Automated modal identification in operational conditions and its application to bridges

The increasing diffusion of long term dynamic monitoring systems for structural condition assessment is currently driving a strong interest towards automated procedures of output-only modal identification. Different approaches have been recently developed in the literature for this purpose, often based on Stochastic Subspace Identification (SSI) methods. Such procedures usually rely on heuristic decisional criteria, hence demanding for independent checks with validation purposes. The aim of this paper is to propose an automated modal identification procedure, belonging to the class of SSI techniques and based on the popular tool of clustering analysis, and to exemplify its application in the operational modal analysis of two bridges, with different levels of complexity from the system identification viewpoint: a historic iron arch bridge and a long-span footbridge. In order to address reliability and accuracy of the procedure, the modal estimates automatically extracted from the data recorded on the two bridges were compared to those independently obtained by using well-established manual techniques. The results demonstrated very good accuracy and robust performance of the developed procedure, also in the case of weakly excited and closely spaced modes, so that its application to process the data collected in continuous monitoring systems seems highly promising.