Abstract
In many practices of the bridge asset management, the life cycle costs are estimated by statistical deterioration prediction models based upon human inspection data. In many applications, it is, however, often the case that the validity of statistical deterioration prediction models is flawed by inadequate stocks of inspection date. In this paper, a systematic methodology is presented to provide the estimates of the deterioration process with bridge managers based upon the empirical judgments by the experts at the early stages, and revised them as the new data are obtained through inspections. More concretely, the Bayesian estimation methodology is presented to improve the Markov deterioration hazard model by the Markov Chain Monte Carlo methods. The paper is concluded by illustrating the application examples of the methodology presented in this paper applied to the data set in the real field.
