The warm mix asphalt was fabricated with different moisture contents (0%, 1%, 2%, and 3%) of limestone aggregates using the Superpave gyratory compactor. The moisture susceptibility of asphalt mixtures with an organic wax additive RH was studied. The samples were compacted and tested using the modified Lottman test AASHTO T283, and the X-ray computed tomography technology was used to capture the internal structure images before and after the freeze-thaw cycles. The test results show that the air voids were distributed in the size range of 0–5 mm3 and 5–10 mm3. The number of air voids decreased with the increase of air void size and increased after freeze-thaw cycles. The air void content can be influenced by the residual moisture in aggregates. The higher the moisture content of aggregates is, the larger the air void content is. So, the air void content is likely to be sensitive to moisture damage. The increase ratio of the air void and moisture content of aggregates had good correlation with the indirect tensile strength and tensile strength ratio of the samples. The indirect tensile strength and tensile strength ratio of the samples decreased linearly, and the samples were sensitive to the moisture damage with the increases of increase ratio of the air void/moisture content in aggregates.
The technology of warm mix asphalt (WMA) features energy conservation and environmental protection, and WMA has been used widely in the industry. Different advantages are identified for WMA compared to hot mix asphalt (HMA), including (1) reduction of fuel consumption during the mixture fabrication; (2) reduction of pollutant emissions (like fumes); (3) improved compaction; (4) the possibility of longer paving seasons; and (5) lower short-term aging thanks to lower production temperatures [
In most cases, the moisture susceptibility of WMA mixtures with different types of additives was studied by conventional test methods used for HMA mixtures. Hesami et al. used hydrated lime to improve the moisture susceptibility of WMA with Sasobit/Aspha-min and limestone/granite. The hydrated lime improved adhesion between asphalt binders and aggregates and reduced the probability of moisture damage in WMA mixtures [
Based on the X-ray computed tomography (CT) technology, the internal structures of asphalt mixtures were analyzed [
The Particle Flow Code (PFC) was used to study the internal structure of asphalt mixtures through X-ray scanning. The aggregate and mastic were simulated, respectively, as linearly elastic and viscoelastic bodies [
Most of the previous studies focused on characterizing the internal structure and developing the finite element 3D numerical model for an asphalt mixture with the X-ray CT technology, but the correlation between internal structure and moisture sensitivity of asphalt mixtures, especially for WMA, is still undeveloped. The motivations of this research are to identify the moisture susceptibility of WMA with an organic wax additive based on X-ray CT technology and find the correlation between the internal structure and moisture sensitivity of WMA.
The objective of this research is to explain the potential for moisture damage from the mesoscale perspective using X-ray CT technology in WMA with organic wax additives. The following tasks were performed in this study in order to achieve the objective: Condition the moist aggregates with different moisture contents (1.0%, 2.0%, and 3.0%). Design WMA mixtures with the dried/moist aggregates according to the Marshall Mix design procedures. The modified Lottman test (AASHTO T283) is used for the WMA samples. The X-ray CT technology is used to analyze the internal structures of the WMA samples and calculate the changes of bulk air void distribution of the WMA samples through freeze-thaw cycles. Establish the relationship between the changes of air void distribution and moisture contents in aggregates. Analyze the correlation between the increase ratios of air voids or moisture content of aggregates and TSR and ITS of WMA. Demonstrate the mesoscale behaviors of moisture susceptibility of WMA.
An SK-90 binder with penetration grade 80/100 used in the WMA was obtained from South Korea. Limestone as a source of mineral aggregate used in the WMA was supplied by Beijing Municipal Road and Bridge Building Material Group Co., Ltd. RH as an organic wax additive is a kind of polyethylene wax and developed by Highway Science Research Institute of Ministry of Communication, which is soluble in the asphalt at the temperatures of above 100°C. RH (3% by mass of SK-90) and SK-90 were blended at 120°C and stirred 5–10 min manually. Physical properties of SK-90, RH, limestone aggregate, and chemical components are listed in Tables
Physical properties of SK-90.
Binder | 25°C penetration (0.1 mm) | 15°C ductility (cm) | Softening point (°C) | 60°C dynamic viscosity (Pa·s) | After RTFOT | ||
---|---|---|---|---|---|---|---|
Mass loss (%) | Residual penetration ratio (%) | 10°C residual ductility (cm) | |||||
SK-90 | 80 | >100 | 53.1 | 140.8 | −0.113 | 61.6 | 10.1 |
Spec. | 80–100 | >100 | >42 | >140 | ≤±0.8 | >50 | >6 |
Physical properties of RH.
