Introduction
The Kansas Department of Transportation (KDOT) has been increasingly using Superior Performing Asphalt Pavements (Superpave) mixtures. The moisture sensitivity (also known as “stripping”) which causes loss of bonding between aggregate and binder is currently evaluated by the Kansas Standard Test Method KT-56 [
1]. KT-56 is similar to the American Association of State Highway Transportation Officials AASHTO T 283 procedure [
2] adopted during the Superpave research but has minor modifications in the conditioning procedure. According to the current KDOT specifications for Superpave mixes, this KT-56 test at best takes four days to run and takes two failing tests to shut down the production. This requirement potentially can result in an eight-day production of Superpave mixtures that could be susceptible to stripping. Another criticism of the current test procedure is that the use of anti-stripping agents could improve the low Tensile Strength Ratio (TSR) or make the mixture meet the minimum TSR requirements. Instead of increasing the conditioned strength, the current test procedure lowers the tensile strength of the anti-stripping agent-treated unconditioned specimen. This problem with the AASHTO T283/KDOT KT-56 procedure has been recognized at the national level. The National Cooperative Highway Research Program (NCHRP) Project 9-34- Improved Conditioning Procedure for Predicting the Moisture Susceptibility of Hot Mix Asphalt (HMA) Pavements [
3] looked into the conditioning procedure in the AASHTO T 283. The project investigated the possibility of correlating the moisture sensitivity of paving mixes measured with the Environmental Conditioning System (ECS)/dynamic modulus (from the Simple Performance Tester) combination to (a) the known field performance of the mixes and (b) their moisture sensitivity measured with the Hamburg Wheel Tracking Device (HWTD) test method and ASTM method D 4867 (moisture sensitivity test) [
4]. The study showed that the HWTD identified moisture sensitivity of six Superpave mixtures out of eight tested (75% success rate). Some state agencies have already adopted “wheel tracking” type tests like the HWTD and the Asphalt Pavement Analyzer (APA) to evaluate moisture sensitivity of Superpave mixtures. Notables are Texas, Colorado, and Georgia.
Research at Kansas State University [
5] demonstrated that the HWTD is capable of identifying moisture sensitivity of Superpave mixtures in Kansas. Cross and Voth [
6] conducted APA tests at the University of Kansas to evaluate the effects of sample preconditioning on rut depths and suitability of APA for determining moisture susceptible mixtures. Their test results suggested that AASHTO T 283 preconditioning had little effect upon the rutting results. Cross and Voth [
6] concluded that testing of samples with dry and soak conditioning may be all that is necessary for developing a test method for predicting moisture susceptibility with the APA. However, Cross and Voth [
6] could not establish good correlation between rut depths and the results obtained from other test methods like TSR values, methylene blue values, and sand equivalent. APA tests were able to detect the influence of liquid anti-stripping agents but could not detect the influence of lime. Cross and Voth [
6] indicated that a 50°C testing temperature could result in more definitive results.
Moisture sensitivity is one of the major problems in HMA pavements, potentially leading to premature pavement distresses. Moisture damage in asphalt pavements can occur either by adhesive fracture, i.e., failure at the aggregate-mastic interface or cohesive fracture, i.e., failure within the mastic. Little and Jones [
7] identified six contributing mechanisms to moisture damage: detachment, displacement, spontaneous emulsification, pore pressure–induced damage, hydraulic scour, and environmental effects on the aggregate–asphalt system.
Various test methods to predict moisture sensitivity have been developed. Tests may be carried out on loose samples as well as on compacted specimens. Common methods for predicting moisture sensitivity of HMA mixes include the boiling water test [8], the Texas freeze–thaw pedestal test [9], the static immersion test (AASHTO T-182), the Lottman test [10], the Tunicliff and Root conditioning test [11], the modified Lottman test (AASHTO T 283), the immersion-compression test (AASHTO T-165), the environmental conditioning system (ECS) [12], the asphalt pavement analyzer (APA) test [13], the Hamburg wheel tracking device (HWTD) test [14], and the fatigue testing [15]. These tests suffer from inconsistency and their results may not correlate well with field experiences. Extensive research is still needed for wheel tracking devices such as the HWTD and the APA before they can be considered as rapid tests useful in evaluating moisture susceptibility.
