Keywords

1 Introduction

The recent negative natural and human-induced events, i.e., include Typhoon in West Japan in 2018, Earthquakes in Central Mexico in 2017, Kumamoto in Japan, Central Italy and Myanmar in 2016, Nepal earthquake in 2015, UK floods in 2015, Balkan floods in 2014 (UNESCO 2018), highlight the relevance of the maintenance process management, which should be approached with a dynamic vision, especially by considering the climate change impacts as well (UNESCO 2017a, b). For this purpose, the knowledge of the goodness (artifacts, monuments, sites, etc.) behaviour evolution plays the main role in the assessment and decision process (Petti et al. 2018c; Baratta et al. 2019). Therefore, real time monitoring techniques are of fundamental importance to support the assessment process of the behaviour of the Cultural Heritage continuously over time, from structural point of view as well.

Furthermore, the recent technological progress in this discipline lead to develop excellent broadband high sensitivity mechanical sensors.

Recently, a wide experimental campaign was carried out on the Trajan arch in Benevento (Italy) (Petti et al. 2017a, b), whose results were applied to the design of an innovative monitoring network of the temple of Neptune in the Archaeological Park of Paestum and Velia (Italy) (Petti et al. 2018a, 2019a).

The monitoring system implemented for the Trajan arch is an application of a general adaptive assessment strategy, with the aim to get an in-deep knowledge of the monument construction details by means of non-destructive investigations. This knowledge was obtained with an optimized design and implementation of a continuous and distributed monitoring of all its important structural elements, in connection with a careful analysis based on an adaptive structural FEM (Finite Element Method) dynamic model of the Arch (Fig. 1). With this process, it was possible to assess both the dynamic and static behaviours of the arch by using anthropic (traffic) and natural (wind) noise, such as to provide reliable description and clarification of the existing damage pattern on its sculptures and structural elements.

Fig. 1
An illustration of layout of the Arch. It has constructing materials travertine, marble Lunense, opus cementum, and thick limestone building blocks labeled beside it.

Assessment of materials and construction details (Petti et al. 2017a, b)

This above-mentioned strategy required the implementation and optimization of a numerical finite element model (FEM) describing the dynamical behaviour of the Trajan Arch based on measurements obtained with an adaptive monitoring system (Barone et al. 2015/1) as described in the following Fig. 2.

Fig. 2
A line graph plots displacement spectral density versus frequency. It plots undulated lines for the central transverse, central longitudinal, and lateral transverse. The peaks are mostly higher for the central transverse at 10 Hertz.

Example of displacement spectral density of selected measurement points on the top of Trajan Arch (I test)

In the case of the Neptune temple in Paestum, several on site and laboratory tests were carried out to design the best strategy to monitor the Temple by means of an advanced sensors network. The tests, part of a joint research among the University of Salerno (IT), the University of Kassel (D) (DFG Project, “Tendon System for Protecting Ancient Column Structures”, TeSSPACS), the University of Roma “La Sapienza” (IT) and the Archaeological Park of Paestum and Velia (PAEVE), consider on site accelerometric and seismic measures, laboratory medium and small-scale tests on columns and blocks sub-structures (Obón Santacana et al. 2017; Petti et al. 2018b, 2019b).

With regard to the on-site monitoring investigations, the structure of the Temple and the soil conditions were mainly investigated respectively in the 2016 and 2019 (Fig. 3), using triaxial accelerometric sensors, characterised by 0-200 Hz frequency range and 0.25 g maximum acceleration scale range.

Fig. 3
An anterolateral view of a temple has many pillars and a Dutch roof. A crisscross boundary surrounds the temple. Some men stand and sit on the brick platforms near the temple.

Monitoring campaign of Neptune temple, 2016

Two test layouts were considered, and, in both cases, the accelerometers were installed on the north side of the Temple. In details, they were positioned between the fourth and fifth columns of the aforementioned side, numbering the columns from East to West. Figure 4 describes a test example.

Fig. 4
2 line graphs with amplitude spectrum and Fourier transform versus frequency, respectively. A rectangular layout and pillars are mentioned at the top. The first and second graphs have a peak at (0, 0.5) and (4, 200), respectively.

