A Comparative Study on the Estimation of Bridges Modal Parameters Through Different OMA Techniques for Structural Health Monitoring

In the context of Structural Health Monitoring (SHM), Operational Modal Analysis (OMA) offers methods for estimating the dynamic characteristics of structures based on vibration responses under operational conditions. However, the performance of each technique depends on the specific modal properties investigated and the parameters set in the identification procedure. This study evaluates the performance of various OMA techniques for estimating dynamic characteristics of structures under operational conditions. The techniques assessed include Frequency Domain Decomposition (FDD), Enhanced Frequency Domain Decomposition (EFDD), Stochastic Subspace Identification in covariance-driven and data-driven formats (SSI-COV, SSI-DATA), and PolyMAX. These methods were implemented in Python and applied to two full-scale bridge structures: a road overpass with a Niagara-type Gerber scheme subject to free vibrations and a 217-m-long, five-span viaduct with Gerber half-joints under standard operational conditions. The measurement systems used in the two cases differ, contributing to a general comparison considering the different sensor configurations. The study compares the effectiveness and robustness of these OMA techniques across different structural scenarios, focusing on various aspects: the ability to correctly identify all principal modes of vibration, comparisons in terms of frequency values and damping ratios, as well as the computational effort required by each technique to achieve the same level of accuracy in the results. This research aims to compare different OMA techniques from various perspectives by applying them to different structures with varying measurement systems.