Stochastic Modeling of Pavement Performance

Abstract

Pavement performance deterioration cannot be predicted precisely because traffic and environmental actions as well as the material properties and geometric variables of pavement systems are uncertain. Therefore, the prediction of the pavement performance should be carried out based on a probabilistic framework. A simple probabilistic approach is developed in this study for predicting pavement performance. The approach is based on a nonhomogenous continuous Markov chain. Its use in conjunction with the flexible pavement deterioration models in the Ontario Pavement Analysis of Cost (OPAC) and in the AASHTO guide is explored. The proposed approach is more efficient than the ones found in the literature since the probability transition matrix in this study depends only on two model parameters, one controlling the intensity of transition and the other controlling the time transformation. The proposed approach seems able to mimic well the pavement degradation process predicted by the OPAC and AASHTO models.

Keywords:

• Pavement
• Deterioration
• Reliability
• Serviceability
• Stochastic process

Acknowledgements

The financial support of the Natural Science and Engineering Research Council of Canada is gratefully acknowledged. We thank the anonymous reviewers for their constructive comments.