Example of transversal accelerometric spectrum and Fourier transform in the frequency domain

The recorded signals analysis highlighted several peaks showing a complex dynamic behaviour due to the construction complexity.

Moreover, several tests (Petti et al., 2019a; Grelle et al. 2014) were carried out in the 2019 to assess the local seismic characteristics of the soil. In this context, indeed, it is fundamental to underline that the knowledge of a monument has to be based not only on the architectural and structural aspects but also on aspects concerning the structure-soil interactions.

For the Archaeological Park of Paestum and Velia, the particular conformation of the soil, geologically characterized by a complex process of evolution of the coast and the river branches, coming from the carbonate Apennine chains, strongly affect the local seismic demand.

This seismic soil characterization highlighted a significant speed inversion of shear waves between the rigid superficial travertine bench, with an overall thickness exceeding 10 m, and the underlying loose deposited, presumably characterized by important pyroclastic fall and/or sedimentation contributions. This behaviour of the stratigraphy realizes a natural physical filter on wide frequency ranges, which presents most of the main natural modes of vibration of the structural elements and in particular in the frequency potentially damaging the temple.

Figures 5 and 6 describe the implemented seismic network and the main obtained results, respectively.

Fig. 5
A schematic of an on-site experiment set-up has dots marked for Stazione 1, 2, and 3 connected like a triangle. A dot for Stazione 4 is marked at the centroid of the triangle. Other marked points are 16.78 and 16.64.

On-site experimentation set-up, 2019

Fig. 6
2 graphs. A graph of H V S R versus frequency plots 4 fluctuating curves. The Theoretical H V S R curve is highlighted. A graph of depth versus shear wave velocity plots a decreasing uneven step function.

On-site campaign: a result of inversion analysis between the experimental HVSR curve (blue) and the theoretical one (red); b speed profile of shear waves and relative stratigraphic column of reference: A = Travertine; B = River-Dunali deposits mixed with pyroclastic deposits; C = Compact or very thickened deposits of probable fluvial and marine origin; D = Basic Seismic Bedrock

Moreover, the research activities carried out investigated the possibilities of using remote sensing technologies, as the satellite interferometric one (Bianchini et al. 2015; Casagli et al. 2017) that can be advantageous considered to assess long-time (weeks-years) global movements and deformations. These techniques allow to measure surface displacements of the ground or structures with millimetre precision thanks to the use of processing techniques called “interferometric”, which consider the phase difference between SAR images (Synthetic Aperture Radar).

2 Neptune Temple Monitoring Network Framework

As part of the joint conservation activities between the Department of Civil Engineering (DICIV) of the University of Salerno and the Archaeological Park of Paestum and Velia (PAEVE), an agreement was stipulated in 2018, aimed at the direct monitoring and the analysis of the safety of the archaeological structures present in the Park, which envisages the realization of an executive project of a dynamic and seismic monitoring network of the Neptune temple. This network has as its main purpose the monitoring of displacements and accelerations induced by natural and human-induced events on the structure, in order to obtain a more in-depth knowledge of its behaviour, and also for the purpose of evaluating the safety of the monument and visitors. In details, modular biaxial and triaxial sensors, based on high sensitivity and large band and monolithic mechanical transducers (Acernese et al. 2010; Barone et al. 2015), are implemented.

As shown in Fig. 7, the 12 points indicated in the map of the Temple will be monitored. The data acquisition is fully locally distributed. An information structure (Totem), that will be installed near the Temple to show details to inform visitors about monitoring activities, to act also as a rack for the management of the equipment, data collection and storage, for an effective real-time pre-analysis and data transfer.

Fig. 7
A model has an allocation of sensors. Triaxial sensors are placed on the base and the top of the temple. Triaxial underground sensors, biaxial sensors placed on the base and the top of the temple, and biaxial underground sensors.

Allocation of the mechanical sensors on the top, on the base and on the foundation layer of the Neptune temple

The passage of the monitoring network cables from the top of the Temple to its base is carried out by means of a cable-duct/sheath with a diameter of 7 cm along one of the columns of the West side of the Temple (Fig. 8).

Fig. 8
A 3-D model of a monitoring system around a large building in a grassy courtyard. It is labeled as a shunting box, collection table, 34 channels dynamic, and a data transfer machine.

