Exploiting Indoor-Induced Vibrations at Castello Normanno-Svevo Aci Castello

Abstract

This study investigates the vibrations at the Castello Svevo-Normanno in Aci Castello (Catania), focusing on its historical and cultural significance. The research aims to analyze vibration levels and frequency distribution to achieve two objectives: protecting historical artifacts and structures through preventive vibration analysis and exploring the use of kinetic energy for powering autonomous systems. The study specifically focuses on the indoor context to understand its unique vibrational characteristics. Measurements were recorded along the X, Y, and Z axes, with detailed analysis of the Z axis using Fast Fourier Transform (FFT) and Power Spectral Density (PSD). The results revealed consistent vibration patterns across all axes, with the Z axis significantly influenced by environmental factors such as wind and sea movement. These findings provide valuable insights for designing optimized energy harvesting systems, electromechanical converters, and monitoring devices suitable for operation in this specific historical context.

1. Introduction

Vibration analysis has emerged as a critical tool for assessing the structural integrity of historical buildings, offering non-invasive insights into their dynamic behavior while preserving cultural heritage [1,2]. Traditional approaches, such as laboratory-based modal analysis, numerical simulations [3,4], and analytical or semi-analytical formulations for beam-like elements on elastic foundations [5], often fail to account for real-world environmental factors (e.g., wind, traffic) that influence structural responses. While these methods provide controlled data, they lack the fidelity required to evaluate aging structures in their natural settings, particularly in seismically active regions like Sicily [6,7]. In [8], the authors conducted a study on vibration analysis at the Castello Ursino Picture Gallery in Sicily, Italy. Their research focused on monitoring vibrations within the gallery and exploring the possibility of using them as a source of energy. They examined how vibrations, including those caused by visitors, could be harnessed. Recent advancements in portable sensor technology and on-site measurement techniques now enable researchers to capture real-time, multi-axis vibration data directly from heritage sites, a shift that addresses gaps in traditional methodologies [9].

This study presents the results of an analysis carried out on the 11th-century Norman castle of Aci Castello, a small town located on the northern coast of Catania, Sicily. The monument represents the main attraction of the area, and every year, it houses a huge number of visitors. Through the ages, the castle became the symbol of Aci Castello and its architecture, overlooking the sea, and naturally merged with the basaltic spur upon which it is built, providing a strongly evocative view. Inside the castle, an archeological museum illustrates the history of the area from the Neolithic to Medieval Age, enriched with a mineralogical collection. In this respect, conservation and valorization of the monument are crucial for improving the touristic economy of the town, on one hand, and for reinforcing the identity and sense of self of the local community.

It is worth noting that a preliminary outdoor study was presented in [10]. Our approach combines triaxial measurements with environmental context, capturing the castle’s dynamic response to natural forces such as wind loads. Unlike previous works that prioritized controlled environments or simplified models, our research focuses on triaxial vibration measurements (X, Y, Z axes) across critical structural zones of the castle.

This approach aligns with the growing emphasis on operational modal analysis (OMA), which extracts structural dynamics under ambient conditions without artificial excitation [11]. Within multi-level assessment approaches for very old masonry constructions, ambient vibration measurements provide a fundamental observational level for identifying global dynamic properties while fully preserving the structure. For historic castles, where material degradation, construction irregularities, and complex soil–structure interaction are often unknown, vibration data acquired under operational conditions support the calibration of numerical models by constraining stiffness distribution and boundary conditions. The analysis of induced vibrations therefore contributes to a more reliable interpretation of the structural behavior of ancient monuments, reducing uncertainties inherent in purely analytical or numerical approaches. For instance, Brigante et al. demonstrated the value of vibration testing for heritage preservation but limited their work to isolated laboratory settings, overlooking site-specific environmental interactions [12]. Similarly, Russotto et al. applied vibration monitoring to Sicilian monuments but used single-axis sensors, reducing data granularity [13].