The objective of this research was to use wheel tracking methods to evaluate moisture sensitivity of HMA (Superpave) mixtures. In cooperation with KDOT, research was carried out jointly at the University of Kansas (KU) and Kansas State University (KSU). Six Superpave mixtures were selected for evaluation. HWTD tests were conducted at KSU. APA tests were conducted at KU. The results obtained from these tests were analyzed to evaluate how well they can detect moisture sensitivity.
Laboratory study
This study was part of the joint research efforts between KU and KSU sponsored by KDOT through the Kansas Transportation Research and New-Developments (K-TRAN) program to develop a rapid test method for evaluating moisture sensitivity of HMA samples. APA tests were conducted at KU while HWTD tests were conducted at KSU. To eliminate possible variations of sample preparation, a series of test samples for APA and HWTD tests were prepared at KSU by the same members of graduate and undergraduate students from KU and KSU. These samples were tested parallel using APA and HWTD testers at KU and KSU, respectively.
Test samples
HMA samples were fabricated using Superpave gyratory compactors. The diameter of all the samples was 150 mm. Table 1 provides the dimensions of these cylindrical samples. For APA tests, the height of samples was 75 mm whereas for HWTD tests it was 60 mm. Six samples were made with anti-stripping agents and six without anti-stripping agents for both APA and HWTD tests. Aggregate, binder, and anti-stripping agents were obtained from KDOT. One design mix each for Districts 2, 3 and 5 and three design mixes for District 6 were used to prepare laboratory samples for the joint research. These mixes involve the most commonly used mixtures, SM-12.5A with PG 64-22 binder, in overlay projects, and SM-19A mixtures with PG 64-22 binder for major modification projects, at KDOT. Two more sets of samples using one of the mixes for District 6 were prepared at KSU as well as KU and tested by the APA at KU to evaluate possible effect of sample preparation at different laboratories.
Table 2 presents the overall project information on the HMA mixes used in different districts, the name of county, and the contractor. Mixing and molding temperatures, aggregate type and ratio, binder type, anti-stripping agent type, and the amount of each mix are also provided.
All samples used for the HWTD tests were soaked in water at 50°C. The time needed for the water to reach 50°C was 30 min. Samples were cut and fit in a mold. Rut depth in the cylindrical sample vs. number of cycle was recorded automatically.
In case of the APA tests, samples were subjected to vacuum saturation (20 in Hg) for six minutes before wet tests. Samples were soaked for one hour after the required temperature was reached. All those samples using 50°C in HWTD were tested at the same temperature in the APA. Manual measurements were taken using a dial gauge. These conditions were selected based on the previous study conducted by Cross and Voth [
6].
Test equipment
Superpave gyratory compactor
The Superpave gyratory compactor is a device which can be used to fabricate test specimens by simulating the effect of traffic on an asphalt pavement. The specimens fabricated with the gyratory compactor can be used to determine the volumetric properties (air voids, voids in the mineral aggregate, and voids filled with asphalt) of Superpave mixes. Peterson et al. [
16] showed the Superpave gyratory compactor with proper selected parameters can produce specimens which simulate the mechanical properties of pavement cores in the field. Thus, the gyratory compactor can be used for quality control/quality assurance. The level or amount of compaction is dependent on the environmental conditions and traffic levels expected at a specific job site. Sample height, number of gyrations as well as pressure to be applied can be set in the Superpave gyratory compactor. The sample height of 75 mm for APA tests and the sample height of 60 mm for HWTD tests were fixed in this study.
Hamburg wheel testing device
The HWTD was originally manufactured by Helmut-Wind, Inc. of Hamburg, Germany. Test samples are typically 260 mm wide, 320 mm long, and 40 mm thick and they are compacted at approximately 7 percent air voids using a plate compactor. They can also be prepared using a Gyratory compactor. Two samples are tested simultaneously. The samples are commonly submerged under water at 50°C even though the temperature can vary from 25 to 70°C. A steel wheel, 47 mm wide and loaded under 705 N makes 50 passes over each sample per minute. The maximum velocity of the wheel is 340 mm/sec in the center of the sample. Each sample is loaded for 20,000 passes or until 20 mm of deformation occurs. Approximately 6-1/2 h are required for one test. The HWTD was used in this study to evaluate the moisture sensitivity.
