PSI - Issue 44

Davide Arezzo et al. / Procedia Structural Integrity 44 (2023) 2098–2105 D. Arezzo et al./ Structural Integrity Procedia 00 (2022) 000 – 000

2101

4

3. Preliminary dynamic identification tests A good comprehension of the building's global dynamics is the first step in the design and development of a monitoring system. At this stage, Ambient Vibration Tests (AVTs) and Operational Modal Analysis (OMA) have a key role. Measurements on the church were carried out using 16 single-axis piezoelectric accelerometers located in 8 points around the ellipse of the tiburium at two different levels in order to capture the global dynamic behavior of the church. As can be seen in Figure 3a and b, two uniaxial accelerometers were placed at each point, from A to H, with measurement directions along the main axes of the church hall. As for the measurement chain, PCB 393B31 accelerometers with a sensitivity of 10V/g, NI 9234 analogue-to-digital conversion modules, a cRIO 9045 and three cDAQ 9185 were used. A distributed sensor network was created by placing one measuring station cRIO near reference point 1B and three mobile cDAQs near the other measuring points. Synchronization was achieved using Time Sensitive Networking (TSN) technology, which provides distributed time synchronization and deterministic communication using standard Ethernet networks. Once the acceleration measurements were performed, the dynamic identification was carried out using the Principal Component - Stochastic Subspace Identification (SSI-PC) technique. The measurements at the two levels were performed at different times and therefore, being asynchronous, the modal shapes needed to be scaled; in this case, the global results were obtained by applying the PoSER (Post Separate Estimation Re-scaling) technique. In Figure 3c, the stabilization diagram and the damping frequency diagram are shown with the stable solutions highlighted by vertical alignments. Overall, 10 vibration modes were identified and, the global results in terms of frequencies, damping ratios and modal shapes are shown in Figure 4a. Finally, in Figure 4b, the relevant AutoMAC matrix is shown. It is interesting to note that, despite the severity of the damage experienced by the church, the dynamics of the structure presents the first three modes very well decoupled: the first two translational modes and the third rotational one. In general, measurements present a low degree of entropy as it could be expected from an undamaged system. 4. Modal updating and optimal sensors placement Initially, the geometry of the church was surveyed from the technical drawings and updated taking into account the severe structural damage. In-situ or laboratory mechanical tests were not carried out on the masonry, and the classification of the different masonry typologies was carried out through a careful visual inspection. In total, four different masonry types (M1, ..., M4) were recognised, as illustrated in Figure 5a.

Fig. 3. AVTs: (a) measurement setup at the lower level; (b) measurement setup at the upper level; (c) stabilization diagram and frequency – damping diagram with highlighted stable solutions.