Focusing attention on structural health monitoring (SHM), structural identification [14], this study plays a crucial role in preserving historical buildings by providing data on their stability and response to environmental factors. In fact, recent studies have demonstrated that advanced monitoring techniques, such as vibration analysis and wireless sensor networks, offer reliable and non-invasive ways to assess the condition of heritage structures without causing damage [15]. For example, research by Drdácký and Urushadze highlights the importance of using dynamic testing methods to capture authentic structural behavior, while other studies emphasize the benefits of triaxial vibration measurements for detecting site-specific influences like wind and seismic activity [16]. By integrating modern SHM techniques [17,18] with on-site measurement approaches, researchers can better understand how historical buildings respond to natural forces, helping to develop more effective preservation and restoration strategies [19,20]. It is worth mentioning that various studies have focused on a multi-level approach, such as the one proposed in [21] for Romanian Orthodox masonry churches in the Banat area, and in [22], where the authors proposed an inventory of masonry churches in the Groningen area.

This paper improves the state of the art by presenting, for the first time to the authors’ knowledge, a study on indoor vibrations induced by natural agents such as wind, wave-generated kinetic energy from the sea, and external sources acting on the studied site. It is worth noting that the novelty lies in the in situ study of this historic site, which has never been previously investigated in terms of indoor-induced vibrations and analyses conducted within the archeological museum.

On-site measurements at Castello Normanno-Svevo Aci Castello revealed unique challenges and insights. For example, the castle’s basaltic rock foundation, volcanic geology, and coastal exposure to humidity introduce complex vibrational modes that cannot be easily replicated in simulations, except under specific assumptions. By comparing our results with earlier studies on Mediterranean historical sites [14], for example, we demonstrate how site-specific factors influence structural behavior, emphasizing the need for context-aware monitoring. The study is of interest not only in terms of conservation and preventive conservation but also for estimating kinetic energy levels that could potentially be harnessed for energy recovery, with the goal of powering autonomous monitoring systems and sensor nodes. By establishing a framework for adaptive preservation strategies in the region, the findings also aim to guide restoration protocols for comparable historical structures, helping to ensure their resilience to natural hazards while maintaining their architectural integrity.

2. Materials and Methods

This study employed a smartphone-based approach to perform on-site vibration analysis at the Castel of Aci Castello in Catania, Sicily. The objective was to collect and analyze real-world vibration data in different parts of the monument, capturing its structural response under natural conditions. The methodology focused on precision and repeatability, ensuring comprehensive and reliable results.

2.1. Data Collection

The environmental conditions during the measurements were documented to ensure proper operating conditions. The temperature ranged from 18.5 °C to 20 °C, with a northern wind direction and a wind speed of approximately 20 km/h. Figure 1 shows a real photo of one of the rooms where the measurements were carried out.

The measurements were conducted using an iPhone 13 Pro Max equipped with the Vibration Analysis application. This application allowed for the real-time recording of vibrations along the X, Y, and Z axes. A compass was used to standardize the orientation of the smartphone, which was placed horizontally on the ground with its top aligned to a fixed direction of 260° west. This alignment ensured consistency across all measurements.

To achieve comprehensive coverage, vibration measurements were conducted in four different rooms inside the castle, as shown in Figure 1 and Figure 2. Within each room, measurements were recorded at three distinct positions. Each measurement session lasted 5 min, during which the smartphone remained stationary on the floor. The measurements were conducted sequentially. All selected locations were within enclosed areas of the castle and were chosen to represent different structural zones in order to analyze variations in the structural response. It is also noted that the castle was closed to the public on the day the measurements were carried out, ensuring minimal external disturbance.

2.2. Data Processing

The raw vibration data collected were exported for analysis. Signal processing was performed using MATLAB^® R2024b, leveraging its robust tools for vibration analysis. Two key techniques were applied:

• An FFT-based method was used to convert the time-domain vibration data into the frequency domain, identifying the dominant frequencies and spectral components [23].
• PSD-based analysis provided a detailed representation of the energy distribution across different frequencies, enabling the assessment of structural dynamics and resonances [24].

3. Results

3.1. Time-Domain Analysis

The measurements in this study were taken in eleven fully enclosed areas, which are numerically indexed (1–11) in Figure 3 to indicate the specific measurement locations. Uppercase letters (A–F) identify the corresponding museum rooms in which these locations are situated.

To study the general behavior of the signal, we plotted the time domain of the cumulative signal. In this way, we can identify the consistency of the vibrations not only at each point by itself but also across all the measurement points. The time-domain analysis helps in identifying whether there is any uncommon behavior at a specific spot.

Figure 4 shows the time-domain waveforms and, in particular, their variations along the X, Y, and Z axes.

