High-Resolution Time-Frequency Representation of Bridge Dynamics Using Hybrid RLMD-SET Framework: Implications for Energy Harvester Optimization

The efficient design of bridge vibration energy harvesters (BVEHs) depends strongly on accurate, high-resolution characterization of structural vibrations under real-world conditions. This study proposes a novel hybrid time-frequency analysis framework that integrates Robust Local Mean Decomposition (RLMD) with the Synchro-Extracting Transform (SET) to enhance the interpretability and precision of bridge vibration signal analysis. RLMD is employed to adaptively decompose raw vibration signals into product functions that represent both external excitations and intrinsic structural modes. SET is then applied to these components to extract detailed time-frequency features with high localization and resolution. The performance of the proposed RLMD-SET framework is evaluated through both numerical simulations and field experiments. A coupled vehicle-bridge dynamic model is developed to simulate vibration responses under controlled loading scenarios. Additionally, field measurements are conducted using interferometric radar on a 32-meter simply supported beam bridge to validate the method under operational conditions. The results demonstrate that the proposed approach accurately identifies and distinguishes vibration events, revealing spatial and temporal characteristics of structural responses with high fidelity. More importantly, it captures the distribution of vibration energy and its relationship to harvesting potential, offering insights for the optimal placement and tuning of BVEHs.