Page 217 - MODELING OF ASPHALT CONCRETE

VEPCD Modeling of Asphalt Concr ete with Gr owing Damage 195

Because the permanent strain at the end of the recovery is used to determine the VP
model coefficients, the predicted viscoplastic strains are in good agreement with the
measured permanent strain at the end of the recovery. Also, it is found that the
viscoelastic strain recovers completely after the rest period.
The most noteworthy observation to be made from Fig. 7-21 is the overprediction of
the viscoelastic strain and, therefore, the total strain. The degree of overprediction is
found to increase as the temperature increases. The same observation has been made
from the monotonic prediction although it is not shown in this chapter to save space.
The fact that the VEPCD model predicts the material’s behavior in tension extremely
well but very poorly in compression suggests that the missing mechanism in the VEPCD
model is unique to compression loading.
In order to investigate the difference between the behavior of asphalt concrete in
tension and compression, the stress versus pseudostrain relationships are examined
first. Figure 7-22 presents the stress versus pseudostrain curves for intermediate to high
temperatures. The stress versus pseudostrain curves for 5°C are shown in Fig. 7-20. It
can be seen from these figures that, as the temperature increases, the stress versus
pseudostrain curve changes from a simple softening shape to a more complex shape.
That is, at 5°C, the stress versus pseudostrain relationship starts along the line of
equality (i.e., viscoelasticity dominates the behavior with minimal microcracking
damage) and then changes to a softening curve, indicating the stiffness reduction due
to microcracking in the vertical direction. At the peak stress or slightly over the peak
stress, the localization starts, which is the beginning of the macrocrack propagation. At
higher temperatures, the stress versus pseudostrain curve starts along the line of
equality, and then the slope changes to an upward direction, indicating the hardening
behavior. The shape of the curve changes finally to represent the softening behavior
followed by the failure at the peak stress. This pattern becomes more evident as the
temperature increases and the rate of loading decreases. It is noted that, in tension, this
pattern was never observed.
The primary mechanisms that govern the constitutive behavior of asphalt concrete
in tension are viscoelasticity, the plastic flow of the binder, and cracking. In compression,
it is well known that the interlocking of aggregate particles is an important factor that
affects the behavior of asphalt concrete. The effect of aggregate interlocking increases as
the binder viscosity decreases, which happens when the temperature increases and the
rate of loading decreases. The primary characteristic of aggregate interlocking is that it
stiffens and becomes more significant as the deformation of the asphalt concrete increases
until the aggregate particles begin to slip. The observations made from Fig. 7-22 are well
supported by the expected behavior of asphalt concrete due to aggregate interlocking.
It must be noted that this behavior cannot be detected in the stress versus strain
plots because of the mixed effects of viscoelasticity and aggregate interlocking. The
benefit of using pseudostrain (i.e., eliminating the viscoelasticity from the plot) is clearly
demonstrated in this figure.
In order to display this hardening and softening behavior of asphalt concrete in
compression more effectively, the apparent pseudo secant stiffness (C ) is calculated and
A
presented in Fig. 7-23. The pseudo secant stiffness in this figure is called apparent because
the true pseudo secant stiffness is calculated using the viscoelastic strain only. In this
figure, the apparent pseudo secant stiffness is calculated using the total strain, which
includes both viscoelastic and viscoplastic strains and the aggregate interlocking effect
on these strains. The first and second numbers shown in the legends of three figures in