Items | Density (g/cm3) | Melting point (°C) | 25°C penetration (0.1 mm) |
---|---|---|---|
Test value | 0.914 | 108 | 0 |
Spec. | — | 98–110 | ≤1 |
Physical properties of coarse limestone.
Items | 10–15 mm | 5–10 mm | Spec. |
---|---|---|---|
Apparent specific gravity | 2.759 | 2.772 | >2.6 |
Bulk volume relative specific gravity | 2.707 | 2.591 | — |
Crushing value (%) | 10.1 | 12.1. | ≤26 |
L.A. abrasion (%) | 11.2 | 10.3 | ≤28 |
Soundness (%) | 2.1 | 2.3 | ≤12 |
Soft stone content (%) | 0.1 | 0.1 | ≤3 |
Physical properties of fine limestone.
Items | 0–5 mm | Spec. |
---|---|---|
Apparent specific gravity | 2.793 | >2.5 |
Clay content (%) | 0.1 | ≤3 |
Sand equivalent (%) | 73 | >60 |
Soundness (%) | 3.1 | ≤12 |
Physical properties of fillers.
Items | Filler | Spec. |
---|---|---|
Apparent specific gravity | 2.793 | >2.5 |
Water content (%) | 0.1 | ≤1 |
Hydrophilic coefficient (%) | 0.03 | ≤1 |
Plasticity index (%) | 1.1 | ≤4 |
Chemical component of limestone aggregates.
Components | CaCO3 | SiO2 | MgCO3 | Fe2O3 | CaO2 | K2O |
---|---|---|---|---|---|---|
Content (%) | 75–77 | 6–8 | 11–13 | 0.2–0.4 | 0.5–1 | 0.4–0.6 |
The asphalt mixtures with RH, an organic wax additive, were formed. The target air void of samples is 4 ± 1.0%. The OAC (optimal asphalt content) of the RH-WMA (WMA with RH additive) is determined in accordance to the Marshall Mix design procedures and American Society for Testing and Materials (ASTM) D1559 (2006), and it is 4.8%. The gradation and performance of the RH-WMA are shown in Tables
Gradation of the RH-WMA.
Sieve size (mm) | 16 | 13.2 | 9.5 | 4.75 | 2.36 | 1.18 | 0.6 | 0.3 | 0.15 | 0.075 |
---|---|---|---|---|---|---|---|---|---|---|
Lower-upper limits (%) | 100 | 90–100 | 68–85 | 38–68 | 24–50 | 15–38 | 10–28 | 7–20 | 5–15 | 4–8 |
Passing ratio (%) | 100 | 96.3 | 76.8 | 49.2 | 34.6 | 24.7 | 19.2 | 12.6 | 11.0 | 5.7 |
Performances of the RH-WMA.
Items | Dynamic stability (times/mm) | Failure strain ( |
TSR (%) |
---|---|---|---|
Test results | 1033.46 | 2194 | 78.39 |
Spec. | ≥800 | ≥2000 | ≥75% |
Based on the literature reviews [ The temperature (110°C) of aggregates was determined according to the viscosity-temperature curve of the control mix binder; the viscosity-temperature curve of the control mix binder is seen in Figure The aggregates were put in the oven (110°C) more than 4 h. The dried aggregates were poured into the bucket mixer (110°C) for keeping heat uniformity more than 30 min. The extra distilled water (1.0%, 2.0%, and 3.0% by mass of aggregates) was added to the bucket mixer and mixed with aggregates (the 90 s) to condition the different moist aggregates.
Viscosity-temperature curve of the control mix binder.
The dried and moist aggregates with 1.0%, 2.0%, and 3.0% were conditioned in the laboratory. RH-WMA samples with dried/moist aggregates were compacted using the Superpave gyratory compactor (SGC). The mixing and compaction temperatures of control samples are 130°C and 120°C, and the mixing and compaction temperatures of RH-WMA samples (moist aggregates with 1.0%, 2.0%, and 3.0%) are 110°C and 100°C. The test samples were cylinders with 64 ± 0.5 mm in height and 100 ± 0.5 mm in diameter. A total of 4 sets of samples were fabricated. Each set contained at least 16 samples, and 64 samples totally were prepared. The bulk air void content of the samples was measured and found to be 4 ± 1.0%. The residual moisture contents of moist aggregates are 0.35%, 0.89, and 1.63%, respectively, which are corresponding to initial moisture contents of 1%, 2%, and 3%. All the RH-WMA samples were kept at room temperature for 24 h before testing.
