Multi-Stage Damage Identification Method Based on Integrated Neural Network

During the operation of the bridge structure, various structural damages occur due to the sophisticated loads. Failure to detect and handle bridge structure damage opportunely may cause serious accidents. In general, the use of a single type of neural network to identify bridge structural damage has its limitations. Therefore, establish the finite element model of a three-span continuous beam bridge structure and analyze the dynamic characteristics. The process of damage identification is divided into three stages which could reduce the problem difficulty. The input parameters of the neural network damage identification are improved in stages, and several neural networks are integrated at each stage. The final identification results show that the method can effectively identify the damage of bridge structures.