Reliability-Based Mechanistic–Empirical Flexible Pavement Design

Abstract

It is well known that pavement design and construction involve many uncertainties and variations from specified design parameters. To address this issue, a reliability concept is often incorporated into the pavement design and analysis process. However, because of the complexity of modern mechanistic–empirical approaches, most of the existing reliability analyses deal only with some overall variability. A genuine reliability analysis considering the variability of specific input parameters is often computationally prohibitive. After a review of reliability approaches used in existing pavement design systems, Rosenblueth's 2n 1 1 method is recommended as a good candidate for modern mechanistic–empirical pavement designs. Implementation of Rosenblueth's 2n 1 1 method in the Texas Mechanistic–Empirical Flexible Pavement Design and Analysis System shows the simplicity and practicality of this approach. A pilot comparison demonstrates that this approach can make rational predictions considering the variability of different combinations of input parameters.