Database Support And Modeling For The Mechanistic Empirical Pavement Design Guide (Mepdg).

The Mechanistic Empirical Pavement Design Guide (MEPDG) procedure is substantially more complex than the 1993 AASHTO Guide and it requires significantly more inputs from the designer, some of which are either not available or are stored in locations not familiar to designers. Many data sets need to be pre-processed before they can be used in MEPDG. Based on the study of all the necessary inputs and analysis parameters required in MEPDG, a comprehensive centralized supporting database for MEPDG is developed to provide pavement designers a tool to pre-process and store the climate, traffic, material, performance, construction, and maintenance data for the state of Arkansas. An external data interface is produced to automatically generate climate, traffic, and material data files that can be directly imported to the MEPDG software. The traffic module of the software interface provides a useful tool to automatically preprocess and import the raw Weigh-In-Motion (WIM) data, check the quality of the original traffic data, and generate the traffic inputs in accordance with the file format requirements in the MEPDG software. In addition, the most significant influencing material inputs, such as the Dynamic Modulus (E*) for flexible pavement and Coefficient of Thermal Expansion (CTE) for PCC pavement, can be retrieved from historic material testing data in the populated database to aid pavement designers. Pavement performance prediction models are one of the most important components in the MEPDG design process. In MEPDG, multivariate regression techniques are adopted to develop the International Roughness Index (IRI) based smoothness prediction models. However, because pavement performance involves complex relationships, some of which don't have analytical solutions, this statistical approach may have limitations. In this research, gray theory, fuzzy gray theory, and Bayesian based prediction techniques are employed to develop alternative models for pavement smoothness with partial known information and under uncertainties, as well as to integrate engineering prior judgments into smoothness prediction. Pavement condition evaluation is crucial for the rehabilitation design procedure in MEPDG to determine priorities for the selection of project strategies. The MEPDG guide gives a generic description of pavement evaluation rather than producing a quantitative overall evaluation index, which is more suitable for project level evaluation. In this dissertation, Extension theory, and gray clustering based approaches are adopted to develop overall pavement condition evaluation procedures following the MEPDG framework so as to aid pavement management and rehabilitation strategies at network level for MEPDG. It is anticipated that the research efforts in this dissertation will improve the management and accessibility of the MEPDG input data and initiate a starting point for the calibration and implementation of the MEPDG design procedure.