Damage Identification For Bridge Based On Artifical Neural Network

The identification, location, and estimation problem of structural damage for bridge has been the popular research in recent years. The damage identification techniques have been widely used in aeronautical, civil,mechanical and nuclear structures. It is a complex technique constructed on the damage theory, sensor technology, signal analysis,computer science and artificial networks. Relative to the traditional technology of structural damage identification,this thesis is mainly involved in the theory and application of detection technique for bridge structure damage based on the neural network.Through analyzing the research findings of structural damage identification based on artificial neural networks and comparing the capabilities of BP networks, this thesis presents that bridge damage technique can be researched based on BP neural networks. Simultaneously, artificial networks using LM arithmetic is constructed and the process is given.The rates of the curvature change at the nodes are obtained by the theory analysis before or after the bridge structural damage and the rates of change contain the information suchs as the location and the degree of the damage. On the base of the theory, numerical simulation of structural damage is carried for a simply supported beam bridge and a continuous beam bridge respectively,and the rate of the curvature change for first mode is picked up and input parameter of BP network is constructed properly. The damage structure is detection by use of the trained neural networks,and the result indicates that good effect is obtained for the rates of the curvature change for first mode as input parameter,so it has certain applied value in the detection of structural damage.