Application of the Bayesian statistical approach to develop a Stone Mastic Asphalt (SMA) pavement performance model

Abstract

Stone mastic asphalt (SMA) has not been widely used in the pavement industry, and there are no detailed design specifications for this type of asphalt. Therefore, long-term performance data is not available widely, and no performance model has been developed for SMA. The main purpose of this study was to integrate expert knowledge (using the Markov-chain process) and experimental data from field investigations to propose a performance model for SMA through the incorporation of the Bayesian technique. The combination of these sources of data resulted in an efficient and effective method to develop a performance model for this type of pavement, which did not have a long-term performance database. As a result, a robust linear performance model was established to predict the service life of SMA. The service life of SMA can be estimated explicitly according to the developed performance model which has been validated using a new set of data.