Research On The Bridge Damage Detection Method Based On Modal Analysis And BP Neural Network

Recently our civil traffic projects develop rapidly; plenty of new and oversize bridges have been established. But the bridges will gradually wear down under the circumstances of the natural environment and applied environment. This will causes substantial loss of people's life and their property. So analyzing the technology of bridge damage is very important.Based on researching at home and abroad bridge damage detection theory and BP neural network, the article making use of the large-scale finite element analysis software ANSYS and the BP neural network toolbox function in MATLAB the vibration modal analysis theory and BP neural network are applied to the bridge damage identification. Through a supported beam is given damage identification, damage bridge and the calibration of damage degree we give the method based on modal analysis and BP neural network.The first, several bridge damage identification methods based on vibration modal analysis are discussed in detail including the theory, method and the applications based on modal frequency, modal vibration shape difference, curvature modal shape, strain modal shape, flexibility difference, flexibility curvature, stiffness matrix, genetic algorithms, neural network, and so on. Through analysis we take advantage of the change ratio of the modal frequencies as the method to achieve the damage location of the bridge.The second, BP neural networks, BP classic algorithm and improved BP algorithm are introduced. And the principle based on BP neural networks is also introduced. I also give the reader how use the BP neural network toolbox functions in MATLAB to achieve the bridge damage identification.The third, Use the large-scale finite element analysis software ANSYS and the BP neural network tool functions in MATLAB to achieve the damage identification, damage location and the calibration of the damage degree of a classic simply supported beam. Numerical simulation and analysis showed that the theory based on the vibration modal analysis and BP neural network is very effective. This will provide methods and means for structural health system better applied to the actual.