Identifying the Bridge Natural Frequency Pattern Under Operational Condition

Abstract

This study investigates variations in the Humber Bridge’s natural frequency and presents waveforms of individual vibration modes using acceleration response data collected on February 1, 2012. Three synchrosqueezed transform analysis methods were primarily employed to process the numerical multi-component and closely-spaced signal to identify the most suitable techniques for real-world applications. The improved multisynchrosqueezing transform and Fourier synchrosuqeezed transform were chosen for analyzing the bridge’s monitored vertical acceleration at Humber. The time-frequency representations and instantaneous frequencies (IFs) generated by these methods are well-aligned, revealing less than a 5% change in the bridge’s vertical natural frequency, with mode 1 exhibiting an 11% variation. The monitored vertical acceleration was then reconstructed, and the results demonstrated the high accuracy of the recovered data, with an L-Infinite norm of 0.0013. Finally, the reconstructed waveforms of each mode, along with the corresponding IFs and spectra, provided a convenient means for further insight into the acceleration data. This study presents new findings, such as the waveform of each vibration mode from the monitored acceleration of the bridge, previously unreported in bridge engineering, and validates the feasibility of cutting-edge synchrosqueezed transform methods for accurately tracking the bridge’s frequency variations under operational conditions. It lays a foundation for further analysis of the bridge’s monitored signals.