Page 39 - MODELING OF ASPHALT CONCRETE
P. 39

Modeling of Asphalt Binder Rheology and Its Application to Modified Binders      17


                    measure and to define a temperature susceptibility parameter. Again several types of
                    relations have been used: slope of log viscosity versus temperature (Traxler and Schweyer
                    1936), log viscosity versus reciprocal of temperature raised to a variable power (Traxler
                    and Schweyer 1936; Cornelissen and Waterman 1956), and log-log viscosity versus log
                    temperature (Fair and Volkmann 1943). The latter measure is probably the most widely
                    accepted and is referred to as viscosity-temperature susceptibility or VTS.
                       A third type of approach used a combination of consistency measures. The penetration
                    index (PI) for example combines penetration and softening point (Van der Poel 1954), and the
                    penetration viscosity number (PVN) combines penetration and viscosity (McLeod 1972).
                       The best indication of how well these different susceptibility measures represent the
                    rheological properties of asphalts is probably found in durability studies that used these
                    measures to define asphalt performance with age hardening. The conclusions about the
                    effect of aging on temperature susceptibility are controversial: Pfieffer (1950) and Blokker
                    and Hoorn (1959), using the penetration index, reported a decrease in temperature
                    susceptibility upon oxidative aging. Halstead and Zenewitz (1961) using plots of log-log
                    viscosity versus temperature, in the range of 60 to 95°F, indicated that some asphalts
                    show a decrease in temperature susceptibility while others show an increase upon aging.
                    McLeod (1972) used the penetration index (PI) and his penetration viscosity number
                    (PVN) and observed that after 9 years of service asphalts showed either no change or a
                    significant increase in temperature susceptibility. However, the PI did not show the same
                    change in temperature susceptibility. Puzinauskas (1979) used the VTS and concluded
                    that the majority of asphalts show an increase in temperature susceptibility. Anderson
                    and his coworkers (1983) used three different parameters (PI, PVN, and VTS) and
                    concluded that PI and VTS show a general increase in temperature susceptibility, while
                    PVN does not show significant changes with aging. The authors, experimenting on a
                    very large number of asphalts, indicated that these measures do not measure the same
                    property, and no simple explanation can justify the difference in the effect of oxidative
                    aging on the values of these parameters. Button and his coworkers (1983) confirmed
                    Anderson’s findings. The authors found that, using the PVN (estimated from pen77 and
                    vis140) as the temperature susceptibility measure, asphalts that were originally highly
                    temperature susceptible became more susceptible with oven aging, while those that
                    were less temperature susceptible became even less susceptible with aging.
                       Others have also used apparent viscosity as the measure of consistency. Moavenzadeh
                    and Stander (1967) measured apparent viscosity at a constant power input (1000 ergs) over
                    a range of temperatures between 10 and 160°C. The authors indicated that aged asphalts
                    are less temperature susceptible in the lower temperature range while more susceptible in
                    the higher temperature range. In other words, the temperature susceptibility shows
                    different changes at different ranges of temperature. Kandhal and his coworkers (Kandhal
                    et al. 1973; Kandhal and Wenger 1975) confirmed Moavenzadeh and Stander findings.
                    Here researchers measured apparent viscosity at 39, 77, 140, and 275°F and observed that
                    temperature susceptibility measured in terms of log-log viscosity versus log temperature,
                    decreased in the lower temperature range with age hardening while it increased in the
                    higher temperature range.
                       No doubt the above review is confusing. Two fundamental reasons can be given to
                    explain the problems with temperature susceptibility parameters:

                         1.  The temperature susceptibility, as shown in linear viscoelastic characterization
                           studies, is both temperature range and time of loading dependent. The relations
                           are not linear nor can be approximated by a linear relation such that a few
   34   35   36   37   38   39   40   41   42   43   44