Evaluation of pavement performance using a probabilistic design approach with application based on accelerated pavement testing data

A problem frequently encountered in pavement design is the uncertainty and variability of certain design parameters. Empirical safety factors are generally employed to account for the uncertainties involved, but this can lead to significant design errors. This work proposes an improved probabilistic design approach to the design of asphalt pavements, based on the principles of a deterministic method, but taking into account the variability of different design parameters. In this work, the methodology is validated using full-scale experimental data from an Accelerated Pavement Testing (APT) facility. By leveraging results from controlled, full-scale loading, this study demonstrates that APT-derived variability serves as a realistic proxy for real-world field stochasticity. The proposed approach successfully predicts the primary stages of pavement deterioration, providing a robust bridge between controlled experimental results and mechanistic-empirical design reliability. These findings offer a validated methodology for improving pavement design methods and advancing the practical application of probabilistic engineering.