PSI - Issue 78

Daniele Storni et al. / Procedia Structural Integrity 78 (2026) 237–244

242

According to the aims of the research, the dynamic behaviour of the tower has been investigated even in response to seismic events. On 12 March 2023, a seismic swarm occurred in the area, with the mainshock reaching a local magnitude M L of 3.0 and an epicentral distance of approximately 35 km from the monitored structure. During that event the OSU seismic station acquired the seismic input that was characterized by PGA values of 0.135 cm/sec² in the E-W direction, 0.098 cm/sec² in the N-S direction, and 0.078 cm/sec² in the Z direction, demonstrating the capacity of this monitoring system to catch light and far events also. The dynamic behaviour of the tower has been evaluated in terms of global displacements as reported in Figure 6. A global displacement of 1.3 mm, 1.1 mm, and 2.4 mm has been calculated at the monitored levels. The analysis is coherent with the modal properties already discussed and identifies average drifts of 0.044‰ at Level 1, 0.028‰ at Level 2, and 0.118‰ at Level 3. The result underline s that the higher displacement demand affects the upper level of the building due to the significant lateral stiffness reduction due to the windows.

Fig. 6. Displacement tracking during the seismic event ML3.0 on 2023-12-07

As already introduced, the tower went under a SHM monitoring campaign. The six modes reported in Figure 5, have been tracked from 2023-09-01 until 2025-02-20 and reported in Figure 7. The tracking identifies a fluctuation of the natural frequencies, mostly concentrated between May and July, probably correlated to sensible temperature variation during that period. It is worth noting that, due to some external issue, the data were not saved between 11/08/2024 and 6/09/2024. During that period, the average external temperature dropped from an average of 30.5°C to 22.4°C. This detail defines the drop that is notable in Figure 7. From a general perspective, Figure 7 does not identify anomalies in the frequency trends that are actually affected by the external temperature. Figure 7 shows the evolution of the first six natural frequencies from September 2023 to February 2025. From September 2023 to January 2025, three daily 30-minute recordings of ambient vibration noise were acquired at 2:00 a.m., 12:00 p.m., and 10:00 p.m. In February and March 2025, 48 recordings of 30 minutes each are available. The structure frequencies are obtained through automatic modal identification and tracked over the monitoring period. The time series are well-defined with daily dispersion contained within 5 – 10 milliHz (mHz), and no long-term drift is observed. Analysis coverage reaches 90% of the observation period. As already investigated by other authors (Ubertini F. et al. 2017, Pellegrini D. et al. 2023, Spina D. e t al. 2024, Sivori D. et al. 2025) the masonry structures are significantly affected by temperature variation and weather conditions. Unfortunately, a weather station was not implemented on the tower and, within the aim of this research, a public database has been adopted for estimating correlations between temperature, humidity, wind speed and frequencies. As reported in Figure 8.a, the temperature affects mainly all the tracked modes. In detail the frequencies from first to third are signed by a gradient of 3.96 mHz/°C, 3.88 mHz/°C and 3.44 mHz/°C. The fourth mode is characterized by a negative slope of -2.83 mHz/°C. The fifth and sixth are signed by 7.47 mHz/°C and 4.73 mHz/°C. In contrast to the others, only the fourth mode is related to the temperature by a negative slope. Figure 8.b reports the humidity-frequency dependency. The frequency-humidity dependency can be considered for fifth and sixth modes only. Finally, Figure 8.c shows no meaningful correlation between wind and frequency only for the third mode.