Asphalt pavement analyzer
This APA machine available at KU was manufactured by Pavement Technology Inc., (PTI) Covington, GA. It was 2.03 m long, 0.9 m wide, and 1.78 m high. The total weight of this machine was 1358.4 kg. This machine had retractable legs with wheels to make it portable and anchored while in use. Although the air consumption is low, the minimum pressure of 827 kPa is critical to maintain the adequate hose inflation.
The APA is designed to simulate a rolling wheel condition by rolling three concave metal wheels on three rubber hoses which can provide the pressure ranging from 0 to 827 kPa to simulate the effect of tire pressures. In this study, 0.44 kN loaded wheels on rubber hoses that had air pressures of 690 kPa were used. The assembly holds the samples directly underneath the rubber hoses to allow the samples to be subjected to the wheel tracking action during the test. The tractable tray allows the samples to be pulled out of the machine for manual measurements and sample installation. The cylindrical rutting test molds were used in this study. Cabin temperature as well as water temperature can be set to reach a desired test temperature. In this study, tests were conducted at 50°C. The water submerging system allows the water to cover the test samples in the submerged-in-water tests. A water heater heats the water to a set temperature and a water pump circulates the water from the lower water tank (reservoir) to the upper water tank (where the sample is conditioned). For manual measurements, the upper water tank is lowered and then the sample tray is pulled out for data recording. When the water tray is lowered, the water is drained down to the lower water reservoir. The upper water tank can be raised again to continue the wet test. This system consists of an air inlet regulator controlling the maximum air pressure to the APA, an electronic regulator controlling the hose pressure, three regulators controlling the left, middle, and right wheel cylinders, and the pressure booster regulator doubling the incoming air pressure. All rut measurements were made manually using a steel plate and a dial gauge. The steel plate is fitted in the slot of the mold, and a dial gauge is placed on the top of the plate to take a measurement. It should be noted that there is a unique position for the plate to fit in the cylindrical mold so that the plate is at the same level for every measurement.
Test procedure
Sample preparation
Cylindrical samples were prepared for the APA and the HWTD wheel tests. One set of samples each for Districts 2, 3, and 5 were prepared whereas three sets were prepared for District 6. Each set consisted of 6 cylindrical samples without any anti-stripping agent and 6 cylindrical samples with an anti-stripping agent for the APA tests and equal number of samples for the HWTD tests. The asphalt binder contents used in this study were from 4.7% to 6.75%. The anti-stripping agent contents were from 0.25% to 0.5%.
Mix and compaction temperatures were selected based on the binder grade. Aggregate was weighed and heated in an oven to the desired mix temperature. An asphalt heater was used to heat the binder. Once the desired temperature was reached, aggregate was mixed with binder in an electrical mixer. A hand scoop was also used to mix and to make sure aggregate was mixed properly. When the anti-stripping agent was used, the required quantity of anti-stripping agent was added and mixed along with the binder and aggregate. The mix was then again stored at compaction temperature for two hours for short-term aging.
The Pine Superpave gyratory compactor was used to fabricate the samples. Gyratory molds, the mix pouring funnel, the scoop were all heated to the compaction temperature. After two hours of short-term aging, the required quantity of mix for one sample was poured in the gyratory mold using the sample pouring funnel. The mold was then placed inside the gyratory compactor chamber and the door was closed. The number of gyrations, the sample height, and the required compaction pressure were set on the control panel. In this study, the compaction pressure was set at 600 kPa. Reaction bearings are used to position the mold at an internal angle of 1.16 degrees and the load was applied to the specimen from upper and lower mold plates. The compactor stopped itself when the set height was reached. Once the machine self-parked, the door was opened and the compacted sample was removed from the chamber and extruded using the hydraulic jack on the right side of the compactor. Six cylindrical samples were prepared in one batch.
