Structural Damage Localization of Steel Arch Bridge Based on Wavelet Packet Transform

In order to solve the damage identification problem of steel arch bridge, a damage detection method based on wavelet packet is presented in this paper. A vertical impact load as excitation source is exerted on the surface of the arch bridge. Thus, the acceleration responses can be directly measured, and applied to identify the damage locations. The acceleration responses of damaged and undamaged arch bridges will be converted to node energy by using wavelet packet transform. We will contrast the changes of node energy of each band before and after damage occurs, and calculate the sum of absolute energy differences of all frequency bands. Thus, a damage index can be created, and through the peak method the suspected damage locations can be inferred preliminarily. Then, Shannon entropy is proposed as cost function to optimize the level number of wavelet packet decomposition. Two-dimensional and three-dimensional truss finite element models of arch bridge are applied to verify the performance of the proposed damage detection method. The simulation results show that the method can identify damage locations. Therefore, the proposed method is effective for the damage identification of steel arch bridge.