Condition Evaluation of Cable-stayed Bridge Based on Model Updating with Neural Network

摘要

In this paper, a method based on finite element model (FEM) updating with neural network is proposed to evaluate the actual state of bridges in operation period. The modal flexibility is introduced as one of the input parameter of neural network and the model updating results indicate the validity of this parameter. After simulating various load combinations on the updated model, the state of bridge is assessed by capacity analysis. Finally, the example of a prestressed concrete cable-stayed bridge indicates that the updated model can reflect the bridge actual state and this evaluation method of updating model with neural network is feasible.