IOMAC'15 6th International Operational Modal Analysis Conference 2015 May12-14 Gijón - Spain
NON-DESTRUCTIVE EXPERIMENTAL TESTS ON A HISTORICAL SLENDER STRUCTURE M. Diaferio1, D. Foti2, C. Gentile3, N.I. Giannoccaro4, A. Saisi5 1
Assistant Professor, Department of Sciences in Civil Engineering and Architecture, Politecnico di Bari,
[email protected] 2 Associate Professor, Department of Sciences in Civil Engineering and Architecture, Politecnico di Bari,
[email protected] 3 Associate Professor, Department of Architecture, Built environment and Construction engineering, Politecnico di Milano,
[email protected] 4 Associate Professor, Department of Engineering for Innovation, University of Salento,
[email protected] 5 Assistant Professor, Department of Architecture, Built environment and Construction engineering, Politecnico di Milano,
[email protected]
ABSTRACT The paper describes the application of conventional accelerometers and innovative microwave remote sensing to operational modal testing of a historical slender structure: the bell tower of the Cathedral of Trani. In more details, the natural frequencies and mode shapes of the investigated tower are identified by applying different output-only techniques to the data obtained by means of the two measurement systems. Full details on the experimental procedures, instrumentation and data analysis are reported in the paper; furthermore, accuracy and limits of the radar-based technology in monitoring a stiff historic building are addressed as well. Keywords: Ambient vibrations, Non-destructive testing, Microwave remote sensing, Operational modal analysis,.
1.
INTRODUCTION
As it is well known, Italy is a country with a huge number of outstanding historical monuments and Cultural Heritage buildings, and a relatively high seismic risk; as a consequence, during the last decades a great attention has been devoted to the assessment and preservation of Cultural Heritage structures. Nevertheless, the vulnerability assessment of a historic building is a step-by-step and multi-disciplinary process, which starts with accurate on-site inspections and concludes with Finite Element (FE) modelling and numerical analyses: in principle, the collected knowledge of global and
local geometry, the construction details and the mechanical characterization of the materials should be synthesized in the FE model, so that the results of numerical simulations should provide the decision makers with an exhaustive picture of the structural condition. On the other hand, the deep knowledge needed for the definition of a reliable model is usually unavailable for a historic construction and the possibility of conducting on-site tests is limited to the non-destructive (or minor destructive) ones. Hence, dynamic tests in operational conditions have acquired an increasing importance in the recent studies [1]-[10], where the modal parameters identified from ambient vibration data have been used to update some uncertain structural parameters of the FE models. Ambient vibration testing and continuous monitoring of the structural response under ambient excitation are especially suitable to Cultural Heritage because of the fully non destructive and sustainable way of testing, that is performed by just measuring the dynamic response under ambient excitation and does not involve additional loads rather than those associated to normal operational conditions. Furthermore, operational modal testing methods are suitable for the analysis of slender structures, that are more sensitive to ambient actions and exhibit a cantilever-like behavior, so that an appropriate modal characterization usually requires a relatively limited number of sensors [1]-[10]. The slender historical structure herein considered is the bell tower of the Cathedral of Trani (Bari, Italy). The original tower dates back to the 12th century and was built in tufa masonry (Stone of Trani). The tower underwent a series of interventions across the centuries and, in particular, it was completely disassembled and reassembled between 1950-1960. The tower is about 60 m high and is connected to the Cathedral through a step supported by a pointed arch (Fig.1). In order to define a numerical model that describes the actual behavior of the tower, the available documentation has been examined but − due to the lack of some structural information − it has been decided to perform dynamic tests in operational conditions and to utilize two different techniques: accelerometers and microwave remote sensing.
Figure 1. Views of the Cathedral of Trani and its bell tower.
