Strain-Level Determination Procedure for Small-Specimen Cyclic Fatigue Testing in the Asphalt Mixture Performance Tester

With an increase in small-specimen cyclic fatigue testing using the Asphalt Mixture Performance Tester (AMPT), researchers have observed that the strain-selection guidelines in AASHTO TP 107-14 that are intended for large AMPT cyclic fatigue tests are inadequate for testing small specimens. The machine compliance factor is significantly different for testing small specimens compared with large specimens because of different required load levels, resulting in a significant offset in the relationship between the input strain and the number of cycles to failure. To this end, this paper presents the development and verification of a phenomenological model that relates strain levels to dynamic modulus and number of cycles to failure for small-specimen AMPT cyclic fatigue tests, as well as the development of a corresponding stepped strain-level determination procedure that takes into account cases when the initially selected strain-level results in an unexpected number of cycles to failure. The final procedure includes a table with input strain levels and step strain increments for a wide range of dynamic modulus values as well as a flow chart to guide the use of the step strain adjustment procedure.

EVALUATION OF THE HIRSCH MODEL FOR DYNAMIC MODULUS ESTIMATION OF ASPHALT MIXTURES

The dynamic modulus of the asphalt mixtures is an important factor in designing or analyzing an asphalt concrete pavement, but it is expensive and time consuming to measure. Therefore, it is important to develop a model to predict this value. In this regard, the Hirsch model is a popular model, however, it is developed based on a range of U.S. asphalt mixtures and standards. Therefore, it is not certain that it can be used for asphalt mixtures based on materials and codes other than U.S. This article investigated whether this model performs satisfactorily with two typical asphalt mixtures in Western Australia (WA) containing 0, 10, 20, and 30% of recycled asphalt pavement. To do so, cylindrical samples were made with materials and locally established standards in Western Australia and then tested in Asphalt Mixture Performance Tester (AMPT) machine to acquire their dynamic modulus and phase angle values in different loading frequencies (0.01 to 10 Hz) and temperatures (4 to 40°C). Meanwhile, the results are estimated by the Hirsch model using some properties of the mixture and binder. The properties of the binder in different test conditions are obtained using a dynamic shear rheometer. The comparison of the results showed that the dynamic modulus underestimation or overestimation error can reach to 50 and 280% respectively. Generally, this model did not perform well in this study.

Practical Method to Determine Strain Level for Cyclic Direct Tension Fatigue Testing in the Asphalt Mixture Performance Tester

This paper presents a practical procedure for estimating strain levels in the asphalt mixture performance tester cyclic direct tension fatigue test and provides practitioners with guidance for selecting a strain value for testing materials with unknown fatigue characteristics. A large variety of plant-produced and laboratory-prepared mixtures were analyzed. These included hot- and warm-mix asphalt, reclaimed asphalt pavement, and recycled asphalt shingles and various gradations and air void contents; all were tested following the AASHTO TP 107 procedure. The experimental results from a segregated material illustrate that AASHTO TP 107 could produce repeatable results and was sensitive to field variations in binder content. Data for satisfactorily performing and poorly performing mixtures were clustered on the plot of cycles to peak phase angle versus actuator strain level and a Black Space diagram evaluated at the AASHTO TP 107 test temperature. Curves of cycles to peak phase angle versus actuator strain level were built from the experimental database, with each curve representing one ideal mixture that followed a particular power function. The dynamic modulus data are used to determine a recommended actuator strain level to start the set of fatigue tests for each cluster.