Sample conditioning and testing
Before running an APA wheel test on wet samples, samples were subjected to 20-in Hg vacuum saturation for six minutes each. Three molds were used in one set of test. Each mold contained two samples. Hence six samples were tested in one set. Water and cabin temperatures were raised and samples were soaked in the APA machine before running the test. All the wet samples were pre-conditioned for at least one hour after the water and chamber temperatures reached the desired temperature. The number of required cycles was set on the APA machine control panel. The air compressor was turned on and the required hose pressure was set on the control panel. The wheel load was applied on the cylindrical samples by switching the green button in the control panel. All rut depths were measured manually and averaged for each set in this study. The relationship of rut depth vs. number of cycles for each test was plotted in a spreadsheet for analysis.
Test results and analysis
Introduction
A summary of APA tests carried out at KU is presented in Table 3. All tests from No. 1 to 12 were conducted using the samples prepared at KSU. The same number of HWTD tests were conducted at KSU. Since these samples were prepared under the same condition, their test results should compare well. The test results for these samples obtained from the HWTD tests as well as the APA tests are presented in this section.
Visual observation of stripping phenomenon
After each test, samples were examined visually to inspect the degree of stripping. Most samples prepared without any anti-stripping agent showed clear or some degree of stripping after the APA tests. The stripping exhibited in forms of debonding of binder and aggregate and washing out of particles as shown in Fig. 1. The samples in Fig. 1 had the same mix design except the use of the anti-stripping agent. In case of samples prepared with an anti-stripping agent, there was no obvious stripping phenomenon. The visual observation of the stripping phenomenon for each mix is provided in the summary of the test results later.
Test results and analysis
The curves from the APA and HWTD tests are presented and discussed in terms of samples for different districts in the following section.
District 2 (2G06015A)
The Superpave HMA mix used in Cloud and Jewel Counties of KDOT District 2 was selected to make laboratory samples. The parameters associated with this mix are provided in Table 2. Figure 2 shows the measured rut depth
vs. the number of cycles from the APA tests. Even though stripping was observed on the sample without any anti-stripping agent, the rut depth was smaller for the mix without any anti-stripping agent than that of the mix with an anti-stripping agent. No definitive stripping inflection point existed for either curve. This result is consistent with what West et al. [
17] found in their study using APA. HWTD test results show that for the sample without any anti-stripping agent, the stripping inflection point occurred around 1500 cycles and the use of the anti-stripping agent increased the number of cycles corresponding to the inflection point and reduced the rut depth. Figure 2 also shows the comparison of rut depths obtained from the APA and the HWTD tests. From the curves, the APA tests could not identify the benefit of using the anti-stripping agent whereas the HWTD tests could identify the benefit. The curves from the HWTD tests show stripping inflection points but those from the APA do not show any inflection point. Visual inspection identified stripping in the sample after the APA tests. The HWTD results matched with visual inspection. In addition, the overall rut depths from the HWTD tests were larger than those from the APA curves, therefore, the HWTD tests were more severe than the APA tests.
District 3 (3G06020A)
The Superpave HMA mix used in Rooks County of KDOT District 3 was used to make the laboratory samples. The parameters associated with this mix are provided in Table 2. Figure 3 presents the measured rut depth vs. the number of cycles for the APA tests. The measured curves for samples with and without an anti-stripping agent virtually overlap. Even though the number of cycles for the sample with the anti-stripping agent was continued up to 20000, no definitive stripping inflection point occurred in either curve. The HWTD test results show that for the sample without any anti-stripping agent, the stripping inflection point occurred around 4,500 cycles. The use of the anti-stripping agent resulted in larger rut depth at the beginning up to 6,500 cycles but showed a long-term benefit after this level. Figure 3 also shows the comparison of rut depths obtained from the APA and the HWTD tests. The curves from the APA tests do not show the benefit of the anti-stripping agent whereas the curves from the HWTD tests show its benefit at a higher number of cycles. The curve from the HWTD test on the sample without the anti-stripping agent shows the stripping inflection point. However, the curves from the APA tests do not show any inflection point for both cases. Visual inspection identified stripping in one of the samples without any anti-stripping agent after APA tests. HWTD results match with visual inspection. The curves from the APA tests, both with and without an anti-stripping agent, overlap with those from the HWTD tests with an anti-stripping agent. Rut depth was larger for the HWTD curve without any anti-stripping agent which shows stripping as well.
