Dynamic strain-based monitoring of a historical Vierendeel truss bridge under changing environmental and support conditions
摘要
Arguably the most widely researched structural health monitoring method for condition assessment and damage detection of civil structures is vibration-based monitoring or VBM. Primarily natural frequencies but also damping ratios and displacement mode shapes are in the core of the majority of the developed methods. However, direct detection of small-scale damage can be cumbersome, given the high sensitivity of natural frequencies to temperature, the difficulty in accurately identifying damping ratios and the complexity of employing dense grids of accelerometers to obtain detailed displacement mode shapes. Strain mode shapes are highly attractive features for VBM as they have demonstrated in several recent case studies a high sensitivity to early-stage damage and a relative insensitivity to temperature, while they can be obtained in dense grids when multiplexed fiber-optic sensors are employed. Further research is required though, as the influence of other sources of variability on strain mode shapes is not yet known. In the present work, a historical Vierendeel truss bridge is monitored for more than a year with 64 fiber-Bragg gratings and its modal characteristics are automatically identified through a hierarchical clustering algorithm. The bridge is known to be sensitive to brittle damage due to the ductile-to-brittle transition of its structural steel for temperatures close to 0