Abstract
Modal parameters, including natural frequencies, damping ratios, and mode shapes, are vital for evaluating the in-service condition of long-span bridges subjected to performance degradation or damages. Operational modal identification techniques requiring only ambient vibration responses are usually utilized to identify the modal parameters of in-service bridges. Due to the measurement incompleteness and modeling errors, uncertainties would inevitably be involved in the identification. The changing operational loads of long-span bridges, e.g. vehicles and winds, could also increase uncertainties. These uncertainty factors could decrease the precision of the operational modal identification and undermine the reliability of structural health monitoring (SHM) for long-span bridges. This study introduces an improved Bayesian modal identification approach using the scaled Fast Fourier Transform data for uncertainty quantification. The proposed method integrates the genetic optimization with the posterior probability density function of modal parameters to ensure the accuracy and robustness of the iteration process. Additionally, an asymptotic estimate interval with the assumption of a high signal-to-noise ratio is defined to improve the computational efficiency. The acceleration responses from a numerical example and a full-scale long-span bridge are adopted to validate the performance of the proposed method. Results show that the improved Bayesian approach can accurately identify the modal parameters and efficiently quantify the uncertainties. The method enhances the reliability of modal tracking for SHM of operational long-span bridges.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have