Study Of Steel Beam Damage Identification Based On Neural Network

As the rapid development of our economy and the improvement of science and technology, our transportation industry is also in rapid development. As an important branch of highway construction, bridge construction is also in constant development. In recent years, every kind of bridge is emerging all over the country. However, due to the oldness, the overload and the deterioration of the natural environment and other factors, many bridges have various degrees of damage. These external damage or the damage hides in the interior can caused security risks. Many accidents of bridges not only caused huge economic losses, but also endanger the personal safety of people seriously. Monitoring the bridge structure, diagnosing the damage of the structure and repairing the damage will be very important for maintenance of bridge.In this paper, modal analysis and neural network are used to analysis the structure on the basis of domestic and foreign methods of structure health diagnosis. First using the large-scale analysis software ANSYS to do the simulation analysis and get the modal parameters of the structure, then using MATLAB to analysis these datas and construct a neural network. Finally, using the trained neural network to detect the datas of the experiment and compare the results. The results show that the method of using modal analysis and neural network can realize the detection of structural damage location and damage degree.The background and significance of the research on the damage identification of the bridge is introduced.At the same time, the progress of research work of damage identification technology is also introduced.The basic principle of modal analysis is introduced. The using of several modal parameters in damage identification is discussed and then comparing the accuracy of these parameters.The theory and analysis method of the software based on finite element ANSYS are introduced. The principle of neural networks is introduced, and the pros and cons of the application of neural network in damage detection are analyzed.Using the simulation software ANSYS to model and analysis the structure of nondestructive and different proportion of damage. The modal datas are combined as input of neural network to train the network. Then using the trained neural network to identify the damage location and damage degree.The main research work of this paper can be summarized as:regarding the curvature mode of specified points and frequency as the input parameters of the BP neural network can point the location accurately and determine the degree with a high precision.