Identifying damage in long-span self-anchored suspension bridges using characteristic indices

Abstract

This study presents a methodology for detecting damage in bridge structures using three damage identification indices: the deviation of displacement influence line (DDIL), the deviation curvature of displacement influence line (DCDIL), and the relative rate of change of wavelet packet energy (RES). The proposed approach is based on the principles of displacement influence line and wavelet packet transform. To demonstrate the effectiveness of the three metrics in determining the location of structural damage, a finite element model of the main bridge in Zhengzhou Taohuayu Yellow River Bridge, a three-span twin-tower all-steel girder self-anchored suspension bridge, is evaluated. The study investigates the impact of various factors, such as damage degree, damage location, signal noise, variable speed vehicles, and random traffic flow, on the usefulness of the metrics. The results indicate that all three indices are effective in detecting damage in bridge structures. The DCDIL demonstrates a distinct, abrupt change at the location of damage, while the DDIL is somewhat less effective. Additionally, the RES curve is particularly sensitive to local damage and can identify damage locations far from the bridge span. However, the study notes that the accuracy and reliability of the RES identification indicators decrease non-linearly in the presence of signal noise. The highest values of the RES curve are positively correlated with acceleration and less closely related to uniform acceleration. A vehicle speed of 20 m/s is recommended for identifying the location of bridge structure damage. Finally, the study suggests that random, unevenly spaced traffic enhances vibration and can more precisely localize the damage location and degree of the bridge structure.