The tower has been instrumented with 28 high sensitivity accelerometers, placed on four different levels (Fig. 3) and several environmental acquisitions were carried out with a sampling time of 256 Hz and 10 minutes of length. The acquisitions have been analyzed with two different operational modal techniques and the first 5 natural frequencies and mode shapes have been estimated. The radar sensor used in this work is an industrially engineered microwave interferometer (IDS, IBISS system) [11]-[12]: electromagnetic signals are emitted at a central frequency of 17.2 GHz with a maximum bandwidth of 300 MHz, so that the radar is classified as Ku-band, according to the standard radar-frequency letter-band nomenclature from IEEE Standard 521-1984. The two main fronts of the tower were surveyed by microwave remote sensing. For each radar position, deflection (radar) data were acquired at a rate of 100 Hz over time windows of 1800 s and 3600 s; due to the presence of an appropriate ambient excitation (gusts of wind) along one of the tower’s main directions, the dynamic deflections in that direction exceed 5-10 times the sensor resolution (0.02 mm) [11], so that significant
results have been obtained also in terms of natural frequencies and modal amplitudes. Although only the fundamental mode has been identified using the radar sensor, the results are in good agreement with the ones obtained from accelerometer data.
2.
DESCRIPTION OF THE CATHEDRAL OF TRANI AND ITS BELL TOWER
The Cathedral of Trani (Fig. 1) is one of the most important examples of Apulian Romanesque style; its construction began at the end of the 11th century and ended in the 1143. In the subsequent years, a tall bell tower neighbouring the Cathedral was built. During the centuries, the tower underwent to some interventions with the aim of increasing its stability: in the 18th century, some openings were walled and in the 19th century ties have been installed. Nevertheless, the most important intervention was dated 1950 when the tower was completely disassembled and reassembled. The available structural drawings and design reports are referred to this renovation but they are not sufficiently detailed, so that direct inspection of the tower was necessary to complete the geometric description of the building. The bell tower, 57 m tall, has a square plane with a side of about 7.5 m and the masonry thickness is about 1.40 m; the tower is composed by six floors and is connected to the church through a step supported by a pointed arch. The bell tower is supported by strong quadruple arcade and has a unique rigid floor on top of the arcades; the corresponding height is about 14.30 m from the ground and this level hosts the opening allowing the communication between the church and the tower. The square cross-section does not change in elevation between the aforementioned floor level till the height of about 47.60 m. Above the height of 47.60 m, the tower section becomes octagonal, with a side of about 2.30 m, and remains constant till the height of about 51.00 m, from which the dome emerges. All the four façades of the tower are characterized by the same openings, even if their shape and dimensions vary along the height of the tower. Four bells are installed at the fourth level.
3.
DYNAMIC TESTING USING ACCELEROMETERS
3.1. Experimental setup The dynamic response of the tower to ambient excitation was recorded using 28 high sensitivity piezoelectric accelerometers (PCB Model 393B31, Fig. 2) connected to 4 DAQs; one switch device collected the Ethernet cables from each group of channels and the digitized data were transmitted to a laptop.
Figure 2. Piezoelectric accelerometers (PCB Model 393B31).
The accelerometers were placed on the four different levels, referred to as 1, 2, 3, 4 in Fig. 3a: six accelerometers are placed at Level 1 (height of 22.55 m), six at Level 2 (height of 35.57 m), eight at Level 3 (height of 42.07 m), and eight at Level 4 (height of 49.26 m). The identification number of the nodes of the ARTeMIS model [13] used in the identification process and the arrows corresponding to the acquisition direction of the sensors are illustrated in Fig. 3b.
Eleven consecutive tests (named 1, 2…, 11 respectively) were conducted on the tower; each test had a duration of 10 minutes with a sampling frequency of 1024 Hz, which has been subsequently decimated by a factor equal to 4 to have a frequency of 256 Hz. It should be considered that every 15 minutes the bells of the tower ringed according to the local time; the bells contribute is evident in almost all the tests. (a)
(b)
Level 4 Level 3 Level 2
Level 1
Level 1
Level 2
Level 3 Level 4 Figure 3. (a) Instrumented levels along the height of the tower; (b) Plan views: sensors location and signal acquisition direction.