From the analysis of the plots, it is possible to observe the consistency of the vibrations regardless of time and location, particularly along the X and Y directions. However, the vibrations along the Z axis exhibit more spikes. The sign denotes the direction of motion: upward/downward along the Z-axis, backward/forward along the X-axis, and left/right along the Y-axis.

From

Table 1. Statistical parameters of vibrational data on X axis.

Spots	Minimum (g)	Maximum (g)	Mean (g)	STD (g)
P1	−2.908 × 10−3	2.738 × 10−3	3.906 × 10−13	8.640 × 10−4
P2	−2.410 × 10−3	2.733 × 10−3	3.681 × 10−6	8.283 × 10−4
P3	−2.673 × 10−3	2.592 × 10−3	4.883 × 10−13	8.479 × 10−4
P4	−2.653 × 10−3	2.413 × 10−3	9.766 × 10−13	8.472 × 10−4
P5	−2.646 × 10−3	2.465 × 10−3	7.812 × 10−13	8.401 × 10−4
P6	−3.020 × 10−3	2.428 × 10−3	1.238 × 10−6	8.428 × 10−4
P7	−2.928 × 10−3	2.291 × 10−3	1.953 × 10−13	8.585 × 10−4
P8	−2.744 × 10−3	3.192 × 10−3	7.813 × 10−13	8.714 × 10−4
P9	−2.669 × 10−3	2.870 × 10−3	8.789 × 10−13	8.718 × 10−4
P10	−2.770 × 10−3	2.769 × 10−3	4.883 × 10−13	8.407 × 10−4
P11	−2.451 × 10−3	2.416 × 10−3	1.367 × 10−12	8.360 × 10−4
Cumulative	−3.020 × 10−3	3.192 × 10−3	4.472 × 10−7	8.496 × 10−4

,

Table 2. Statistical parameters of vibrational data on Y axis.

Spots	Minimum (g)	Maximum (g)	Mean (g)	STD (g)
P1	−2.395 × 10−3	2.365 × 10−3	4.883 × 10−13	7.552 × 10−4
P2	−2.188 × 10−3	2.268 × 10−3	1.207 × 10−6	7.696 × 10−4
P3	−2.279 × 10−3	2.817 × 10−3	3.388 × 10−21	7.311 × 10−4
P4	−1.847 × 10−3	2.166 × 10−3	7.813 × 10−13	7.157 × 10−4
P5	−2.068 × 10−3	2.357 × 10−3	2.148 × 10−12	7.753 × 10−4
P6	−2.249 × 10−3	2.252 × 10−3	6.263 × 10−7	7.606 × 10−4
P7	−2.263 × 10−3	2.300 × 10−3	1.172 × 10−12	7.272 × 10−4
P8	−2.790 × 10−3	2.353 × 10−3	2.051 × 10−12	7.243 × 10−4
P9	−2.142 × 10−3	2.131 × 10−3	4.883 × 10−13	7.200 × 10−4
P10	−2.195 × 10−3	2.062 × 10−3	1.660 × 10−12	7.139 × 10−4
P11	−2.154 × 10−3	2.149 × 10−3	3.906 × 10−13	7.066 × 10−4
Cumulative	−2.790 × 10−3	2.817 × 10−3	5.279 × 10−8	7.306 × 10−4

and

Table 3. Statistical parameters of vibrational data on Z axis.

Spots	Minimum (g)	Maximum (g)	Mean (g)	STD (g)
P1	−3.133 × 10−3	2.467 × 10−3	9.766 × 10−14	7.829 × 10−4
P2	−2.211 × 10−3	2.657 × 10−3	1.668 × 10−5	7.724 × 10−4
P3	−6.291 × 10−3	4.528 × 10−3	7.813 × 10−13	8.056 × 10−4
P4	−2.487 × 10−3	2.640 × 10−3	1.074 × 10−12	7.568 × 10−4
P5	−2.774 × 10−3	2.857 × 10−3	8.789 × 10−13	7.753 × 10−4
P6	−2.334 × 10−3	2.075 × 10−3	2.068 × 10−6	7.764 × 10−4
P7	−2.614 × 10−3	2.620 × 10−3	1.953 × 10−13	7.596 × 10−4
P8	−4.316 × 10−3	5.236 × 10−3	1.855 × 10−12	10.950 × 10−4
P9	−2.225 × 10−3	2.673 × 10−3	2.930 × 10−13	7.671 × 10−4
P10	−2.260 × 10−3	2.226 × 10−3	7.516 × 10−6	7.394 × 10−4
P11	−2.266 × 10−3	2.343 × 10−3	8.789 × 10−13	7.872 × 10−4
Cumulative	−6.291 × 10−3	5.236 × 10−3	1.021 × 10−6	8.068 × 10−4

, it is possible to observe the consistency of the vibrations along the X and Y axes. Similarities can also be noticed among the three signals in the mean and standard deviation values. The mean acceleration values are reported as absolute values. Additionally, the vibration magnitude is generally bounded between ±0.003 g, with some exceptions in the Z direction, where it reaches up to 0.006 g.

