Stochastic finite element analysis of moisture damage in hot mix asphalt

Moisture damage is one of the major problems that can be faced by a pavement during the design life. It can tremendously reduce a pavement's strength and consequently its life. Moisture sensitivity testing of asphalt mixtures is critical for ensuring performance expectations are met. Moisture susceptibility is most commonly tested using the modified Lottman test. The shift towards mechanistic design calls for the utilization of a more fundamental test to evaluate moisture damage. The evolution of unconfined dynamic modulus and creep (flow number) tests as performance tests for inclusion in the Superpave mix design process make these candidate tests for inclusion in moisture sensitivity testing. The challenge in moisture sensitivity testing is the ability to capture the various mechanisms that cause moisture damage. Previous research has recommended the use of the dynamic modulus test for moisture damage evaluation. The dynamic modulus test results can be used to develop master curves that can be used to predict pavement performance at any temperature and/or frequency.An objective of this study was to identify the appropriate test that can identify whether a mix is moisture susceptible or not. Indirect tensile test, dynamic modulus test and flow number test were investigated to satisfy this objective. Another objective was to use finite element modeling to evaluate the moisture susceptibility and variability of a mixture.In the present study, sixteen field procured mixtures were subjected to five different modes of moisture conditioning: (1) unconditioned without water submersion testing, (2) unconditioned with water submersion testing, (3) moisture saturation with water submersion testing, (4) moisture saturation with freeze/thaw conditioning without water submersion testing, and (5) moisture saturation with freeze/thaw conditioning and with water submersion testing. These samples were tested for flow number.Dynamic modulus tests were performed on both moisture conditioned and unconditioned samples. The results were used to develop mastercurves. The dynamic modulus results were used as input to a finite element model in which stochastic variation of the results were incorporated in the model. The model was validated by the results from the flow number test.The methodology was applied on sixteen projects and the results were compared to the results achieved using the AASHTO T283 methodology and dynamic modulus test results. The dynamic modulus test results show consistency with AASHTO T283 in identifying moisture sensitivity of a mixture. This dissertation outlines a method for evaluating hot mix asphalt moisture susceptibility utilizing dynamic modulus testing and is compatible with the proposed performance testing for accompanying Superpave volumetric mix design. The results of the proposed mixture dynamic modulus moisture susceptibility method can also be used in the new M-E PDG for evaluating the moisture susceptibility effects of the tested mixtures. This in part allows for the evaluation of this environmental effect in the M-E PDG.The results show that the dynamic modulus test has good potential to identify the moisture susceptibility of the material provided that it is combined with the field and loading conditions. The flow number test also showed good potential when it was analyzed using the Ohio State model. The data showed consistency but a comparison to field performance is needed to identify whether the results are correlated to field performance or not. The finite element analysis showed that the results' variability increase with moisture conditioning and that moisture conditioned samples are more susceptible to rutting. Finite element model is a good tool to be combined with the dynamic modulus test to be able to evaluate the moisture susceptibility based on site condition.