ITS (psi) 80.3 73.9 Min 50 Unconditioned ITS (psi) 92.3 95.9 None TSR 0.87 0.77 Min 0.70 Note: 1While there is no threshold proposed in NCHRP Project 01-55 for binder content, SHAs typically require a minimum of 6 percent binder content for their PFC mixtures. Table 4-4. Aggregate Gradations of LMLC and PMLC Specimens for South Carolina Project. Sieve Size Percent Passing SCDOT Specification 01-55 Proposed 12.5-mm Gradation Std. Metric JMF LMLC (Batch) PMLC (Ignition) 1" 25.0 mm 100 100 100 100 100 3/4" 19.0 mm 100 100 100 100 100 1/2" 12.5 mm 99 97 98 92 – 100 80 – 100 3/8" 9.5 mm 69 64 69 55 – 75 35 – 60 #4 4.75 mm 19 25 25 15 – 30 10 – 25 #8 2.36 mm 8 13 16 5 – 15 5 – 10 #30 0.600 mm 5 8 10 #100 0.150 mm 3 5 6 #200 0.075 mm 1.6 3.3 4.4 0 – 4 2 – 8

18 ALDOT Project The ALDOT PFC resurfacing project is on I-65 from MP 359 through MP 366 in Limestone County, Alabama. The existing PFC surface of the pavement section was placed in 2016, but it raveled out within two years. A portion of the pavement section was replaced in 2019. Due to the raveling issues observed in the pavement section paved in 2016, changes to the PFC mix design were made by the contractor and ALDOT for the PFC mixture that was placed in 2019. Tables 4-5, 4-6 and 4-7 compare the aggregate materials, gradations and basic properties for the PFC mix designs used in the 2016 and 2019 resurfacing operations. The noticeable changes include (a) no RAP in the 2019 mix design compared to 10 percent RAP in the 2016 mix design and (b) increasing the total asphalt content from 6.0 to 6.6 percent. These changes increased the virgin SBS-modified PG 76-22 binder content from 5.4 to 6.6 percent (i.e., a 1.2 percent increase). Table 4-5. Aggregates Used in Mix Designs for the 2016 and 2019 Paving Operations. Aggregate Cold Feed (%) 2016 Paving 2019 Paving 1/2" Slag 25 8 3/8" Slag 0 11 #7 Sandstone 59 50 #8 Sandstone 5 30 Baghouse Fines 1 1 RAP 10 0 Fiber 0.3 0.3 Table 4-6. Design Gradations for Mix Designs for the 2016 and 2019 Paving Operations. Sieve Size Percent Passing ALDOT Specification 01-55 Proposed 12.5-mm Gradation Std. Metric 2016 Paving 2019 Paving 1" 25.0 mm 100 100 100 3/4" 19.0 mm 100 100 100 100 1/2" 12.5 mm 90 85 85 – 100 80 – 100 3/8" 9.5 mm 65 62 55 – 65 35 – 60 #4 4.75 mm 19 20 10 – 25 10 – 25 #8 2.36 mm 10 10 5 – 10 5 – 10 #200 0.075 mm 2.7 2.0 2 – 4 2 – 8

19 Table 4-7. Mix Design Properties for the 2016 and 2019 Paving Operations. Mix Properties 2016 Paving 2019 Paving ALDOT Specification Total AC (%) 6.0 6.6 6.0 – 9.0 Virgin PG 76-22/SBS 5.4 6.6 RAP AC 0.6 0.0 No. of Gyrations 50 50 50 %Gmm 86.6 81.3 Mix Temperature 345oF 345oF Delivery Temperature 300oF 300oF A 1.8-mile portion of the pavement section that was paved in 2016 remained in place (i.e, not replaced in 2019), but it was scheduled to be replaced in late July 2020. Even with the changes made to the PFC mix design in 2019, ALDOT engineers were still concerned with the durability of the PFC mixture (Figure 1) to be placed in 2020. Thus, to help ALDOT evaluate the PFC mix design that was updated in 2019, raw materials were sampled from the asphalt plant (as no plant mix was available) and tested in the NCAT laboratory following the NCHRP Project 01-55 performance-based mix design procedure. Figure 4-1. Replacement PFC Mix (After Second Paving in 2019). During a conference call with ALDOT engineers, the research team learned that the plant mixes produced for this project required an extended silo storage and hauling time—about 2 to 3 hours—due to some night paving constraints. In addition, the target mixing temperature was 345oF, and the target delivery temperature was 300  20oF. Thus, the 2019 PFC mix design was tested with two short-term aging (STA) periods: (a) mixed at 345oF and aged for 2 hours at 325oC, which is typically followed during mix design, and (b) mixed at 345oF and aged for 4 hours at 325oC to simulate the extended silo storage and hauling time observed in the paving project. The results are summarized in Table 4-8 and compared with the respective crieteria proposed in NCHRP Project 01-55.