3.2. Results The structural identification of the bell tower of the Trani Cathedral has been carried out by means of the techniques of the OMA (Operational Modal Analysis) based on output-only measured data. The OMA methods herein used are: the EFDD (Enhanced Frequency Domain Decomposition, [14]) that operates in the frequency domain, and the data-driven SSI (Stochastic Subspace Identification, [15]) technique that operates in the time domain. In the examined case, the identification of the structural modal parameters was carried out by means of ARTeMIS Extractor software [13]. It is woth mentioning that in all the tests with the bells swinging, the accuracy of the identification was lower than in the other cases, due to the significant influence of the external action. An example of the data analysis is shown in Figs. 4 and 5 for the two considered EFDD and SSI-data techniques.
Figure 4. Typical results of the EFDD technique.
Figure 5. Typical results of the SSI-data technique.
Mode 1: f = 2.04 Hz
Mode 2: f = 2.26 Hz
Mode 3: f = 7.04 Hz
Mode 4: f = 7.60 Hz
Mode 5: f = 9.20 Hz
Figure 6. Identified mode shapes.
All the available datasets have been analyzed applying both the EFDD and the SSI-data techniques; the estimated natural frequencies are almost the same for the two procedures and for all the examined datasets. The results are summarized in Table 1.
Table 1. Mean values and standard deviation of the identified natural frequencies for all the eleven tests. Mode number
EFDD method Mean value Standard [Hz] deviation
SSI method Mean value Standard [Hz] deviation
1
2.04
0.004
2.03
0.002
2
2.26
0.012
2.28
0.013
3
7.03
0.020
7.07
0.112
4
7.60
0.024
7.68
0.088
5
9.16
0.173
8.94
0.064
The mode shapes corresponding to the identified frequencies are shown in Fig. 6. The first two natural frequencies correspond to the first flexural modes along the principal directions, the third mode corresponds to the second flexural mode along the axis parallel to the church façade, while the fourth mode involves torsion and the fifth one is the second flexural mode along the axis orthogonal to the church façade.
4.
DYNAMIC TESTING USING MICROWAVE REMOTE SENSING
4.1. Description of the interferometric radar and experimental procedures As previously pointed out, the radar sensor used in this work (Fig. 7) is an industrially engineered microwave interferometer (IDS, IBIS-S system) and consists of a sensor module, a control PC and a power supply unit. The sensor unit is a coherent radar (i.e. a radar preserving the phase information of the received signal) generating, transmitting and receiving the electromagnetic waves. The sensor unit is installed on a tripod equipped with a 3D rotating head (Fig. 1), allowing the sensor to be orientated in the desired directions. Data control and acquisition are completely managed through a laptop via a standard USB interface; the control PC is provided with the software for the system management, data storage, basic signal processing and preliminary view of the results in real time.
Figure 7. View of the microwave interferometer (IDS, model IBIS-S).
The main information provided by the microwave interferometer [11]-[12] is the synthetic image of the scenario and the displacement time histories of the points in the scenario that are characterized by a good electromagnetic reflectivity. Each radar image represents a distance map of the intensity of radar echoes coming from the reflecting targets: for example, each discontinuity of a structure − such as the "corner zones" corresponding to the intersection of girders and cross-beams in the deck of bridges − represents a good reflecting target and can be identified as a relative maximum of the echo amplitude. The displacement of each reflecting target is evaluated from the phase variation of the back-scattered microwaves associated to that target at different times. The main technical characteristics of the radar sensor are the following: maximum range (distance) resolution: 0.50 m; maximum sampling frequency: 200 Hz; maximum operational distance: > 500 m; displacement accuracy: < 0.02 mm. It should be noticed that the maximum operational distance and the displacement accuracy are strongly related to the intensity of the reflected signal. It is worth underlining that the microwave interferometer has only 1-D imaging capabilities, i.e. different targets can be unambiguously detected if they are placed at different distances from the radar; hence, measurement