Fatigue cracking and moisture damage are two important modes of distresses in asphalt pavements. Recently, the Dynamic Mechanical Analyzer (DMA) was used to characterize fatigue cracking and evaluate the effects of moisture damage on the Fine Aggregate Matrix (FAM) portion of asphalt mixtures. The FAM specimens should be properly fabricated to represent the composition and structure of the fine portion of the mixture. The objective of the first phase of this study was to develop a standard test procedure for preparing FAM specimens such that it is representative of the mixture. The method consists of preparing loose full asphalt mixtures and sieving through different sizes. Then, the ignition oven was used to determine the binder content associated with the small size materials (passing on sieve #16). Sieve #16 is used to separate fine aggregates from the coarse aggregates. The applicability of this new method will be evaluated using a number of asphalt mixtures. The objective of the second phase of this study was to develop software to analyze the data from DMA test. Such software will enable engineers and researchers to perform the complex analysis in very short time. This is Microsoft Windows ® based software, executable in any hardware configuration under this operational system.

abstract: The fatigue resistance of asphalt concrete (AC) plays an important role in the service life of a pavement. For predicting the fatigue life of AC, there are several existing empirical and mechanistic models. However, the assessment and quantification of the ‘reliability’ of the predictions from these models is a substantial knowledge gap. The importance of reliability in AC material performance predictions becomes all the more important in light of limited monetary and material resources. The goal of this dissertation research is to address these shortcomings by developing a framework for incorporating reliability into the prediction of mechanical models for AC and to improve the reliability of AC material performance prediction by using Fine Aggregate Matrix (FAM) phase data. The goal of the study is divided into four objectives; 1) development of a reliability framework for fatigue life prediction of AC materials using the simplified viscoelastic continuum damage (S-VECD) model, 2) development of test protocols for FAM in similar loading conditions as AC, 3) evaluation of the mechanical linkages between the AC and FAM mix through upscaling analysis, and 4) investigation of the hypothesis that the reliability of fatigue life prediction of AC can be improved with FAM data modeling. In this research effort, a reliability framework is developed using Monte Carlo simulation for predicting the fatigue life of AC material using the S-VECD model. The reliability analysis reveals that the fatigue life prediction is very sensitive to the uncertainty in the input variables. FAM testing in similar loading conditions as AC, and upscaling of AC modulus and damage response using FAM properties from a relatively simple homogenized continuum approach shows promising results. The FAM phase fatigue life prediction and upscaling of FAM results to AC show more reliable fatigue life prediction than the fatigue life prediction of AC material using its experimental data. To assess the sensitivity of fatigue life prediction model to uncertainty in the input variables, a parametric sensitivity study is conducted on the S-VECD model. Overall, the findings from this research show promising results both in terms of upscaling FAM to AC properties and the reliability of fatigue prediction in AC using experimental data on FAM. / Dissertation/Thesis / Doctoral Dissertation Civil and Environmental Engineering 2016

CoordenaÃÃo de AperfeiÃoamento de Pessoal de NÃvel Superior / Fatigue cracking is one of the most common distresses in flexible pavements in Brazil and in the world. Fatigue damage in asphalt mixtures originates as a small scale phenomenon. Recently, some authors have used the fine portion of the asphalt mixtures, called Fine Aggregates Matrix (FAM), as an intermediate step to estimate the hot mix asphalt (HMA) behavior. By definition, the FAM is composed of fine aggregates, mineral filler and asphalt binder. There are many volumetric issues of these kind of mixtures. In Brazil, the developed research regarding FAM uses sieve #10 (2.00mm) in the FAM sample preparation, since the #16 sieve (most used in the literature) does not belong to the set of national standardized sieves. Recently, some authors have used the concept of continuum damage mechanics in viscoelastic media in order to quantify the damage during the fatigue testing. In this theory the cracks are considered homogeneous dispersed in the medium, and are represented from internal state variables, determined experimentally. The objective of this study is to evaluate the influence of the variation of one of the FAM constituents, the aggregates, by varying its Maximum Nominal Size (MNS), on damage resistance of these asphalt mixtures. To achieve the objective, Linear Amplitude Sweep (LAS) and time sweep (TS) tests at controlled stress and controlled strain modes of loading were performed on the mixtures and the results were analyzed using the simplified viscoelastic continuum damage (S-VECD) methodology creating C vs S curves. Through the results and a failure criterion based on energy, cyclic tests can be simulated at different loading conditions for WÃhlerâs curves construction. The FAM dynamic moduli obtained are very similar to the frequency used to induce damage. However, for lower frequency, the intermediate FAM presented lower stiffness value. It was concluded that FAM with different MNS present very distinct damage performance, and the decrease in the MNS results in an increase in the mixture damage performance for low strain amplitudes. Also, the LAS testing protocol does not allow failure prediction due to the absence of failure in this type of test for FAM. / O Trincamento por fadiga à um dos problemas mais comuns em pavimentos flexÃveis no Brasil e no mundo. Danos por fadiga em misturas asfÃlticas se origina como um fenÃmeno de pequena escala. Recentemente, alguns autores tÃm utilizado a porÃÃo fina das misturas asfÃlticas, chamado Matriz de Agregados Finos (MAF), como um passo intermediÃrio para estimar o comportamento do Concreto AsfÃltico (CA). Por definiÃÃo, a MAF à composta de agregados finos, fÃler mineral e ligante asfÃltico. Hà muitas questÃes volumÃtricas neste tipo de mistura. No Brasil, a pesquisa desenvolvida sobre MAF usa peneira #10 (2,00mm) na preparaÃÃo de amostras, uma vez que a peneira #16 (mais utilizado na literatura) nÃo pertence ao conjunto de peneiras normatizadas nacionalmente. Recentemente, alguns autores tÃm utilizado o conceito de mecÃnica do dano contÃnuo em meios viscoelÃsticos, a fim de quantificar o dano durante o ensaio de fadiga. Nesta teoria as trincas sÃo consideradas homogÃneas dispersa no meio, e sÃo representados como variÃveis ​internas ​de estado determinadas experimentalmente. O objetivo deste estudo à avaliar a influÃncia da variaÃÃo de um dos constituintes da MAF, os agregados, variando seu tamanho mÃxima nominal (TMN), na resistÃncia ao dano destas misturas asfÃlticas. Para alcanÃar o objetivo, ensaios de Linear Amplitude Sweep (LAS) e Time Sweep (TS) à tensÃo controlada e à deformaÃÃo controlada foram realizadas nas misturas e os resultados foram analisados ​​usando o mÃtodo simplified viscoelastic continuum damage (S-VECD) para a criaÃÃo de curvas C vs S. AtravÃs dos resultados e um critÃrio de falha com base em energia, os testes cÃclicos podem ser simulados em diferentes condiÃÃes de carregamento para a construÃÃo da curva de WÃhler. Os mÃdulos dinÃmicos obtidos das MAFs sÃo muito semelhantes aos da frequÃncia utilizada para induzir danos. No entanto, para menor freqÃÃncia, a MAF intermediÃria apresentou menor valor de rigidez. Concluiu-se que as MAFs com diferentes TMNs apresentaram desempenhos de dano muito distintos, e à diminuiÃÃo do TMN da mistura resulta em um aumento da resistÃncia ao dano para baixas amplitudes de deformaÃÃo. AlÃm disso, o protocolo de teste LAS nÃo permite a previsÃo de falha, devido à ausÃncia de falha neste tipo de ensaio para MAF.