3D view of the monitoring system set-up

More in details, the monitoring system of the Neptune temple in Paestum was conceived starting from the consideration that an effective evaluation and correlation of the static and dynamic structural behaviour of structures and sites, with natural and anthropic actions, requires the optimization and validation of adaptive structural finite elements models (FEM). Those models need a high quality and accurate experimental data, the latter acquired with modular distributed monitoring systems, adaptive both in terms of architecture and in terms of sensitivity and bandwidth of sensors. This innovative approach required a dual design action, aimed at both the implementation of the monitoring system and at the selection and specialization of sensors.

The monitoring system was designed to ensure not only compliance with the band and sensitivity requirements of the initial system for the different degrees of freedom (horizontal and vertical), but also with expandability and versatility (e.g., increase of the sensors number and/or repositioning), if the structural behaviour analysis requires actions aimed at the optimization of its performance. In order to optimize the signal-to-noise ratio, the acquisition will be obtained near the data production points (sensors), making, in this way, the system not sensitive to couplings with environmental noises. Figure 9 shows the block diagram of the data acquisition, storage, distribution and analysis system designed for the Neptune temple in Paestum.

Fig. 9
An illustration titled modular data acquisition and storage network. It has the connected components of base-level D A Q, top-level D A Q, totem section, D A Q network, and museum network.

Block scheme of the monitoring system of the Temple of Neptune in Paestum

The key element of the monitoring system is the local, compact and robust DAQ (Data Acquisition), based on the FieldDAQ™ technology of National Instruments, state of-the-art in the field both in terms of data quality and system robustness. The system consists of 34 degrees of freedom (sensors) acquired through 5 FieldDAQ™ voltage modules FD-11603 (8 voltage channels at ± 10 V, resolution up to 24 bits, simultaneous sampling frequency up to 100 kS/s, absolute accuracy of 0.019 V from −40 to 85 °C, less than 0.2% of the total measuring range).

In order to satisfy possible new requirements, the monitoring system can be easily expanded to the measurement of environmental conditions (temperature, pressure, humidity, wind speed and direction), relevant for the study of the dynamic slow-motion behaviour of the Neptune temple structure by cross-correlation analysis.

The data acquisition and storage processes are managed and synchronized by a Pc server (4 Tbyte) acting as frame builder and data storage system, downloading the data on local disks for backup, data distribution and presentation. The synchronization of the acquisition, better than 1 µs on the whole network of DAQ nodes, is carried out via Ethernet by the Pc server using the NI-DAQmx software. The time is provided by GPS card allowing, in this way, the definition of a timestamp on each frame, necessary for a correct reconstruction of signals.

The network topology, based on the IEEE 802.1AS profile (Ethernet TSN), is a daisy chain, chosen in consideration of the relatively low number of initial degrees of freedom (34) and of the relatively small distances among sensors. This network topology guarantees direct expansion of the number of sensors by increasing the number of DAQs, up to a maximum of 15 for a total of 120 channels. The simple addition of a TSN switch (compliant with the IEEE 802.1Q standard to support the Rapid Spanning Tree protocol (RSTP) and the IEEE 802.1AS standard to support synchronization of DAQs on individual branches) allows the extension of data acquisition and synchronization to other temples and structures of the historical area of Paestum, through the creation of a star configuration of daisy chains.

A direct connection to the museum Ethernet network is implemented mounting remotely the PC server storage disks through NFS, allowing both real time data backup and data presentation on monitors in museum rooms, using open software presentation standards (e.g., GEOPSY, typically used in seismology).

This technological approach leads possible also an easy development of dedicated presentation software, offering the visitor the opportunity to visualize via mobile apps in an effective and direct (even interactive) way the global and/or local dynamical behaviour of the temple of Neptune, also in connection with the daily evolution of the environmental conditions.

For what concerns sensors, the key element of the monitoring system, as it will be clear in the following, they have also been adapted to this specific application, which requires high sensitivity \(({10}^{-8} m/\sqrt{Hz})\) in the band \(\left(1 mHz-100 Hz\right)\), together with small dimensions, weights and costs.

The design and implementation philosophy characterizing the data acquisition, storage, analysis and presentation system was extended to the sensors as well.