In the next section, an in-depth analysis is performed by conducting FFT analysis on the vibration on the Z axis to understand the reason behind the observed spikes.

3.2. Fast Fourier Transform Analysis

The FFT breaks down a signal to its frequency components which would provide an understanding of the dominant frequencies of the signal. We therefore study the FFT of the measured vibrations. Two of the signals contain spikes (spots 3 and 8) and the other 9 spots have more consistent signals.

The plotting of the FFT curves is on a logarithmic scale to allow us to focus on the influence of low frequencies on the vibrations. Figure 5, Figure 6, Figure 7, Figure 8, Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14 and Figure 15 show the FFT waveforms of the vibrations on Z axis in spots 1 to 11, respectively. The overlaid line is provided as a guide for the eye.

3.3. Formatting of Mathematical Components

In this section, we present the PSD plots of the signals to identify frequency ranges with high energy. Figure 16, Figure 17, Figure 18, Figure 19, Figure 20, Figure 21, Figure 22, Figure 23, Figure 24, Figure 25 and Figure 26 show the PSD plots of the vibrational signal from spots 1 to 11, respectively. Also in this analysis, the overlaid line is provided as a guide for the eye.

4. Discussion

Based on the time-domain analysis, the vibrational signals within the castle exhibit consistent behavior across various measurement intervals and locations, particularly along the horizontal axes. This consistency indicates that the signals remain relatively uniform in both timing and spatial distribution. However, at the third and eighth measurement locations, noticeable spikes were observed on the vertical axis.

Further examination using FFT reveals that all eleven signals display similar spectral patterns, with no single dominant frequency emerging. This finding suggests that the observed vibrations arise from multiple sources at different frequencies. The PSD analysis also shows a uniform distribution of energy in the 0–50 Hz range.

Taken together, these results imply that the vertical-axis spikes recorded at the third and eighth locations are most likely attributable to transient events (e.g., human activity) rather than any resonant or persistent frequency component.

Observing the vibrations in the castle is crucial not only for monitoring purposes [25,26] but also for enabling energy-harvesting strategies [27,28,29,30]. In particular, vibrational energy harvesting through piezoelectric devices is gaining momentum due to their capacity to convert mechanical energy into electrical energy [31,32]. In terms of extractable power to supply a measurement node, the potential power source is in the order of tens of nW by using a microscale piezoelectric energy harvester [33,34]. Clearly, appropriate power management circuits are required in this case to intermittently power the measurement node.

The results presented in this study align well with the application of nonlinear energy harvesting techniques, which exploit broader frequency and energy distributions [35,36,37,38]. Even though the observed vibration amplitudes are small and concentrated primarily within the low-frequency range, these conditions still allow for feasible low-voltage generation [35,36,37,38,39,40,41].

The use and selection of appropriate technology, especially in light of the measurements and results obtained, are crucial, including MEMS solutions with SOI substrates and energy-converting materials [42,43].

Such an approach can be highly beneficial for powering sensors [35,44,45,46] and other monitoring equipment, demonstrating how integrating piezoelectric-based solutions within the castle monitoring system can improve both sustainability and autonomy in long-term structural health management.

5. Conclusions

This paper studies the vibrations inside the Aci Castello historical monument. The measurements were taken in closed spaces to isolate any wind effect from the measurement tool, which was an iPhone 13 Pro Max in our use case. This study compares the spots with the highest vibration spikes and studies their frequency component in comparison with spots with constant behavior. The results helped in understanding the general behavior of the vibrations in the castle as well as the frequency dominance and energy distribution. These results may be enhanced by taking simultaneous measurements across multiple spots and for longer periods to check the consistency of the measured behavior. Further applications may address the use of vibration for energy-harvesting purposes using nonlinear systems.