District 5 (5G06016A)
The parameters associated with this mix are provided in Table 2. Figure 4 presents the measured rut depth vs. the number of cycles from the APA and HWTD tests. Similar to District 2 samples, the APA tests did not detect any effect of the anti-stripping agent. Even though the visual inspection showed that the sample without any anti-stripping agent had more obvious stripping behavior than that with the anti-stripping agent. The rut depth for the sample with the anti-stripping agent was larger than that without any anti-stripping agent. In addition, there is no obvious stripping inflection point for either curve. The HWTD test results show that the curve for the sample without any anti-stripping agent shows a stripping inflection point while that with the anti-stripping agent does not show a clear inflection point. The use of the anti-stripping agent increased the rut depth at the beginning but reduced the rut depth after about 2100 cycles. Figure 4 also shows the comparison of rut depths obtained from the APA and the HWTD tests for this mix. The curves from the APA tests do not identify the benefit of using the anti-stripping agent in the reduction of rut depth whereas the curves from the HWTD tests identify some benefit at the higher number of cycles. The curve for the sample without any anti-stripping agent from the Hamburg test shows a stripping inflection point but the curves from the APA tests do not show any inflection point. Visual inspection observed stripping in the sample after the APA tests. The HWTD results match with visual inspection. Again, the overall rut depths in the Hamburg tests were larger than those in the APA tests.
District 6 (6G06011A)
The parameters associated with this mix are provided in Table 2. Figure 5 presents the measured rut depth vs. the number of cycles from the APA tests, which shows unusually large rut depth for the sample with the anti-stripping agent. No curves show a stripping inflection point. The anti-stripping agent in this case might have helped soften the binder thus increasing the rut depth. The measured rut depth vs. the number of cycles from the HWTD tests is also shown in Fig. 5. Curves for these tests do not show any stripping inflection point. Similar to the APA tests, the rut depth for the sample with the anti-stripping agent is much larger than that without any anti-stripping agent. The anti-stripping agent in this case might have softened the binder thus increasing the rut depth. Therefore, both APA and HWTD tests do not show any stripping inflection point. Visual inspection after APA tests showed that the samples did not have any stripping. Test results matched with visual inspection results. Both tests show larger rut depths in samples using the anti-stripping agent. This behavior might be due to binder softening.
District 6 (6G06016A)
The Superpave HMA mix used in Meade County of Kansas District 6 was used to make laboratory samples. The parameters associated with this mix are provided in Table 2. Figure 6 presents the rut depth vs. the number of cycles from the APA and HWTD tests for this mix. It is shown that the samples with and without an anti-stripping agent performed equally in the APA tests. No obvious stripping inflection point can be identified in either curve from the APA tests. The measured rut depth vs. the number of cycles from the HWTD tests is also shown in Fig. 6. The curves for the HWTD tests do not show any stripping inflection point. The use of the anti-stripping agent reduced the rut depth. Figure 6 shows the comparison of the results from the APA and the HWTD tests for this mix. Both tests did not show any stripping inflection point. Visual inspection after the APA tests showed that the samples do not have any stripping. Test results match with visual inspection results. The HWTD results show a smaller rut depth in the samples with the anti-stripping agent whereas the APA results show both curves almost overlapping each other. The curves for the samples with the anti-stripping agent tested in the HWTD almost overlap those obtained from the APA tests. Hence rut depths obtained from the HWTD and the APA match reasonably well.
District 6 (6G07002A)
The Superpave HMA mix used in Meade County of Kansas District 6 was used to make laboratory samples. The parameters associated with this mix are provided in Table 2. One sample with an anti-stripping agent and one without any anti-stripping agent were prepared at KSU for the APA tests. The same mix was used to prepare an equal number of samples at KU to investigate any possible effect of sample preparation at two different laboratories.