The mechanical architecture of the sensors is based on an innovative technological platform, developed by the Applied Physics Research Group of the University of Salerno (UNISA) Folded Pendulum, international state-of-the-art platform for the design and implementation of monolithic mechanical oscillators (Acernese et al. 2010; Barone et al. 2015).

This innovative technological platform, already protected by four families of international patents, has allowed the development of linear and angular (tiltmeter) position, speed and acceleration state-of-the-art sensors. These sensors, very innovative in the field of seismology, can be calibrated in frequency, are fully scalable in terms of sizes and weights with band in the low frequency region as large as \(\left( {10^{ - 7} - 10^2 {\text{Hz}}} \right)\), can be designed to guarantee very high sensitivities \(\left( { < 10^{ - 12} m/\sqrt {Hz} } \right)\) in the band, \(\left( {10\,{\text{ mHz}} - 100 \,{\text{Hz}}} \right)\), even lower than the minimum seismic noise that can be measured on the earth's surface defined by the classical models (Peterson 1993; McNamara and Buland 2004). Moreover, they can operate in ultra-high vacuum (UHV) and cryogenic environments (<3 K), still remaining simple, compact, light and, with suitable designs, applicable to very different fields, including the cultural heritage one (Barone et al 2015; Barone and Giordano 2017).

The advantage of using a particularly sensitive and broadband instrumentation is that it allows a continuous evaluation of the dynamic behaviour of the structures, highlighting also effects of environmental forcing, such as those due to wind, to ground heating, to thermal expansion of the structures due to the different illumination from the sun during the day.

The sensors used in the initial monitoring network of the Neptune temple in Paestum, models SE-10H (horizontal) and SE-10 V (vertical) with a LVDT readout module, were slow motion seismometers (no feedback) in uniaxial configuration. These new sensors, produced by a privately owned company that develops sensors and systems for scientific and engineering applications for monitoring and control called ADV3S Company (Advanced Scientific Sensors and Systems, http://www.adv3s.com/en/homepage/), are based on the model of an aluminium folded pendulum monolithic oscillator internationally known as GE15, a choice that guarantees the implementation of biaxial or triaxial configurations by means of suitable combinations of uniaxial sensors Fig. 10.

Fig. 10
2 graphs. A scatter graph plots amplitude versus frequency. The line through the dots has a peak at (8,1000). A line graph plots acceleration versus frequency for N L N M, McNamara, and commercial L V D T, and a G E 15 mechanical oscillator.

GE15 mechanical oscillator transfer function calibrated at the frequency of 3.75 Hz (left); sensitivity (acceleration) of the GE15 monolithic oscillator equipped with different types of readouts, compared with the minimum levels of earth’s seismic noise (Peterson (NLNM) and McNamara models) (right); GE15 sensor equipped commercial LVDT readout system (bottom right)

Being the GE15 mechanical oscillator limited by its thermal noise only, its sensitivity is practically determined by the readout system, that in this specific case is a commercial LVDT. Nevertheless, a LVDT readout upgrade with an optical readout allows to get a sensitivity improvement of three orders of magnitude \(( < 10^{ - 11} m/\sqrt {Hz} )\) and a band extension in the low frequency region.

Figure 10 shows the transfer function of the GE15 mechanical oscillator calibrated at the resonance frequency of 3.75 Hz, the different sensitivities in acceleration achievable with sensors based on GE15 monolithic oscillator according to the type of readout used and a picture of the basic architecture of the seismometers SE-10H and SE-10 V, based on the GE15 oscillator and on a commercial LVDT (Barone et al, 2015; Barone and Giordano 2017).

3 Conclusions

The paper presents an innovative network for monitoring the dynamic behaviour of the local seismic response of the Neptune temple in Paestum. The system uses an excellent broadband high sensitivity mechanical sensors network that allows to carry out an effective assessment of the static and dynamic behaviour of the temple structure, also in connection with the soil-structure interaction, also monitoring in real-time the effects due to human-induced and natural events, including, in example, those related to climate change phenomena. The network will allow a real-time data collection and storage, guaranteeing an effective dissemination of the information through open source platforms, key element for carrying out a research framework aimed at a synergic centralization of he contribution of researchers on the analysis of the complex aspects that describe monumental structures and for a share of further research activities.