The test CT was industrial X-ray CT provided by NIKON Corporation, Japan, which essentially consisted of the following components: X-ray source, a radiation detector and collimator, image acquisition system, scanning system, data acquisition and processing system, auxiliary system, and so forth, as shown in Figure
The schematic diagram of the industrial computed tomography (CT) system.
Specifications and scanning parameters of the industrial CT.
Parameters | Energy | System spatial resolution | Maximum power (small/standard) | Focal size (small/standard) | Pixel size | Scan time |
---|---|---|---|---|---|---|
Spec. | 600 KeV | 3.0 lp/mm | 700 W/1500 W | 0.4 mm/1.0 mm | 1024 × 1024 | ≤25 min |
The linear attenuation coefficient of a material is not constant and varies as a function of its density. The attenuation of X-ray penetrates the materials with different penetrating energies. The tomography images of materials without overlapping are obtained by reconstruction algorithms or the distributions of linear attenuation coefficients of X-ray on cross-sections of the materials. The CT image is a map of the spatial distribution of the linear attenuation coefficient, where bright regions correspond to high values of the coefficient, vice versa. The differences in densities of two-dimensional slices were used for identification.
The moisture susceptibility of the RH-WMA with dried/moist aggregates was carried out according to the modified Lottman test method (AASHTO T283). For each type of RH-WMA, two sets of unconditioned and conditioned samples were separated. The dry indirect tensile strength (dry ITS) and wet indirect tensile strength (wet ITS) were measured, and the tensile strength ratio (TSR) was also calculated.
In accordance with AASHTO T269 and AASHTO T 166, the maximum theoretical specific gravity and bulk specific gravity of the RH-WMA with dried/moist aggregates were measured, and the air void percentage of the RH-WMA was calculated.
The industrial CT was used to scan the RH-WMA with dried/moist aggregates through freeze-thaw cycles with a 0.1 mm vertical gap, and their grayscale images were captured. The digital image processing and visual 3D reconstruction technology were used to observe and determine the change in content, size, and distribution of air voids, and the increase ratio of air voids was also calculated.
An imaging processing technique is a process of converting an image into a digital form and applying various mathematical procedures to extract significant information from the image. VG Studio MAX 2.2 was used for batch processing the images of the 64 RH-WMA samples obtained from industrial CT, and the 2D images of the RH-WMA samples thereby were imported into the VG Studio MAX 2.2 software for reconstructing the 3D visual model. As the properties of aggregate, mastic, and air voids are three different phases in asphalt mixture, they can be separated, and a suitable gray intensity threshold value is determined to separate air voids from aggregate and mastic. After 3D visual model reconstruction, the air voids of the RH-WMA might be seen as a defect, and the defect analysis function of the VG Studio MAX 2.2 was selected for recognizing and analyzing air voids. Thus, the air void distribution of the samples was observed, and parameters, such as air void content, size, distribution, and connectivity, were calculated accurately.
It is complicated and important to choose a suitable gray intensity threshold in image processing. The quality of images obtained through industrial CT related to aggregate characteristics. OTSU, Gaussian mixture model (GMM), and fuzzy C-means (FCM) methods were considered and compared in order to accurately determine the gray intensity threshold value and identify air voids of RH-WMA samples [
The 2D and 3D air void segmentation images of the RH-WMA are shown in Figure
2D and 3D air void segmentation images of the RH-WMA.
According to the abovementioned method for obtaining the air voids of the RH-WMA samples, VG Studio MAX 2.2 was used to calculate the air void sizes and air void contents of the RH-WMA samples with dried/moist aggregates before and after freeze-thaw cycle, respectively, as presented in Table
Air void contents and distributions of the RH-WMA samples through freeze-thaw cycles.