Figure 7 presents the measured rut depth vs. the number of cycles from the APA tests on the samples prepared at KSU and KU. It is shown that the samples without any anti-stripping agent had slightly larger rut depths than those with the anti-stripping agent for the samples prepared at KSU. However, the use of the anti-stripping agent reduced the rut depth more significantly than that without any anti-stripping for the samples prepared at KU. To investigate the stripping inflection point, the APA tests were run up to 20,000 cycles, the number of repetitions commonly used for the HWTD tests. However, no stripping inflection point was observed even up to 20,000 cycles. The curves for the samples prepared at KU show slight stripping inflection points for both samples with and without the anti-stripping agent. It was observed visually that the samples without the anti-stripping agent showed more stripping behavior than those with the anti-stripping agent. Therefore, the samples prepared at two different laboratories had different results, especially for the sample without any anti-stripping agent.
Figure 8 presents the measured rut depth vs. the number of cycles for all the samples prepared at KSU and tested by the APA and the HWTD. The APA test results show almost no difference in the rut depths with the number of cycles and no inflection point for both samples. However, the HWTD test results show that the samples without any anti-stripping agent had a stripping inflection point at approximately 6,500 cycles. Samples with the anti-stripping agent did not have any stripping inflection point. Similar to other results, the use of the anti-stripping agent increased the rut depth for the short-term but reduced the rut depth for the long-term, especially after the inflection point of the sample without the anti-stripping agent. Visual inspection of these samples showed stripping after the APA tests. Rut depths from both APA and HWTD match reasonably well. The APA did not detect any stripping inflection point. However, the HWTD test results show a more obvious stripping inflection point for the samples without any anti-stripping agent. The use of the anti-stripping agent improved the performance of the sample against stripping.
Summary of results
Table 4 summarizes the test information and results obtained using the APA and HWTD tests. The table tabulates the district, project design number, mix type, laboratory designation number, stripping inflection point, benefit of anti-stripping agent, and visual inspection of samples after APA and HWDT tests. Test results show that APA tests could not result in any stripping inflection point no matter whether an anti-stripping was used or not. However, after each APA test, stripping might be observed visually on the sample. HWDT tests resulted in stripping inflection points for most of the samples without any anti-stripping agent but did not show any stripping inflection point for most of the samples with an anti-stripping agent. Therefore, the HWTD test method is more effective in detecting potential moisture sensitivity of an HMA sample than the APA test method.
Conclusions
Totally 26 sets of hot-mix asphalt (HMA) samples were prepared at the Kansas State University and two more sets of samples were prepared at the University of Kansas. These samples were tested by the asphalt pavement analyzer (APA) and the Hamburg Wheel Tracking Device (HWTD) to rapidly evaluate moisture sensitivity of HMA samples. The following conclusions can be made:
1) Visual observation detected stripping in four out of six mixes without any anti-stripping agent in the APA tests. Loss of bonding between aggregate and binder occurred in these samples. Sand particles were washed out of binder in a few stripped samples.
2) None of the APA tests conducted in this study showed a clear stripping inflection point. However, the HWTD tests did show such a stripping inflection point in all four mixes without any anti-stripping agent. The HWTD tests did not show any stripping inflection point for two mixes without any anti-stripping agent, which did not show stripping during visual inspection.
3) The use of an anti-stripping agent showed the benefit in reducing rut depth in some cases of the APA tests but not in all. However, most of the HWTD test results showed this benefit in the later stage of the tests. In the earlier stage, the samples with an anti-stripping agent had larger rut depths than those without any anti-stripping agent. In one mix (6G06011A), the use of the anti-stripping agent increased rut depth in both APA and HWTD tests.
4) Both APA and HWTD tests take around 4 to 6 h to complete. These methods are faster than conventional test methods used to determine moisture sensitivity. Both APA and HWTD tests can show stripping behavior visually. In addition, Hamburg tests can detect stripping behavior in HMA mixes based on the rut depth vs. the number of cycle curve. Thus the HWTD tests are more effective as a rapid test method in case of determining moisture sensitivity.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(1301KB)
