An enhanced stochastic subspace identification incorporating time variables and its application to large-span sea-crossing suspension bridges

Abstract

Structural health monitoring and assessment are essential to large-span sea-crossing bridges, and modal identification is the most effective method for achieving them, with the stochastic subspace identification (SSI) method dominating. However, when analyzed in open traffic scenarios with SSI for a large-span sea-crossing bridge, the contour of the vibration mode is distorted. Related affecting factors such as the measuring conditions, user-defined parameters, and traffic are explored, and traffic is verified as the essential component. To address this issue, an enhanced SSI algorithm incorporating time variables is created to automatically identify and rectify vibration mode shapes. Feature heterotopic signals are first matched with a common reference signal to perform modal identification. Advanced particle swarm optimization is applied to estimate critical parameters. The SSI algorithm with the density-based spatial clustering of applications with noise (DBSCAN) algorithm is then utilized to extract the modal characteristics. Both simulated misalignment and real records are applied to verify the suggested approach's performance. Comparative analysis with state-of-the-art techniques, such as the error norm, cross correlation, phase transform-β, mean phase deviation, and genetic algorithm, shows that the recommended approach is dependable and successful at identifying the large-span sea-crossing bridge's vibration mode shapes and computationally more accurate and efficient.