Time-temperature equivalence and unified characterization of asphalt mixture fatigue properties.

• A new fatigue model of asphalt mixture is established, which can normalize the fatigue characteristics under various temperatures and several loading frequencies. • Time-temperature equivalence of fatigue life is demonstrated and fatigue master curves was formed. • The time-temperature equivalence happens on fatigue and strength at the same time in the new model. This work presents a new method for evaluating the time-temperature equivalence and unified characterization of the fatigue performance for asphalt mixtures. For this purpose, direct tensile strength tests and direct tensile fatigue tests were carried out on asphalt mixtures at different temperatures and loading rates. A time-temperature-dependent stress ratio was established. By exploring the correlation between stress ratio and fatigue life, a new fatigue model considering the effect of temperature and loading time was built. The time-temperature equivalent of fatigue performance was confirmed by analysis of the fatigue dates under several temperatures and loading frequencies. The results showed that the direct tensile strength of the asphalt mixture changed non-linearly with the variation of the loading rate or temperature. The stress ratio was also related to temperature and loading rate. The time-temperature stress ratio was linearly related to fatigue life in a double logarithmic coordinate system, and a new S N fatigue curve was formed. The fatigue life under several temperatures and various loading frequencies can be described in one fatigue model and the time-temperature equivalent can be explored. The master curve of fatigue life or stress ratio was built. The time-temperature superposition principle was applicable for fatigue life estimation and stress ratio calculation. The fatigue life under different loading frequencies and various temperatures of the pavement in service can be completely estimated. The relative error between the test results and the calculated values is less than 30%. [ABSTRACT FROM AUTHOR]