Performance prediction models for flexible and rigid pavements – state-of-the-practice review for implementation in North America
Yongsung Koh, Yujia Lu, Robert Wiggins, Y. S. Kim, Issam I. A. Qamhia, Erol Tutumluer, Jeb S. Tingle, Timothy A. Parsons, Michael J. Harrell
Abstract
The durability of pavement infrastructure depends on accurate distress and service life predictions. Empirical, mechanistic, and neural network models have been extensively developed to predict pavement distresses considering structural, traffic, and environmental factors, yet their accuracy varies due to inherent assumptions and diverse failure mechanisms. This paper presents a comprehensive review of pavement distress prediction models with an emphasis on selecting suitable mechanistic-empirical (M-E) models for integration into the Joint Evaluation and Design Integrated 2D (JEDI2D) pavement analysis platform developed by the U.S. Army Engineer Research and Development Center. The review critically evaluated existing predictive models for primary pavement distresses, including fatigue cracking, rutting, thermal cracking, roughness in flexible pavements, and transverse cracking, faulting, and roughness in rigid pavements, considering stabilised layers. Each model was assessed in terms of practical implementability, credibility, and complexity of required input parameters to identify the most suitable models associated with each distress type, such as the AASHTOWare Pavement ME models, which demonstrate robust validation, straightforward implementation, and established credibility. Furthermore, key research gaps were highlighted to guide future enhancements, emphasising the necessity of local calibration and advanced modelling techniques to improve accuracy and applicability for pavement management systems.