文摘
W?hler’s law constitutes the foundation for the fatigue models used in mechanical–empirical (M–E) pavement design methods worldwide (Perraton, Correspondances entre les coefficients des modèles de fatigue dans les méthodes mécanistiques-empiriques de dimensionnement de chaussées souples, 2011). Often, pavement design software will consider constant parameters in describing bituminous mixtures W?hler’s law and focus on the calibration process to correlate modeling response to the fatigue cracking network observed on roads. To better define specific bituminous mixtures fatigue behaviour to be considered for modeling in an M–E pavement design method, W?hler curves obtained using the classical fatigue criterion (N f50%) were compared to those established by considering a more appropriate failure criterion (N fII/III). Fatigue measurements on GB20 asphalt concrete mixes with 2 different types of bitumens (mix A: PG58-28 and mix B: PG64-28) were investigated. In reference with N fII/III and N f50% values, W?hler curves obtained for mix A are similar even though a clearly different tendency appears for mix B fatigue results; these results illustrate the inability of the classical fatigue criterion (N f50%) to properly define the beginning of the micro-crack process in the material associated to the transition point between phases II and III in the relationship between the stiffness and the number of cycles of a sample tested in fatigue. Moreover, fatigue test results confirm that the fatigue life links to phases II/III threshold is strongly associated with a critical micro-crack density in the sample.