Page 118 - MODELING OF ASPHALT CONCRETE

96 Cha pte r F o u r

Control Mode In the original ASTM protocol, the dynamic modulus test was conducted
in load control by applying constant compressive cyclic loading throughout the
frequency sweep. This same approach was planned to be adapted to the new test
protocol. However, the applied constant stress causes the subsequent resilient strain to
increase as frequencies decrease in the frequency sweep due to the viscoelastic nature
of the mix. The proposed new test protocol considers 150 microstrains as the limit for
the linear viscoelastic region. However, at warm temperatures the low-frequency strain
can be up to three times larger than strain at high frequencies, when constant stress is
used. Then, at high temperatures the magnitude of the strain may exceed 150 microstrains
at low frequencies although the strain at high frequencies stays under 50 microstrains.
This may cause some damage accumulation in the specimen, in addition to a creep
imposed to the cyclic straining pattern. This needs to be accounted for in the data
analysis when obtaining modulus and phase angle values.
Therefore, some European researchers (Di Benedetto and de la Roche 1998;
Doubbaneh 1995) prefer to do the cyclic testing in strain control mode by applying
tension-compression loading. However, specimens need to be glued to the platens to
apply tension. This loading model gives zero average stress, which enables the steady-
state straining pattern and, thus, eliminates the creep. The strain amplitude can also be
controlled to stay below desired limit during testing. In pure tension testing, strain
keeps creeping as a function of time without achieving steady-state pattern.

Rest Period in the Frequency Sweep A rest period between frequencies in the frequency
sweep is not discussed in the proposed new test protocol. However, some of the
controllers used in the past and perhaps some current ones cannot produce continuous
frequency sweeps; each frequency must be programmed separately in such a manner
that there will be some lag time or rest period between each frequency. The rest period
helps to prevent specimen from heating during cyclic testing, although for modulus
tests the number of cycles is usually limited to fewer than 200 and heat increase may not
be a problem. However, the rest period allows some of the transient strains to recover
during testing which may have some effect on the measured modulus values and
selection of suitable data analysis methods.

Compressive Axial versus SST-Shear Modulus

Correlation of Modulus and Phase Angle Values
A study conducted by Witczak et al. (2000) compared SST frequency sweep and uniaxial
∗
dynamic modulus tests and concluded that the shear modulus value |G | obtained
∗
from the SST test was not a true fundamental property such as |E | due to the
instrumentation and specimen size problems. Thus, according to Witczak et al., test
results from the SST device can be categorized as representing an index value of shear
modulus of the asphalt mixture, not a fundamental material property.
∗
∗
This means that the |G | cannot be used directly to replace the |E | in the pavement
design applications where mixture stiffness is needed. Equation (4-8) gives a conversion
model developed by Christensen, Pellinen, and Bonaquist (2003), which can be used to
∗
∗
convert |E | to |G |. Both moduli values have units of psi in the equation:
|G ∗ |= 0 .0603 |E ∗ | . 1 0887 R = . 0 93 (4-8)
2