Comparative Analysis of Performance Prediction Models for Flexible Pavements

Abstract

In the case of flexible pavements, efficient and realistic prediction models for the long-term process of their degradation are essential. The failure of flexible pavements due to cracking, surface deformation, disintegration, and surface defects is the result of the reactions of the pavement system to the complex mechanical, thermomechanical, or chemical loads of car-traffic systems, environmental systems, or internal elements of the pavement system itself. Because of the degradation process’ complexity, uncertainty, and variability with time, studies of the estimation of the pavement lifetime and the evaluation of pavement conditions should be conducted in a comprehensive and phenomenological manner. Pavement degradation exhibits considerable variation, which is associated with uncertainties and is dependent on intrinsic and extrinsic factors. Three comparative models of deterministic, probabilistic, and stochastic view are applied to pavement degradation data to demonstrate and compare them with respect to the uncertainties associated with the nature of pavement degradation. These three models have strengths and weaknesses in terms of usability, implementation, and the amount of information required. In particular, this study focuses on and illustrates the use of a gamma process model for the prediction of performance degradation and the lifetime of a pavement. The gamma process model is used to obtain the conditions and lifetime as probability distribution curves with time varying parameters and is suitable for determining the optimal replacement lifetime and percentile lifetime of the pavement.