20 Table 4-8. Test Results of LMLC Specimens for Alabama Project. Test LMLC, STA (2 hrs @325oF) LMLC, STA (4 hrs @325oF) NCHRP 01-55 Criteria Draindown (%) 0 n/a Max 0.3 Air Voids (Corelok) (%) 17.4 17.7 15 – 20 Air Voids (Dimensional) (%) 19.9 20.2 17 – 22 Gmm 2.511 2.490 n/a Cantabro Loss (%) 17.2 26.5 Max 20 Conditioned ITS (psi) 69.2 80.7 Min 50 Unconditioned ITS (psi) 84.6 87.3 n/a TSR 0.82 0.92 Min 0.70 All the test results shown in Table 4-8 passed the criteria proposed in NCHRP Project 01-55 except for the Cantabro loss for the test specimens that were short-term aged for 4 hours at 325oF. The high Cantabro loss (above 20 percent) can cause a premature raveling issue in the PFC mixture and could be attributed to the extended silo storage and hauling time at the elevated temperature in combination with the absorptive slag and sandstone aggregate materials used in the mixture. To reduce the Cantabro loss, the following options were considered: 1. Reducing the mixing and delivery temperatures by 20-25oF (i.e., from 345/325 to 325/300oF) while keeping other variables the same. Also, a WMA additive was used, which can be important in case of unforeseen delays during hauling. 2. Increasing the total binder content by 0.5 percent (i.e., from 6.6 to 7.1 percent) while keeping other variables the same. 3. Option 2 plus reducing the mixing and delivery temperatures by 25oF. 4. Option 3 plus using a WMA additive. Table 4-9 summarizes the Cantabro test results for the four options discussed above; the following observations can be offered based on the results:  For the 2019 mix design, the extended silo storage and hauling time (i.e., additional 2 hours) has a significant effect on the Cantabro test results (i.e., increasing from 17.2 to 26.5 percent), which can affect the long-term durability of the mix in the field.  For Option 1, reducing the mixing and delivery temperatures by 25oF (with a WMA additive) can lower the Cantabro loss from 26.5 to 18.3 percent, which is below the threshold of 20 percent proposed in NCHRP Project 01-55.  For Option 2, increasing the total binder content by 0.5 percent alone does not improve the Cantabro test results.  For Option 3, a lower Cantabro loss of 17 percent is largely attained by reducing the mixing and delivery temperatures by 25oF and is less affected by the higher binder content.  For Option 4, adding a WMA additive to those considered in Option 3 can further reduce the Cantabro loss.

21 Table 4-9. Evaluation Results for the Four Potential Options to Reduce Cantabro Loss. Option AC (%) WMA (%) Mixing Temp (F) Aging Temp (F) Aging (hr) Air Void (%) Cantabro Loss (%) 2019 Mix Design 6.6 0 345 325 2 17.6 17.2 6.6 0 345 325 4 17.3 26.5 Option 1 6.6 0.5 325 300 4 18.3 18.3 Option 2 7.1 0 345 325 4 16.1 27.8 Option 3 7.1 0 325 300 4 17.7 17.0 Option 4 7.1 0.5 325 300 4 17.1 14.6 Based on the results obtained from the laboratory experiment, the following suggestions were presented to ALDOT:  The production temperature should be reduced to 325oF to lower binder absorption and avoid possible damage to binder and molecular polymer chain.  The delivery temperature can be kept at 300  20oF.  The contractor can start with load No. 3 at the transverse joint instead of the first load to account for any variability in plant start-up temperatures. The remaining 1.8-mile section was repaved in July 2020 using the same mix design approved in 2019. However, the target production temperature was reduced to 325oF, and the mix was kept in silo for less than two hours. (During initial plant start-up there was a spike in production temperature as high as 390°F. The first five loads of mixture produced were rejected at the plant due to excessive temperature.) The delivery temperature at the paving site was around 300oF. A sample of the plant mix was taken during production for laboratory testing. Table 4-10 compares the laboratory test results for the PMLC specimens and LMLC specimens that were conditioned for 2 hours at 325oF to simulate the production condition. The draindown and indirect tensile strength test results of the PMLC specimens met the proposed thresholds. The binder content for the PMLC mixture was 0.3 percent lower than the LMLC mixture, and its air voids were lower than the proposed thresholds, which resulted in a lower Cantabro loss of 5.7 percent. The lower air voids of the plant mix were likely due to the finer gradation in the plant mix, as shown in Table 4-11. The gradation of the plant- produced mixture was outside of the ALDOT specification bands for 3/8” and #4 sieves. Table 4-10. Test Results of Lab and Plant PFC Mixtures for the Alabama Project. Test LMLC Mixture (2 hrs @325oF) PMLC Mixture (Reheated) NCHRP 01-55 Criteria Binder Content (%) 6.6 (JMF) 6.3 (Ignition) Draindown (%) 0.0 0.0 Max 0.3 Air Voids (Corelok) (%) 17.4 14.3 15 – 20 Air Voids (Dimensional) (%) 19.9 16.8 17 – 22 Gmm 2.511 2.544 n/a Cantabro Loss (%) 17.2 5.7 Max 20 Conditioned ITS (psi) 69.2 64.6 Min 50 Unconditioned ITS (psi) 84.6 85.7 n/a TSR 0.82 0.75 Min 0.70

22 Table 4-11. Design and Plant Mix Gradations for the Alabama Project. Sieve Size Percent Passing Std. Metric Lab Mix (JMF) Plant Mix (Ignition) ALDOT Specification 1" 25.0 mm 100.0 100.0 3/4" 19.0 mm 100.0 100.0 100 1/2" 12.5 mm 85.0 87.7 85 – 100 3/8" 9.5 mm 62.0 65.2 55 – 65 #4 4.75 mm 20.0 23.2 10 – 25 #8 2.36 mm 10.0 12.5 5 – 10 #200 0.075 mm 2.00 2.22 2 – 4