Scanning vertical and flexural-torsional frequencies of thin-walled girder bridges using a two-axle vehicle and a signal enhancement approach

In research on using vehicles to extract bridge frequencies, single degree-of-freedom vehicles and twodimensional bridges cannot fully simulate vehicle-bridge interaction and bridge frequency extraction. Therefore, a method of using two-axle vehicles to extract the vertical and flexural-torsional frequencies of threedimensional (3D) bridges is proposed. First, the feasibility of this method is theoretically verified by analytical methods. Subsequently, a signal enhancement approach, combining successive variational mode decomposition (SVMD) and a designed window function, is proposed. SVMD performs modal decomposition on vehicle signals, while the window function reduces noise in vehicle signals and enhances bridge signals, resulting in a low-noise spectrum. The investigations indicate that bridge frequencies extracted by the proposed method exhibit a relative error of <5 %, which meets engineering requirements. Moreover, the method is insensitive to vehicle parameters and is not limited by two-axle vehicle types. Notably, vertical acceleration spectra of two-axle vehicles, filtered using the signal enhancement approach, can resist the effect of road roughness noise on bridge frequency identification. This study further advances the vehicle scanning method and offers a practical approach to bridge health monitoring.