Moisture content (%) | 0–5 (mm3) | 5–10 (mm3) | 10–50 (mm3) | 50–100 (mm3) | 100–300 (mm3) | 300–500 (mm3) | >500 (mm3) | Note |
---|---|---|---|---|---|---|---|---|
0 | 2346 | 168 | 21 | 9 | 10 | 6 | 2 | Before freeze-thaw cycle |
1 | 2154 | 201 | 17 | 12 | 9 | 4 | 1 | |
2 | 2187 | 147 | 19 | 10 | 12 | 9 | 3 | |
3 | 2301 | 216 | 28 | 13 | 8 | 5 | 1 | |
|
||||||||
0 | 2024 | 298 | 25 | 11 | 10 | 6 | 2 | After freeze-thaw cycle |
1 | 1854 | 366 | 32 | 15 | 11 | 5 | 2 | |
2 | 1911 | 280 | 41 | 14 | 14 | 10 | 3 | |
3 | 2004 | 441 | 46 | 17 | 12 | 8 | 3 |
Air void contents and their sizes of the RH-WMA through freeze-thaw cycles.
From Figure
Compared with the air void contents of the mixtures in the same air void size through freeze-thaw cycles, it is noticed that the air void content of the samples increased after freeze-thaw cycles except for the size of 0–5 mm3 with the same moisture content of aggregates. The air void content in a range of 5–10 mm3 after freeze-thaw cycles increased by 78%. It is worthy to note 5–10 mm3 was an important size on the air void distribution. In general, the air void contents of the RH-WMA samples after freeze-thaw cycles increased by approximately 46.2%. This is caused by the state change of residual moisture in aggregate, such as melting and freezing, through freeze-thaw cycles.
The 5–10 mm3 size significantly impacted the increment of the air void. The relationship between the moisture content of aggregates and 5–10 mm3 air void contents of the RH-WMA samples was studied, and the results are exhibited in Figure
Moisture content in aggregates over an increment of air void contents in 5–10 mm3 size.
Figure
According to the AASHTO T269 and AASHTO T166, the air voids of the RH-WMA samples were measured in the laboratory and denoted as MAV. The air voids of the mixtures were calculated by the images with X-ray CT technology and named as CTAV. In Table
The increase ratio of air voids of RH-WMA samples through freeze-thaw cycles.
Moisture content (%) | Before freeze-thaw cycles (%) | After freeze-thaw cycles (%) | Increase ratio of air voids (%) | Relative error (%) | |||
---|---|---|---|---|---|---|---|
MAV | CTAV | MAV | CTAV | MAV | CTAV | ||
0 | 4.01 | 2.93 | 4.53 | 3.33 | 12.63 | 13.94 | 9.40 |
1 | 3.16 | 2.35 | 3.64 | 2.75 | 15.22 | 16.92 | 10.05 |
2 | 3.59 | 2.57 | 4.22 | 3.03 | 17.98 | 18.15 | 0.94 |
3 | 3.70 | 2.31 | 4.495 | 2.88 | 21.63 | 24.92 | 13.20 |
Deviation | — | — | — | — | — | — | 0.0524 |
Figure
Moisture content in aggregates over increase ratio of air voids through freeze-thaw cycles.
From Table The CTAV value was lower than the MAV value before and after freeze-thaw cycles in the RH-WMA samples with dried and moist aggregates. An explanation for this phenomenon was the difference of air void distribution between the RH-WMA samples and scanned samples. The air voids in the top and bottom regions of the samples were calculated in MAV, but not computed in CTAV. In addition, the relative error between the CTAV and MAV of the mixtures through freeze-thaw cycles was 0.0939 and 0.0921, respectively. There is a good correlation between CTAV and MAV. The CTAV and MAV of the samples with different moisture contents of aggregates before freeze-thaw cycles were lower than those after freeze-thaw cycles. The freeze-thaw cycles affected the changes of CTAV and MAV of the samples. There is a linear relationship between the increasing ratios of CTAV and MAV and moisture content in aggregate, and
Due to a high heterogeneity in the surface region of the RH-WMA samples, the samples were cut 4 mm from the top and bottom in height and 4 mm in diameter. Thus, the samples with 56 mm in height and 96 mm in diameter were obtained, as shown in Figure
RH-WMA cut samples.
Table
The increase ratio of CTAV of the RH-WMA cut samples through freeze-thaw cycles.
Moisture content (%) | CTAV before freeze-thaw cycles (%) | CTAV after freeze-thaw cycles (%) | Increase ratio of CTAV (%) |
---|---|---|---|
0 | 1.00 | 1.07 | 5.78 |
1 | 0.97 | 1.05 | 7.71 |
2 | 1.24 | 1.35 | 8.24 |
3 | 1.31 | 1.43 | 9.56 |
Moisture content in aggregates and increase ratio of CTAV through freeze-thaw cycles.
