One of the main requirements of designing perpetual pavements is to determine the endurance limit of Hot Mix Asphalt (HMA). The purpose of this study was to validate the endurance limit for HMA using laboratory beam fatigue tests. A mathematical procedure was developed to determine the endurance limit of HMA due to healing that occurs during the rest periods between loading cycles. Relating healing to endurance limit makes this procedure unique compared to previous research projects that investigated these concepts separately. An extensive laboratory testing program, including 468 beam tests, was conducted according to AASHTO T321-03 test procedure. Six factors that affect the fatigue response of HMA were evaluated: binder type, binder content, air voids, test temperature, rest period and applied strain. The endurance limit was determined when no accumulated damage occurred indicating complete healing. Based on the test results, a first generation predictive model was developed to relate stiffness ratio to material properties. A second generation stiffness ratio model was also developed by replacing four factors (binder type, binder content, air voids, and temperature) with the initial stiffness of the mixture, which is a basic material property. The model also accounts for the nonlinear effects of the rest period and the applied strain on the healing and endurance limit. A third generation model was then developed by incorporation the number of loading cycles at different locations along the fatigue degradation curve for each test in order to account for the nonlinearity between stiffness ratio and loading cycles. In addition to predicting endurance limit, the model has the ability to predict the number of cycles to failure at any rest period and stiffness combination. The model was used to predict fatigue relationship curves for tests with rest period and determining the K1, K2, and K3 fatigue cracking coefficients. The three generation models predicted close endurance limit values ranging from 22 to 204 micro strains. After developing the third generation stiffness ratio model, the predicted endurance limit values were integrated in the strain-Nf fatigue relationships as a step toward incorporating the endurance limit in the MEPDG software. The results of this study can be used to design perpetual pavements that can sustain a large number of loads if traffic volumes and vehicle weights are controlled.

Perpetual Pavements, if properly designed and rehabilitated, it can last longer than 50 years without major structural rehabilitation. Fatigue endurance limit is a key parameter for designing perpetual pavements to mitigate bottom-up fatigue cracking. The endurance limit has not been implemented in the Mechanistic Empirical Pavement Design Guide software, currently known as DARWin-ME. This study was conducted as part of the National Cooperative Highway Research Program (NCHRP) Project 9-44A to develop a framework and mathematical methodology to determine the fatigue endurance limit using the uniaxial fatigue test. In this procedure, the endurance limit is defined as the allowable tensile strains at which a balance takes place between the fatigue damage during loading, and the healing during the rest periods between loading pulses. The viscoelastic continuum damage model was used to isolate time dependent damage and healing in hot mix asphalt from that due to fatigue. This study also included the development of a uniaxial fatigue test method and the associated data acquisition computer programs to conduct the test with and without rest period. Five factors that affect the fatigue and healing behavior of asphalt mixtures were evaluated: asphalt content, air voids, temperature, rest period and tensile strain. Based on the test results, two Pseudo Stiffness Ratio (PSR) regression models were developed. In the first model, the PSR was a function of the five factors and the number of loading cycles. In the second model, air voids, asphalt content, and temperature were replaced by the initial stiffness of the mix. In both models, the endurance limit was defined when PSR is equal to 1.0 (net damage is equal to zero). The results of the first model were compared to the results of a stiffness ratio model developed based on a parallel study using beam fatigue test (part of the same NCHRP 9-44A). The endurance limit values determined from uniaxial and beam fatigue tests showed very good correlation. A methodology was described on how to incorporate the second PSR model into fatigue analysis and damage using the DARWin-ME software. This would provide an effective and efficient methodology to design perpetual flexible pavements.