DAMAGE DETECTION OF CONCRETE BRIDGE WITH WAVELET PACKET DECOMPOSITION AND SUPPORT VECTOR MACHINE
Assessing the health condition and detecting possible damage of a bridge are important issues for bridge engineers. A vibration-based technique combining the wavelet packet decomposition (WPD) and the support vector machine (SVM) is proposed for identifying and location damage in bridges. Assuming that the bridge is under broad-band ambient excitation, the WPD is first applied to the measured acceleration responses to obtain energy spectrum at different frequency bands. The wavelet packet energy spectrum is sensitive to the location and the severity of damage and hence can be used as feature vectors. For damage pattern classification, the SVM has a high accuracy and excellent generalization capability. This method is numerically demonstrated on a three-span highway bridge. Results show that the combined WPD and SVM technique can correctly identify both the location and the severity of damage occurs in the bridge.
Chih-Chen Chang Hao-Xiang He
Department of Civil Engineering, Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, P.R. China
国际会议
长沙
英文
502-508
2007-11-19（万方平台首次上网日期，不代表论文的发表时间）