There are a number of components of pavement engineering, including pavement management, pavement analysis and design, and pavement materials. Historically, the field of pavement management has been interested in monitoring post-construction condition, timing of preventive maintenance and rehabilitation treatments, and economic analysis of alternatives. On the other hand, the field of pavement analysis and design has dealt with optimizing pavement structure; with optimum structure, a pavement system is expected to survive during its service life for given traffic and climate conditions. The performance of pavement materials has been improved to achieve the long-lasting and lower-maintenance pavement systems. A data-driven comprehensive approach considering all aspects of pavement engineering together could be a future direction for advancing pavement engineering practices.

In order to achieve a data-driven comprehensive approach considering all aspects of pavement engineering together as outlined above, a data-driven and efficient pavement design, analysis and management concept has been proposed in this study. To serve as elements of this concept, several models related to pavement structural response models, pavement performance prediction models, and pavement remaining service life (RSL) models have been developed. First, to enable faster three-dimensional finite element (3D-FE) computations of design stresses, artificial neural network (ANN)-based surrogate computational pavement structural response models were developed. These models produce an estimate of the top-down bending stress close to that computed by 3D-FE analysis in rigid airport pavements in a fraction of the time. Second, longitudinal cracking mechanisms of widened jointed plain concrete pavements (JPCP) were demonstrated and their longitudinal cracking potential was evaluated using numerical analysis. Third, the Federal Aviation Administration’s (FAA) current rigid airfield pavement design methodology has been evaluated in great detail to better identify research gaps and remaining needs with respect to cracking failure models so that recommendations could be made as to how current methodology could be improved to accommodate top-down and bottom-up cracking failure modes. Fourth, a detailed step-by-step methodology for the development of a framework for pavement performance and RSL prediction models was explained using real pavement performance data obtained from the Iowa Department of Transportation (DOT)’s Pavement Management Information System (PMIS) database.