ABSTRACT
ABSTRACT Municipal sewer systems are deteriorating in both structure integrity and performance with increasing service age. In this paper, the degradation of sewer structure condition is first discretized into five states based on Manual for Sewer Condition Classification (MSCC) developed by Water Research Center (WRc) in Great Britain; then a time-homogeneous Markov Chain model is developed to describe the transition between two successive condition states where the transition is represented by the transition intensity (also called transition rate). A confident determination of the transition intensities (model parameters) relies heavily upon the availability of a broad database of sewer pipe condition data. For small to medium municipalities, the condition data are typically limited, which has become one of the main challenges in development of municipality-specific deterioration models. To tackle this issue of data shortage, an empirical Bayesian updating approach is adopted for estimation of the intensities. The Bayesian approach mainly follows two steps: first, pool all condition data that are available to form a general database which is then used to determine the prior of the intensities using the maximum likelihood method; next, the Bayesian inference approach is utilized to update the prior intensities using specific condition data from the municipality of interest. The obtained posterior intensities are thus municipality-specific. A case study is illustrated in the end to elaborate the implementation of the Bayesian approach and the application of the developed Markov Chain model in prioritization and optimization of wastewater linear assets within the framework of risk-informed life-cycle cost management.
