Page 352 - Mechanics of Asphalt Microstructure and Micromechanics

T
344 Ch a p t e r e n

very important in the simulation of composite materials besides the properties of the
individual component.
In both the physical and simulation of the test, the stiffness of the asphalt binder is
fairly small compared with samples with aggregates. The stiffness of each sample was
calculated at a displacement of 0.01 mm from both physical and simulative test results.
The relationship between stiffness and aggregate volume fraction, shown in Figure 10.19,
demonstrated a trend of stiffening with aggregate content. It also indicates an increased
rate of stiffness enhancement with the increase of aggregate volume fraction.

10.3 Digital Test for Simulative Test (APA Test)
Following the same philosophy, the digital specimen and digital test techniques can be
applied to simulative tests such as the Asphalt Pavement Analyzer (APA) test. The sim-
ulative tests mean applying a repeated scaled (could be full scale) load on a scaled
wheel (could be full scale) onto a scaled pavement section (could be full scale) to induce
pavement responses and to evaluate its performance. The simulative tests include the
full-scale accelerated pavement test, model mobile load simulator, the Hamburg test
and the APA test. In this section, the APA test is used as an example for developing its
digital counterpart, the digital specimen, and digital test. Figures 10.20a, b, and c illus-
trate the three sequent steps in building the digital specimens, applying boundary con-
ditions and loading conditions. Figure 10.20d illustrates how a digital specimen is taken
from a 3D image.
In this application, a slightly different method is adopted. Test results are compared
statistically. Two specimens (150 mm diameter) for each of the three mixes were cored
from the WesTrack project (Wang et al., 2001). The six specimens were then scanned us-
ing XCT. During the construction process, specimens were manufactured and then
tested in an APA. The digital tests are then based on the digital specimens from the six
scanned cores. During the test, a rolling wheel applied load to the specimens through a
pressurized hose placed on the specimens, and the specimens with the holding mold
were fixed in a temperature-controlled chamber, so that the specimens could be tested
under the specified temperature condition. For this study, the applied load was 100 lbs.
The pressure in the hose was 100 psi, and the testing temperature was 40°C. For a stan-
dard APA test, a typical 8000 cycles of loading are applied to each specimen. Five digital
specimens were cored from different locations of the same scanned specimens as illus-
trated in Figure 10.20d. The resolution of the images was 0.29 mm/pixel. The spacing
between two images was 1 mm. Based on the resolution of the image, the particle size
smaller than 0.29 mm is invisible.

10.3.1 Computational Efficiency
The objective of the image processing and analysis step is to decrease the image size to
improve the computational efficiency. The size of the original images is 512 512 pix-
els. If the original images are used for simulation with 1 pixel (voxel) as one element,
the computation time will be too expensive. Figure 10.21 shows the relationship be-
tween the computation time and the number of elements. The computation time drasti-
cally increases with the increase of the number of elements. Therefore, the re-digitized
images with reduced pixel numbers were used in the simulations. Image analysis soft-
ware, Image-Pro Plus, was used to process the images. An analysis was conducted to