Bridge Optimal Maintenance Using the Multi-attribute Utility Assessment Method

With the rapid development of highway networks, the bridges not only need to meet the safety requirement but also social and economic development goals.In order to plan maintenance strategies to maintain the maximum performance of the bridge during its service life, it is necessary to develop a methodology that evaluates the comprehensive performance of the bridge effectively.In this paper, the multi-attribute utility incorporating the fuzzy evaluation method (MAUFE) is proposed to assess the bridge performance.Furthermore, the Markov model and LSTM algorithm are utilized to build a bridge degradation prediction model.The optimal model is established to maximize the MAUFE results.Various performance indicators are integrated by using a multi-attribute utility function.The bridge condition, life-cycle cost, and sustainability index are selected to assess the comprehensive performance of the bridge.The fusion weight of the performance index is determined by combining the analytical hierarchy process and entropy weight method (AHP-EW).The fuzzy comprehensive evaluation method is adopted to construct the fuzzy membership matrix of different indicators, and the comprehensive performance of bridges is calculated.A data-driven framework is used to predict the bridge's performance based on actual observation data.The mathematical model of bridge life-cycle maintenance decision-making under specified objectives and constraints is established, which is solved by a genetic algorithm to obtain an optimal maintenance strategy.Finally, the representative bridges are selected to assess the life-cycle comprehensive performance.And the maintenance strategies over thirty years of a particular bridge under different evaluation indicators as objectives and constraints are planned by using the proposed optimal model.The results show that the assessment method can fully reflect the decision-makers attitudes and cover many aspects within the performance assessment process.The prediction model is capable of accurately predicting the bridge's health status.The maintenance schedule that makes each evaluation indicator better can be obtained with the objective of comprehensive assessment model.The maintenance decision-making model can maximize the comprehensive performance of the bridge within the limited budget.