Health Monitoring System And Intelligent Diagnosis Methods Of A Long-Span Cable-Stayed Hybird Girder Bridge

With the expeditious development of our civil traffic projects, plenty of oversize and novel bridges have constantly rushed, the number of new and old bridges has been increased. To keep our people in a safe condition upon their lives and treasure, immediate and effective detection towards the positions and extent of damage on bridges and get an immediate idea of the its'health condition in daily use has been brought as a problem into the researchers'focus.The Hebei Street cable-stayed hybird girder bridge in Tianjin was investigated in the dissertation. Take into account of its structural characteristic, the bridge structural health monitoring systerm is constituded and a lot of efforts have been put on the study of the subsystem of health diagnosis. According to the theory of hierarchical damage identification, the course of health diagnosis is supposed to be divided into two steps, i.e. the substructural damage identification and the substructural components damage identification. The damage identification of the bridge in whole span is simulated, and an intelligent diagnosis method for damage identification is proposed in substructural hierarchical damage identification by using the combination of the Probablistic Neural Network (PNN) and the Radial Basis Function (RBF) Neural Network. The first step is to detect the damaged substructure using the PNN, and the second step is to detect the damaged substructural components with the RBF Neural Network. Especially in the second step, an assembled static and dynamic damage sensitive index is presented according to the structural characteristic and the structural response indices including the available dynamic and static response indices. Then, a model of RBF Neural Network is constituted and is used to simulate three damage conditions, i.e. single damage and double or three damages which occured simultaneously.The results indicate that the intelligent diagnosis method combining the PNN and the RBF Neural Network can meet the requirement of health diagnosis subsystem for damage identification with great efficiency. The assembled static and dynamic index shows a sensitive performance during identification of the cable-stayed hybird girder bridge, and it is suitable for the process of damage diagnosis of long-span hybird girder bridge structure. In short, this research could give some useful references towards the constitution of health diagnoses system of long-span complex bridges.