Probabilistic Modeling of Performance-Related Pay Adjustment for In-Place Air-Void Contents of Asphalt Pavements

Abstract

There is a growing interest in development of performance-related specification for pavement construction. This paper presents the development of performance-related pay adjustments with probabilistic models of pavement performance, with application to in-place air-void contents of asphalt pavements. Quality-assurance data were collected from construction database and pavement-performance data were extracted from pavement-management system. Performance-related pay adjustments were formulated using lifecycle cost analysis considering two different scenarios of maintenance strategy and the variations of pavement overlay life. The results show that the Bayesian approach with Markov-chain Monte Carlo methods can capture unobserved variations in pavement-condition data and relate the quality measure of in-place air-void contents to the expected pavement life with high goodness of fit. The analysis results indicate that there may be significant variations in the model parameters for estimating the expected pavement life due to deviations in acceptance-quality characteristics. This implies that addressing the variations in pavement-performance modeling is a critical issue in deriving performance-related pay adjustments. The probabilistic modeling results facilitate considering the level of reliability in decision making of pay adjustments.