errors may arise from the multiplicity of contributions coming from different points placed at the same distance from the radar. Furthermore, the sensor measures displacement along the line of sight (LOS) only so that the evaluation of actual deflections requires the prior knowledge of the direction of motion. The two main fronts of the tower were surveyed by microwave remote sensing. For each radar position, deflection (radar) data were acquired at rate of 100 Hz over time windows of 1800 s and 3600 s. It is worth mentioning that during the tests, the tower was excited by gusts of wind in the direction orthogonal to the façade of the Cathedral (Fig. 8a), whereas the dynamic excitation was much lower in the other main direction and represented only by micro-tremors. The picture in Fig. 8a exemplifies the radar position on site and highlights that the tower is characterized by the presence of string courses approximately corresponding to the levels of the internal floors, which are placed at the height of +21.00 m, +27.72 m, +35.10 m, +41.00; furthermore, one cornice is present at the height of about +48.22 m. The range profile of the scenario obtained in the test in the direction orthogonal to the façade of the Cathedral (Fig. 8a) is presented in Fig. 8b: it is observed that well defined relative maxima of the radar image are detected at the heights of +21.40 m, +27.80 m, +34.80 m and +47.00 m, corresponding with good accuracy to the levels were the string courses and the top cornice are present (Fig. 8a). The inspection of the range profile also reveals the presence of: (a) another well defined peak at the height of +12.10 m, conceivably corresponding to the lower string course and (b) two closely spaced peaks at +38.80 m and +41.50 m, probably associated to the reflections of the upper string course and neighboring window. The deflection response associated to those relative maxima turned out to be characterized by very low amplitude of the
deflection or by not appropriate quality of the radar signal. The accuracy of the deflection responses corresponding to the targets shown in the range profile of Fig. 8b is confirmed by the inspection of the displacements in time domain (Fig. 9) as well as by the operational modal analysis. (a)
(b) 60
z = 21.40 m
Amplitude (dB)
50
z = 27.80 m z = 34.80 m
40
z = 47.00 m 30
20
10
0 0
20
40
60
80
range (m)
Figure 8. (a) View of the microwave interferometer and position of the good reflecting targets detected along the height of the tower; (b) Range profile of the scenario in the dynamic test along one main direction of the tower.
4.2. Results Figs. 9a-d show samples of the displacement time-histories measured at the different levels (+21.40 m, +27.80 m, +34.80 m and +47.00 m, respectively) of the tower during the experimental survey in the direction orthogonal to the façade of the Cathedral (Fig. 8). It should be noticed that, as it has to be expected at the low level of excitation that existed during the tests, the recorded displacements are very low, with the maximum amplitudes generally ranging between 0.10 mm (lower levels) and 0.20 mm (top cornice); in addition, the deflection time histories seems rather "noisy". On the other hand, the deflection amplitude exceeds 5-10 times the sensor resolution (0.02 mm, [11]): hence, significant results have to be expected in terms of natural frequencies and modal amplitudes at the different target points. (a) +21.40 m
(b) +27.80 m
(c) +34.80 m
(d) +47.00 m
Figure 9. Samples (600 s) of the deflection measured by the radar in the test along the direction orthogonal to the façade of the Cathedral at: (a) z=+21.40 m; (b) z=+27.80 m; (c) z=+34.80 m and (d) z=+47.00 m.
Figs. 10a and 10b show the results of the application of the well-known FDD and SSI-data methods to the displacement data collected in the time window of 3600 s. The two output-only techniques clearly identify the fundamental mode of the tower in the investigated direction. The identified value of the natural frequency are 2.021 Hz (FDD) and 2.026 Hz (SSI-data) and exhibit excellent agreement with the values identified from conventional accelerometer data; the corresponding mode shapes are compared in Fig. 10c. (a)
(c)
(b)
Figure 10. Identification of the tower fundamental frequency from radar data: (a) Singular Value lines (FDD); (b) Stabilization diagram (SSI-data). (c) Comparison between the fundamental mode shape identified using the FDD (−−) and the SSI-data (−−) methods.
