Advanced comparative modal analysis using stochastics in objective functions for bridge damage detection

Abstract Structural monitoring is crucial for ensuring the safety and longevity of bridges. Among the available non-destructive methods, finite element model updating (FEMU) stands out as a powerful tool for damage detection. This paper introduces a comparative reliability analysis aimed at identifying the most effective objective function for FEMU-based damage detection in bridges. The study evaluates three objective functions (frequency, modal shape, flexibility) within a genetic algorithm framework to optimize the match between a damaged model and its calibrated counterpart. To this end, advanced analytical and stochastic simulations were carried out to connect dynamic and optimization analyses. To validate the approach, both the stochastic and analytical models were examined using a real case study. The results showed that modal flexibility, which combines mode shapes and natural frequencies, achieved the highest reliability, outperforming the other objective functions in about 85.0% of the tested cases across all scenarios. A major advantage of this approach is that the required dynamic properties could be obtained by monitoring the bridge in situ, allowing the results to be used in FEMU-based assessments to support the maintenance of bridges and other infrastructure systems.