Mechanistic Composition–Specific Fatigue Life of Asphalt Pavements

Abstract

Fatigue life is a key factor for designing an asphalt pavement, which is usually determined by a fatigue test with a failure criterion. Most of the existing criteria lack a direct identification of fatigue failure as the cracked area reaches its maximum value, which is specific for each asphalt mixture and depends on its own composition. This paper fills this gap by proposing the usage of damage density, and derives the fatigue life based on the visco-elasto-plastic form of Paris’ law with damage density (or modified Paris’ law) for an asphalt pavement. Furthermore, it provides both the test-based calculation and nontest estimation for the fatigue life. As the starting of fatigue damage, the crack initiation point, also known as endurance limit, is first derived to be a simplified version of the energy-based crack initiation criterion for visco-elasto-plastic materials. As the ending of fatigue damage, the number of load cycles to failure, or fatigue life, is derived from the modified Paris’ law and calculated using the data from the two mechanics-based test protocols, one for microcracking and the other for macrocracking. To derive the fatigue life from the modified Paris’ law, a characteristic curve that relates the pseudo J-integral to the damage density by the coefficients r and q is defined in this study. It takes into account the loading level and structure of a tested specimen or a field pavement. Based on the data analyzed, a strong correlation is found between r and q. In addition, a prediction model is developed for q using simple mixture and pavement design information. The simplified energy-based crack initiation criterion allows a quick estimate of endurance limits using the mixture modulus, surface energy, and air void content. The fatigue lives of 47 asphalt mixtures with different compositions are calculated using the test data, which are reasonable and vary with the binder type, air void content, aging period, and loading level. Moreover, with the discovery of r and q, the fatigue life can be readily estimated without testing by knowing the asphalt binder type and content, aggregate gradation, air void content, mixture modulus, surface layer thickness, and the traffic level. The case studies are presented as examples to demonstrate the sensitivity of the proposed approach as well as the procedure to determine fatigue lives for field pavements. The predicted fatigue lives were verified by field fatigue observations.