Bridge Technical Condition Assessment, Prediction And System Development

Bridge deterioration in highway networks caused by heavy traffic and severe environment condition becomes more and more significant, which would lead to a high frequency of maintenances and repairs. The traditional management methods based on experience are proved to be more and more inefficient. Thus, many kinds of bridge management systems (BMSs) are developed to assist highway managers in making systematic and cost-effective decisions regarding the MR&R (maintenance, rehabilitation & replacement) of their bridge networks.The current BMS is not only expected to store bridge information, such as inventory information, inspection information, maintenance information, but also to assess the bridge condition and to predict the bridge performance, moreover, it is also expected to determine the optimal strategies within budgetary constraints for maintenance, rehabilitation, and replacement of the bridges. In this thesis, the main work is summaried as follows:(1) The framework of the BMS is built, and a regular database is also constructed to store the information of the bridges, such as inventory information, inspection information, maintenance information and so on. Moreover, bridge members, bridge distress are given detailed classification.(2) The condition evaluation module of BMS is studied and 3 evaluation models are presented. The first one is a computer evaluation method based on that recommended in"Code for Maintenance of Highway Bridges and Culverts"; For the second method, the bridge members are divided into series-connected members, paralell- connected members and additional members, then an evaluation model is built; In the third method,"group-decision"is introduced in AHP (Analytical Hierarchy Process), the expert knowledge is endued with a reliable weight to establish a group-decision improved AHP evaluation model.(3) Three performance prediction models, i.e. polynomial regression model, grey prediction model,and CBR(Case-based Reasoning) model are introduced, and some application examples for the above models are given. Due to the shortages of these models in performance prediction, Markov process prediction model is established in this thesis finally.(4) B/S (Browser/Server) mode is applied in system development, the interface and database for the system are also presented in the thesis.