Use of Markov Chain Model Based on Actual Repair Status to Predict Bridge Deterioration in Shanghai, China

Abstract

Bridge condition prediction is crucial in preparing conservation budgets for the maintenance of bridges. A bridge management system has been formally used and promoted for urban bridge management in Shanghai, China, since 2004, and 16,623 bridge records have been accumulated. Although there are many data records, predicting bridge deterioration precisely is difficult because the data composition is complicated and the maintenance history is varied. Therefore a Markov chain model was applied as a decision aid to consider the different conservation strategies. More than 66,000 data records were used to calibrate the model. The modeling considered two conservation regimes: ( a) routine maintenance and minor repair and ( b) medium and major repair. The repair rate was obtained through an actual conservation survey. The influence of spatial distribution was also considered. Bridge conservation efforts were uneven at the city level. The condition of bridges in the central city is much better than that of those in suburban areas, although the proportion (55.6%, 2014) of suburban bridges is larger. Furthermore, based on the present status of bridges, conservation efforts have been insufficient generally and even worse in suburban areas. The medium and major repair levels have had a significant impact on deck systems and superstructure but quite a small impact on substructure and the whole bridge. Thus the present conservation efforts cannot improve the overall bridge condition fundamentally. As a result, the condition of bridges in Shanghai has deteriorated rapidly.