Semi-Markov Models for the Deterioration of Bridge Elements

Abstract

Many bridge management systems (BMSs) use a Markov chain model to forecast the deterioration process. The Markov property may be considered to be restrictive when modeling the deterioration of transportation assets, primarily because of the memoryless property and assumption of exponential distribution for sojourn times in the condition states. This study addresses some of the limitations that arise from the use of purely Markov chain deterioration models for transportation infrastructure by introducing alternative approaches that are based on the semi-Markov process. Two semi-Markov approaches for modeling the deterioration of certain bridge elements are developed. These are compared against a previously developed semi-Markov approach, the traditional Markov chain deterioration approach, and the change in the average actual condition indices of bridge elements that deteriorated for 8 years after being constructed. The results obtained from this study indicated that semi-Markov models are feasible and more flexible than the traditional Markov chain models when forecasting the future deterioration of transportation infrastructure. The semi-Markov models that were developed produced condition indices that were closer to the values of the average actual condition indices of bridge elements when compared against the values based on the traditional Markov chain model.