Characteristic identification of long-span arch bridge in different states based on high-rate sensor dynamic monitoring

Global navigation satellite system (GNSS) and accelerometers are effective for real-time structural monitoring. However, background noise and low sampling rates restrict their performance in field measurements of realistic full-scale large structures, particularly impeding the widespread application of high-rate GNSS. Furthermore, few studies involve the characteristic investigation of long-span structures under different states using high-rate sensors. Therefore, this study innovatively employs high-rate GNSS and acceleration to accurately measure the dynamic characteristics of a realistic long-span arch bridge in damaged and after-repaired states based on a high-rate GNSS static test. An optimized variational mode decomposition (OVMD) method is developed. The core of OVMD is to adaptively achieve optimal decomposition of signals and accurate selection of information components using the designed relative root mean square error (RRMSE) and dual adaptive threshold criterion, which enables signal reconstruction while effectively separating noise. Subsequently, the random decrement technique (RDT) approach is applied to capture the structural dynamic characteristic variations from denoised data. The experimental results demonstrate that the OVMD method significantly reduces the noise and outperforms wavelet soft threshold (WST) and complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN). The failure probabilities of the repaired bridge calculated from different distributions derived from high-rate GNSS semi-static displacement are at low levels, with a maximum value of 0.0005. In the same state, the relative errors of frequencies captured from two high-rate denoised signals are less than 3.45%. Notably, the first-order frequency exhibits distinct variations before damage, during damage, and after repair. The damping ratio variation characteristics in different states are successfully estimated via the RDT and agree well with the stochastic subspace identification (SSI) method. High-rate sensor technologies, notably high-rate GNSS, provide promising solutions for assessing the changes in bridge characteristics.