5. CONCLUSIONS The study focuses on the evaluation of the modal parameters of the historical bell tower of the Trani Cathedral through the application of high sensitivity accelerometers and a microwave remote sensing. The dynamic tests were performed in operational conditions and the extraction of modal parameters from ambient vibration data was carried out by using the FDD/EFDD the data driven SSI techniques. The results obtained considering the data recorded by the two measurement systems are in good agreement and encourage about the effectiveness of the tests and the accuracy of the estimated modal parameters. On the other hand, the application of microwave remote sensing to historic structures − although promising and in principle very attractive − turns out to be characterized by some potential issues: (a) historic structures (such as towers or churches) generally exhibit few "corner zones" effectively reflecting electromagnetic waves; (b) the actual deflection of the surveyed structure needs to be larger than the radar sensitivity.
ACKNOWLEDGEMENTS
The support of IDS (Ingegneria Dei Sistemi, Pisa, Italy) in supplying the IBIS-S radar sensor employed in the tests is gratefully acknowledged. REFERENCES [1] Ivorra, S., & Pallares, F.J. (2006). Dynamic investigations on a masonry bell tower. Engineering Structures , 28 (5), 660-667. [2] Gentile, C., & Saisi, A. (2007). Ambient vibration testing of historic masonry towers for structural identification and damage assessment. Construction and Building Materials, 21(6), 1311-1321. [3] Diaferio, M., Foti, D., & Sepe, V. (2007). Dynamic identification of the tower of the Provincial Administration of Bari, Italy. In: Proceedings of the 11th Int. Conf. on Civil, Structural and Environmental Eng Computing, (paper n. 2), Malta.
[4] Ramos, L., Marques, L., Lourenco, P., De Roeck, G., Campos-Costa, A., & Roque, J. (2010). Monitoring historical masonry structures with operational modal analysis: two case studies. Mechanical Systems and Signal Processing, 24(5), 1291-1305. [5] Tomaszewska, A., & Szymczak, C. (2012) Identification of the Vistula Mounting tower model using measured modal data. Engineering Structures, 42, 342-348. [6] D'Ambrisi, A., Mariani, V., & Mezzi, M. (2012). Seismic assessment of a historical masonry tower with nonlinear static and dynamic analyses tuned on ambient vibration tests. Engineering Structures, 36, 210-219. [7] Diaferio, M., Foti, D., & Giannoccaro, N.I. (2014) Optimal model of a masonry building structure with wooden floors. In: Proceedings of the 2014 Int. Conf. on Civil Engineering (CIVILENG 2014), Santorini Island, Greece. [8] Diaferio, M., Foti, D., & Giannoccaro, N.I. (2014). Non-destructive characterization and identification of the modal parameters of an old masonry tower. In Proceedings of the IEEE Workshop on Environmental, Energy, and Structural Monitoring Systems, Naples, Italy. [9] Diaferio, M., Foti, D., Giannoccaro, N.I., & Ivorra, S. (2014). Optimal model through identified frequencies of a masonry building structure with wooden floors. International Journal of Mechanics, 8, 282-288. [10] Gentile, C., Saisi, A., & Cabboi, A. (2015) Structural identification of a masonry tower based on operational modal analysis. Int. Journal of Architectural Heritage, 9, 98-110. [11] Gentile, C., & Bernardini, G. (2010). An interferometric radar for noncontact measurement of deflections on civil engineering structures: laboratory and full-scale tests. Structure and Infrastructure Engineering, 6(5), 521-534. [12] Gentile C., Saisi A. (2015). Dynamic testing of masonry towers using the microwave interferometry. Key Engineering Materials, 628, 198-203. [13] SVS (2012). ARTeMIS Extractor 2011 release 5.3. http://www.svibs.com. [14] Brincker, R, Ventura, C.E., & Andersen, P. (2001). Damping estimation by Frequency Domain Decomposition. In: Proceedings of the 19th Int. Modal Analysis Conf. (IMAC-XIX). Orlando, USA. [15] Van Overschee, P., & De Moor, B. (1996) Subspace identification for linear systems: theoryimplementation-applications. Kluwer Academic Publishers.
