Markov Chain Modeling of Pavement Surfacing

Suitable performance models are needed for estimating medium- and long-term budgetary requirements to keep bituminous surfacing of the road network in serviceable condition. Long-term deterioration models are used for predicting both the proportions of the network that will require resurfacing and in what years. These predictions can then be used in estimating required budgets. Such models also can indicate the network level of service that can be achieved within various budget scenarios. The research reported here used subjective condition data to develop deterioration models for five types of pavement bituminous surfacing: chip seal, geotextile seal (chip seal with geotextile underlay), dense-graded asphalt, open-graded asphalt, and ultrathin asphalt. The condition data were collected through visual inspection surveys, which were conducted every 2 to 3 years by experienced personnel who followed detailed guidelines. Cracking, stone loss, texture loss, deformation, and patching data were collected for dense-graded, open-graded, and ultrathin asphalt. For chip seal and geotextile seal, binder condition and level data were also collected. The condition ratings of the distresses were then combined into a single measure, the surface inspection rating. Markov chains were used to develop deterioration models of surface inspection rating as a function of surface age for the five surfacing types. These models were validated, and the reasonableness of their predictions was assessed according to practical experience and engineering judgment. The models proved to be practicable because of ease of development and reasonableness of predictions.