A Bridge Damage Detection Method Based on Vibration Sensor Network and Frequency-domain Paired T-test

This paper proposes a method for identifying damage in bridge structures by combining frequency-domain analysis with hypothesis testing. Vibration signals are continuously collected through a strategically deployed sensor network on the bridge. The method first transforms the vibration signals into the frequency domain, extracting characteristic frequencies from both the healthy and monitored states. The paired-sample t-test is then applied to determine whether the frequency changes are statistically significant, thereby identifying potential damage in the structure. This approach integrates the sensitivity of frequency characteristics with statistical objectivity, allowing for efficient damage detection without relying on complex modal analysis. Experimental results demonstrate that the proposed method performs effectively across a variety of damage scenarios, showing robustness and strong practical applicability. It is particularly suitable for long-term monitoring, periodic evaluation, and rapid post-event screening of bridge structures. Additionally, this method can be applied to stiffness-sensitive structures such as railway bridges, urban overpasses and light rail tracks, offering a quantitative assessment of their condition to ensure operational safety and stability.