PSI - Issue 62

Gallipoli Maria Rosaria et al. / Procedia Structural Integrity 62 (2024) 940–945 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

941

2

1. Introduction We propose a combination of different experimental methods to analyze seismic ambient noise recordings in order to determine the main structural parameters of an infrastructure and their variation or change over time. The proposed approach could be used for Structural Health Monitoring (SHM). Some of the applied methods are well established in their application to infrastructures, others are still in their early stages, together they allow a more reliable interpretation of the results. An additional attractive aspect is the rapid and non-invasive nature of acquiring 20-30 minutes of seismic ambient noise data: A seismic array installed at the points of greatest structural interest is sufficient without diverting or blocking traffic flow or disrupting the operation of a road infrastructure, allowing on-demand monitoring. Four methods are used: Fourier Amplitude Spectra (FAS), Frequency Domain Decomposition (FDD), Ambient Noise Deconvolution Interferometry (ANDI), and NonPaDAn analysis. Currently, Operational Modal Analysis (OMA) is the most robust approach for experimentally estimating the modal parameters of a structure/infrastructure by recording only the natural and freely available excitation, i.e. the seismic ambient noise (wind, traffic, microtremors, etc.). The FDD, one of the OMA methods, has proven to be very accurate in detecting natural modes even for complex structures and very reliable in dynamic identification (Rainieri and Fabbrocino 2014). The FDD has been widely used to study the dynamic behavior of bridges (e.g. Meixedo et al. 2021, Zonno and Gentile 2021, Borlenghi et al. 2022, to name only the most recent articles) and buildings (e.g. Bindi et al. 2015; Petrovic et al. 2015; Karapetrou et al. 2016; García-Macías and Ubertini 2019). Although Ambient Noise Deconvolution Interferometry (ANDI) has already been used to study the dynamic behaviour of buildings (e.g. Petrovic et al. 2015; Bindi et al. 2015; Todorovska et al. 2020), cultural heritage (e.g. Lacanna et al. 2019; García-Macías and Ubertini 2019) and soil-foundation-structure interaction (Petrovic et al. 2018; Petrovic et al. 2019; Skłodowska et al. 2023), there are only a few applications to infrastructure. The integration of ANDI and FDD (OMA-based) has significant potential to monitor structural typologies other than buildings, such as bridges, and to identify their changes in structural characteristics due to ageing and damage (e.g. Garcia-Macias and Ubertini, 2021). Therefore, we believe that the results of these techniques analysing seismic ambient noise data allow the estimation of the main structural parameters (eigenfrequencies and equivalent damping, modal shape, wave propagation velocities) and interaction with the soil with sufficient reliability. We have applied this combination of different rapid and non-invasive experimental approaches to two different types of viaducts both located in the Basilicata region (Italy): the Gravina bridge on the SS.655-Bradanica, and the Monticello viaduct on the SS.407-Basentana (Albano di Lucania). 2. The Gravina bridge The “Gravina” is a recently built bow -string bridge in Matera (southern Italy), located on the Strada Statale 655 “Bradanica”. The foundations are located on the outcropping lithological units of the Bradanica, which rest directly on the calcarenites of Gravina Fm (upper Pliocene - lower Pleistocene). From a geomorphological point of view, the area is typical of a canyon about twenty meters deep. The bridge structure consists of a pair of inclined tubular steel arches with a diameter of 1.7 m, connected by nine circular beams. The deck consists of a steel-concrete composite profile, which extends over 144 m and has a total width of 18.80 m. The suspension system was realized according to the Langer method and consists of 19 pairs of stay-cables, with a fixed end on the arch side and adjustable on the deck end. The structure is isostatic on the outside and the restraint system is achieved by 8 elastomeric isolators: four of them are located under the arch bases, while the other are placed under the main girders (Fig. 1). Seismic ambient noise was recorded in two configurations: I (Fig. 1a, red dots) and II (Fig. 1a, green dots) under operational conditions. The monitoring instrumentation consisted of two sensor arrays with simultaneous acquisition: a) five 4.5 Hz velocimeters (SARA Electronic Instruments - SR04DA) with a sampling rate of 200 Hz on the bridge deck (Fig. 1b) and five velocimeters (Lunitek - Sentinel GEO) with a sampling rate of 250 Hz on the arch (Fig. 1b).