Resumen
Fatigue failure in asphalt pavements, caused by continuous loading, primarily occurs at the interface between the asphalt binder and the aggregate (adhesive failure) or within the asphalt binder itself (cohesive failure). This study conducted variable stress fatigue tests on asphalt binders to investigate the fatigue damage evolution mechanism that aligns most closely with actual road conditions. By altering test conditions such as stress amplitude and loading frequency, the study summarized the patterns of change in the fatigue performance parameters of binder samples and analyzed their fatigue life. The research methods employed are of significant value for refining the existing asphalt fatigue evaluation systems. Indicators such as dissipated energy ratio, cumulative dissipated energy ratio, rate of change in dissipated energy, and the G-R constant n × Gn* were used as criteria for fatigue failure. The fatigue life of samples under different stress levels was calculated, and the applicability of each dissipated energy fatigue indicator was evaluated. The evaluation indicators, like dissipated energy ratio and cumulative dissipated energy ratio, failed under large stress outside the non-linear viscoelastic range, whereas the rate of change in dissipated energy and the G-R constant n × Gn* accurately determined the fatigue life of samples in fatigue tests at all stress levels, covering a broader range of applicable stresses. In variable stress fatigue tests, the rate of change in dissipated energy and the G-R constant were used as indicators for determining fatigue failure. Under a small-to-large loading mode, the second stage of the sample?s fatigue life was too short, causing the G-R constant curve to not reach its peak, and hence it could not accurately determine the sample?s fatigue life. Under a large-to-small loading mode, there are sufficient loading cycles before fatigue failure occurs, allowing the G-R constant curve to reach its peak, demonstrating that the G-R constant remains applicable in this loading mode. Under both small-to-large and large-to-small loading modes, the fatigue failure point of the samples could be determined based on the rate of change in dissipated energy curve. It is recommended to use the rate of change in dissipated energy as the failure criterion for samples in variable stress fatigue tests on asphalt binders.