Reducing Inconsistency of HMA Flexure Fatigue Testing

Abstract

Beam fatigue testing of hot-mix asphalt (HMA) in the laboratory has been used for several decades by many researchers around the world. A total of two standard procedures are currently available: one that controls constant haversine deflection and another that controls constant sinusoidal deflection. The literature shows a large variability in shift factors between laboratory and field fatigue results. Also, recent literature shows inconsistent waveforms when a haversine deflection is used, which could be one of the reasons causing this large variability. The paper examines the results of the standard procedures and provides a rational explanation of the differences between them. Because of the inconsistency of the ASTM flexural fatigue test standard, it incorrectly produces a larger number of cycles to failure than the AASHTO test, with ratios of approximately 10, which varies depending on the stiffness of the material. Also, specimens tested according to ASTM flexural fatigue test standard, with twice the input tensile strain of specimens tested according to AASHTO flexural fatigue test standard at the same temperature, produced approximately the same number of cycles to failure. An approximate method is recommended to correct old ASTM flexural fatigue test standard results by assuming that the obtained number of cycles to failure corresponds to approximately 55–65% of the tensile strain that was inputted to the machine. More studies need to be conducted to verify the results using other materials and test conditions.