Damage Detection of Box Girder Based on Impact Load and the Improved Wavelet Packet-Permutation Entropy
摘要
Damage identification, both size and location, in any orthotropic steel bridge girders is important for timely maintenance and avoiding any catastrophic failure. Previous experimental studies have shown that the measured vibration acceleration data can be used to detect the damage of orthotropic plates of steel box girders when subjected to the sinusoidal excitation. In this paper, wavelet packet decomposition (WPD) is used to deal with the measured vibration acceleration response, and then the permutation entropy (PE) of the reconstructed signal is estimated. Then the damage index is defined according to the permutation entropy difference (PED) between the damage state and the health state, which is used to determine the actual location and relative size of the damage. Through finite element model of the case structure and vibration acceleration responses under different hammer loads are considered. This study demonstrates that the vibration response under different impact loads can also identify the severity and location of damage.