Mechanistic-based characterisation of fatigue resistance of alternative mix designs

The population and economy in the State of Qatar have been increasing significantly in the past 10 years. Accordingly, traffic loading has also increased rapidly, which merits consideration of the design and construction of long-lasting pavement structures that require minimal maintenance. This study started with an investigation of the feasibility and performance of the current asphalt pavement materials and structures being used in the country. This investigation utilised the analysis approach implemented in the Mechanistic-Empirical Pavement Design Guide (M-E PDG) software. The results evidenced how effective it is to replace the conventional unmodified 60-70 Pen bitumen with polymer-modified PG76-10 bitumen for pavements in Qatar and other countries in the region with similar climatic conditions. In addition, the results showed that the use of perpetual pavement structures is a viable option economically and that they are much more accommodating of increase in traffic loading, without causing excessive damage, than conventional pavement structures. The study also concentrated on the assessment of the long-term performance of different full-scale perpetual trial sections by conducting several field tests. The field performance evaluation results showed slightly low resistance to rutting, high IRI values, to some extent, and low stiffness during summertime when the temperature is high. These distresses and deteriorations are expected given the huge traffic loading and the big difference in temperature between seasons in Qatar. Then several field cores, field mixtures, and laboratory mixtures were tested and evaluated in order to assess the performance of different asphalt concrete mixtures against rutting, fracture, temperature susceptibility and fatigue damage. The conducted tests were useful to characterise and assess the performance of the mixtures against several major distresses. The results indicated that resistance of asphalt concrete mixtures to rutting was mainly affected by the bitumen grade, aggregate source and aggregate gradation. A well-designed mixture that uses polymer-modified bitumen (PMB) can achieve the high rut-resistance of asphalt mixtures either with Gabbro or limestone aggregates. The use of polymer-modified bitumen reduced the temperature and frequency susceptibility on the stiffness and rut-resistance. According to these results, it is obvious that rutting and cracking would not be major distresses for asphalt mixtures in Qatar if the mixtures were designed following a Superpave mix design with the appropriate content of polymer-modified bitumen. However, given the current mixture design system in Qatar, which utilises very low bitumen content, fatigue damage or cracking in general is a main distress, and its characterisation should be investigated in depth. The traditional methods to interpret fatigue tests data are not sufficient to characterise and evaluate mixtures against fatigue damage. Therefore, two advanced fatigue characterisation approaches were performed on the raw data obtained from the fatigue test of specimens prepared from different mixtures. The analysis of the fatigue tests focused on calculating the dissipated energy (DE) and obtaining damage characteristics curves following a comprehensive viscoelastic continuum damage (VECD) approach. The test results showed that the use of the VECD approach has major advantages over obtaining DE results only. However, the predicted fatigue life (Nf) for each asphalt mixture was affected by the uncertainty associated with fatigue tests as well as with model parameters. Therefore, it was important to develop a probabilistic analysis approach that accounts for the uncertainty and the variability associated with fatigue tests and analysis, respectively. To that end, a novel probabilistic analysis approach has been proposed in the last chapter of this study for predicting the performance of asphalt mixture against fatigue damage. The VECD characterisation approach was used in the development of this probabilistic analysis model. The random variables (RVs) of the fatigue life (VECD-Nf) model (|E*|LVE, a, b and α) were generated following normal distribution functions. However, it is suggested that more specimens should be tested in the future to specify the true distribution functions for the RVs. In conclusion, it is clear that the fatigue life results of the probabilistic analysis approach were much more consistent and reliable than those of the deterministic analysis approach. This probabilistic approach coupled with VECD results is very practical and useful for engineers and will be beneficial to predict fatigue cracking resistance of asphalt mixtures in the field