Probability-based Optimal Maintenance Strategies For Network-Level Bridge

The aging and deterioration of existing bridge caused by severe service condition and deferred maintenance activities provided a great challenge for bridge management in China. It is important to develop a multi-criteria evaluation method for deteriorating bridge and to determine the prioritization of available bridge maintenance funds to keep acceptable level of performance throughout the whole life-cycle of a bridge network. In this article, a GA(genetic algorithm)-based optimization method in conjunction with reliability analysis is proposed to obtain optimal maintenance strategies for individual bridge and network-level bridge. The main research work can be summarized as follows:Firstly, the time-dependent factors such as the corrosion of reinforcement, degradation of concrete strength, increasing of vehicle loads are analyzed. The time-dependent reliability analysis based on JC method is proposed to predict the deterioration state of the structure in terms of ultimate limit state and serviceability limit state.Secondly, a GA-based optimization method integrated with the time-dependent reliability analysis is proposed to obtain optimal maintenance strategies for the existing bridges. The study involves the determination of target reliability level, time-dependent reliability analysis, bridge maintenance and retrofit activities, and solution of a GA-based multi-objective combinatorial optimization problem. The effectiveness of the integrated method for optimal maintenance strategies is verified through an investigation on a T-beam RC bridge with a span length of 16m.Thirdly, the paper presents a comprehensive mathematical model for a probability-based bridge network-level performance evaluation using network theory. The overall performance of the bridge network is evaluated by three criteria, i.e. network connectivity, user's satisfaction and performance of critical bridges.Fourthly, the study conducts the time-dependent reliability analysis for bridge network-level performance evaluation. A bridge network-level model for one part of ZhongShan region is established, and the overall performance of this bridge network is evaluated by the above-mentioned three criteria.Finally, the concept of the multi-objective combinatorial optimization and dynamic programming procedure are introduced.a GA-based optimization method in conjunction with time-dependent reliability analysis and bridge network-level performance evaluation criteria is proposed to obtain optimal maintenance strategies. The effectiveness of the bridge network-level optimal maintenance strategies is verified through an investigation on the bridge network of ZhongShan region in the service lifetime of 30 years.