Figure
The visual 3D model was established for changes of air voids, and the transparency was processed by VG Studio MAX 2.2. The air void distribution of the samples was determined, and it is assumed that the transparency of the 3D model was zero. The air void sizes in the same cross section of the RH-WMA samples with dried/moist aggregates before and after freeze-thaw cycles were observed. The difficult part of the process is that determination of the same cross section of the samples before and after freeze-thaw cycles. The iron wire was needed to mark and identify the same cross section position of the samples before and after freeze-thaw cycles.
The mixtures were kept in the 40°C oven for about 4 h to ensure the consistent conditions at the time of scanning and the consistent threshold value during image processing were used. The 3D images were converted into two-valued processing by Image-Pro Plus. The area with a projection of the air voids was quantified, and the changes of air voids before and after the freeze-thaw cycles were illustrated, as shown in Figure
Comparisons of air void size of the RH-WMA before and after freeze-thaw cycles. (a) The air void size before freeze-thaw cycles. (b) The air void size after freeze-thaw cycles.
The air void sizes of the samples with dried/moist aggregates increased after freeze-thaw cycles. The air void sizes became larger compared to the analyzed images of the mixtures with dried/moist aggregates before freeze-thaw cycles. The change of air void size in RH-WMA with 3% moisture content was prominent. The more residual moistures in aggregates are, the larger the change of air void size of the RH-WMA is.
The indirect tensile test was conducted on the RH-WMA with dried/moist aggregates after freeze-thaw cycles. The results for TSR of the mixtures and increase ratio of CTAV are shown in Table
ITS/TSR of the RH-WMA uncut samples with dried/moist aggregates.
Moisture content (%) | Dry ITS (kN) | Wet ITS (kN) | TSR (%) | Increase ratio of CTAV (%) |
---|---|---|---|---|
0 | 7.11 | 5.00 | 78.39 | 13.94 |
1 | 6.16 | 4.28 | 73.54 | 16.92 |
2 | 5.61 | 3.64 | 68.99 | 18.15 |
3 | 5.28 | 3.04 | 61.70 | 24.92 |
Figures
ITS/TSR versus moisture contents of an aggregate of the RH-WMA samples.
ITS/TSR versus increase ratios of CTAV of the RH-WMA.
Figure
The air void contents of the RH-WMA samples decreased as the air void size increased, and air void size was mostly distributed in the range of 0–5 mm3 and 5–10 mm3. It indicated that the morphological change of residual moisture in aggregates and freeze-thaw cycles affected on the air void distribution. During the freeze-thaw cycles, the water-ice cycles slightly expanded the volumes of air voids due to the volume changes from water to ice. From the 3D segmentation images of the RH-WMA samples, a high percentage of air voids was present in the top and bottom regions. The percentage of air voids in the middle region was small and uniformly distributed. The air void distribution related to many factors, such as aggregate gradation and method of compactions. The higher the moisture content of aggregates is, the larger the increment of air void content is. The higher increase ratio of CTAV and MAV of the RH-WMA samples was also observed. The initial moisture contents in aggregates can definitely affect the performance and volume characteristics of asphalt mixtures. If the ITS and TSR are lower, the increase ratio of CTAV is larger. The moisture susceptibility of the RH-WMA samples was affected by changes in air voids. The ITS and TSR of the RH-WMA samples decreased with the increase of moisture contents of aggregates. There is a good linear relation between ITS/TSR of the RH-WMA samples and the moisture contents of aggregates through freeze-thaw cycles. The RH-WMA samples with the high moisture content of aggregates tend to be more sensitive to moisture damage. In addition, it is recommended that the moist aggregates are stored in the dry area and different heating steps are used for fully drying.
The data that support the findings of this study are available from the corresponding author upon reasonable request.
The authors declare no conflicts of interest.
This study was sponsored by the National Natural Science Foundation of China (51478028 and 51778038) and Program for Changjiang Scholars and Innovative Research Team in University (IRT-17R06). The authors appreciate the funding support from Beijing Key Laboratory of Traffic Engineering, Beijing University of Technology, under grant no. 2018BJUT-JTJD007. The authors also appreciate Mr. Gao Jinqi for his help in the experiments.