Pavement Rehabilitation Project Ranking Approach Using Probabilistic Long-Term Performance Indicators

Abstract

A probabilistic performance-based approach for generating a priority ranking of pavement rehabilitation project candidates is presented. The deployed probabilistic approach uses a discrete-time Markovian model to predict pavement conditions at the network level. The expected future distress rating associated with a particular pavement project is determined as the mean of a compound uniform probability density function derived from the corresponding future state probabilities. The expected future distress ratings for a particular pavement project are then used to construct the corresponding performance curve. The generated performance curves form the basis for developing an effective mechanism for prioritizing potential rehabilitation projects. A priority ranking system is presented with two alternatives. The first priority ranking alternative requires a fixed analysis period wherein project candidates are evaluated by using three different long-term performance indicators derived from the corresponding performance curves. The deployed performance indicators include the area under the performance curve, the average distress rating, and the terminal distress rating. The second priority ranking alternative requires a fixed terminal distress rating for the pavement network under consideration. The time required for each project to reach the terminal distress rating is determined as the rehabilitation scheduling time, which is used to establish project priority ranking. Pavement rehabilitation project candidates are then scheduled according to their priority rankings and available budget. The developed probabilistic approach is demonstrated by considering a project sample from the two-lane highway system in Ohio.