Page 67 - MODELING OF ASPHALT CONCRETE

Modeling of Asphalt Binder Rheology and Its Application to Modified Binders 45

The same lack of correlation is observed, as the R-squared value is lower than
20 percent. Similar values of correlation were observed for the individual aggregates.
∗
These results show clearly that the G sind of the binders is a poor indictor of the fatigue
life of the mixtures (N50), as defined by the number of cycles at which the initial mixture
modulus reduces by 50 percent. These findings started the process of looking for a
better indicator of the role of binder in fatigue damage and resulted in putting more
emphasis on the development of a better binder test for the evaluation of binder fatigue
behavior, as will be discussed next.

Development of New Tests for Binder Damage Behaviors

Rutting Test
Several protocols were investigated to select a test procedure and a rheological parameter
that could be used as a more effective indicator of the role of binders in mixture rutting
∗
than the parameter G /sind. The selection process was based on two main hypotheses:
1. The strains in the binder domains within the mix are significantly larger than
those at which the binders are subjected to in the DSR.
2. The cyclic loading with complete reversal in strain or stress is not appropriate
for rating the binder’s contribution to rutting resistance caused by cyclic
irreversible loading (also called non-steady-state cyclic deformation or more
simply repeated creep).

The first hypothesis is based on the data collected in a previous study by the author
with coworkers (Bahia et al. 1999), which indicated that modified binders vary
significantly in their strain dependency. It is also based on the finding that mixture
rheological behavior was found highly sensitive to strain level (Monismith 1994). The
second hypothesis was based on the concept of the RSCH and the review of literature
related to the concept of energy dissipation.
To test these hypotheses different testing protocols were used to find a relationship
with mixture rutting performance. The protocols included strain sweeps, stress sweeps,
time sweeps at constant strain and time sweeps at constant stress. In addition, a repeated
creep test was developed and conducted to measure permanent strain behavior of
binders. The analysis of the results led to believe that these strain sweeps and time
sweeps are not promising. None of the cyclic-reversible tests showed a clear potential
to differentiate between binders and to relate strongly to mixture performance.
As a result of this finding a detailed review of the dissipated energy concept and the
derivation of the binder parameters were initiated. This review indicated that although
the cyclic reversible loading could be used to estimate the total energy dissipated during
a loading cycle, for viscoelastic materials that combines permanent deformation and
delayed elasticity, this type of test does not allow the separation between these two
different types of dissipated energy. As shown in Fig. 2-15, during the cyclic reversible
loading only the total energy dissipated is possible to estimate. The rutting mechanism,
as described by many research efforts and measured in the field, does not include
reversible loading required to bring the pavement material to zero deformation. As
shown in Fig. 2-16, rutting is in fact a repeated creep mechanism with sinusoidal loading
pulses. In this case the pavement layer is not forced back to zero deflection but would