A critical review of the fatigue life prediction of asphalt mixtures and pavements

Abstract

New pavement construction techniques and the increased use of recycled materials have led to unexpected and premature pavement failure in recent years. The pavement's exposure to daily and seasonal extreme temperature and repeated vehicular loads accumulate damage. Pavement cracking occurs once the cumulative damage surpasses the material's cracking threshold. Fatigue crack is the most common pavement cracking type. Over the past four decades, researchers have carried out numerous experiments and analyses to understand pavement cracking. This paper aims to provide an overview of fatigue cracking and discuss various fatigue test methods for characterizing asphalt concrete mixtures. The article also discusses the most common phenomenological and mechanistic models for predicting the fatigue life of asphalt concrete pavements based on different fatigue test results. The paper details the implementation of the commonly used numerical models found in numerical simulation software and their prediction ability for the fatigue life of a pavement structure. Two major flaws in current evaluation methods are the sensitivity of experimental results and the lack of reliability of some predictive models. Multiscale asphalt material characterization is the ongoing practice for determining the most appropriate performance evaluation tool. However, proceeding with future research objectives is unrealistic until the accuracy of the tests and reliability of the predictions can be verified against actual field results. This critical review of the fatigue life predictions of asphalt mixtures and pavements should help to refine or redefine the right course of action for future research. • Summarize the fatigue cracking, fatigue test methods, constitutive models, and numerical simulation software programs. • Study the theory, application, advantages, and disadvantages of mechanistic and phenomenological models. • Comparing the three prediction approaches reveal that mechanistic models accurately rank pavement mixture performance according to actual pavement conditions.