Details

Autor(en) / Beteiligte

• Chioccarelli, Eugenio
• Giunta, Daniele
• De Angelis, Alessandra
• Bilotta, Antonio

Titel

Discussing optimal sensor placement for ambient vibration test of an existing bridge

Ist Teil von

• Engineering structures, 2024-07, Vol.310, p.118000, Article 118000

Ort / Verlag

Elsevier Ltd

Erscheinungsjahr

2024

Quelle

Access via ScienceDirect (Elsevier)

Beschreibungen/Notizen

• The operational modal analysis (OMA) allows the identification of the dynamic (i.e., modal) characteristics of an existing structure, analyzing the data recorded by sensors, usually ad-hoc positioned on it. The number of sensors and their locations affect the accuracy of the dynamic identification results: as the number of sensors increases, the quality of the results improves, but the installation and operational costs of the monitoring system also increase. The methods of optimal sensor placement (OSP) were developed in literature to identify a set of sensor locations, from a larger candidate set, to perform OMA without losing information. These methods are based on numerical analysis of a finite element model that surrogates the structure to be monitored before installing any instrument. Among the OSP techniques, the effective independence (EFI) method is one of the best performing according to literature; it associates a ranking coefficient to each possible sensor location as a function of its usefulness in detecting the dynamic properties of the structure. The aim of this work is to highlight pros and cons of the OSP technique analyzing the results of an actual ambient vibration test for dynamic identification of an existing prestressed concrete viaduct. The adopted sensor network was designed starting from an optimization algorithm based on the EFI method, and the recorded signals, freely available, were used for both the dynamic identification of the structure and the experimental validation of the optimized sensor network. Results show that, for the analyzed structure, the optimized sensor network, constituted by four accelerometers, allows the identification of the same natural frequencies of the structure identified by a larger network of thirty accelerometers adopted as benchmark. Moreover, the optimized network shows the best performance in terms of frequency identification with respect to several alternative layouts characterized by the same number of instruments in different positions.