Hierarchical Markov Chain Monte Carlo Simulation for Modeling Transverse Cracks in Highway Pavements

Abstract

Transverse cracks are distresses whose initiation and propagation affect the stability and structural integrity of the highway pavement. They are caused by load, moisture, temperature, construction defects, or a combination of these. Cracking leads to inherent pavement defects that are exacerbated by moisture infiltration as well as the formation of other pavement distresses like roughness with their attendant problems to roadway agencies and users. One feature of transverse cracks is the spontaneity of their initiation and the uncertainty of their propagation. As a result, modeling the formation and spread of transverse cracks requires a framework within which uncertainty can be expressed explicitly. This research seeks to use a statistical approach to model the propagation of transverse cracks on road pavements in Delaware. The aim is to investigate how hierarchical Markov-chain Monte Carlo (MCMC) simulation performs in estimating and predicting the spread of transverse cracks without neglecting their associated uncertainty. Hierarchical MCMC models use the Bayesian approach, which accounts for uncertainty in pavement distresses.