Ambient vibration testing at N. Sra. do Carmo Church ... · Ambient vibration testing at N. Sra. do...

6
Structural Analysis of Historical Constructions - Modena, Lourenço & Roca (eds) © 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9 Ambient vibration testing at N. Sra. do Carmo Church, preliminary results M.A. Baptista, P. Mendes, A. Afilhado & L. Agostinho ISEL - In stituto Superior de Engenharia de Lisboa, Lisboa, Portugal S. Lagomarsino Dipartimento di Ingegneria S/rutturale e Geotecnica - Università di Genova, l/alia L. Mendes Victor Centro de Geofisica da Universidade de Lisboa, Lisboa, Portugal ABSTRACT: The city of Lagos (Algarve, Portugal) is located in an area of modera te seismic risk. In recent years particular attention has been devoted to the historical centre of the city. The focus of this study is the evaluation ofthe seismic vulnerability of one ofthe most important historical monuments ofthe city: the church of N. Sra. do Carmo, built in the 16th century, destroyed by the 1755 Lisbon earthquake was partially rebuilt afterwards. To characterize the dynamic behaviour ofthis structure we developed a Finite Element Model (FEM) and performed a set of ambient vibration tests. In this paper we present a preliminary FEM of the structure and its comparison with experimental data. This work was developed in the framework of project CARAVELA - Instituto Politécnico de Lisboa. INTRODUCTION The city ofLagos (SW Portugal) is located in an area ofmoderate seismic risk. In historie times the city was severely damaged by large earthquakes and, for the well known 1755.11.01 Lisbon event, the macro seis- mie intensity in Lagos was estimated as IX-X, similar to what was observed in the city of Lisbon. Histori- cal documents report earthquakes in Lagos at 382 BC, 1356, 1504, 1719, 1722. More recently, the 1969.02.28 event, located south of the Gorringe Bank, with a magnitude near 7.9, caused considerable damage to this city. Instrumental seismicity close to Lagos is not yet well characterized, in spite of the significant develop- ment of the Portuguese seismic network in Algarve, which is now able to locate a large number of small magnitude events in the southern Portuguese shelf. Recently the local authorities devoted particular attention to preserve and rebuild the main historical buildings of the town centre. In this study we present the preliminary results on dynamic characterization of the church of N" Sr" do Carmo or "Igreja das Freiras" ("nuns chur ch", see Figure 1). The church was built in the 16th century and was partially destroyed by the catastrophic earthquake of 1755 . 11 .0 I - known as the Lisbon earthquake. It was Figure I. View ofthe church ofNa sa do Carmo. Figure 2. Frontal view. partially rebuilt after that event but since then no significant retrofit was perforrned. The municipality of Lagos approved recently a major rehabilitation project for this monumento The works will take 18 month s, di vided in 3 stages: repair 483

Transcript of Ambient vibration testing at N. Sra. do Carmo Church ... · Ambient vibration testing at N. Sra. do...

Structural Analysis of Historical Constructions - Modena, Lourenço & Roca (eds) © 2005 Taylor & Francis Group, London, ISBN 04 1536 379 9

Ambient vibration testing at N. Sra. do Carmo Church, preliminary results

M.A. Baptista, P. Mendes, A. Afilhado & L. Agostinho ISEL - Instituto Superior de Engenharia de Lisboa, Lisboa, Portugal

S. Lagomarsino Dipartimento di Ingegneria S/rutturale e Geotecnica - Università di Genova, l/alia

L. Mendes Victor Centro de Geofisica da Universidade de Lisboa, Lisboa, Portugal

ABSTRACT: The city of Lagos (Algarve, Portugal) is located in an area of modera te seismic risk. In recent years particular attention has been devoted to the historical centre of the city. The focus of this study is the evaluation ofthe seismic vulnerability of one ofthe most important historical monuments ofthe city: the church of N. Sra. do Carmo, built in the 16th century, destroyed by the 1755 Lisbon earthquake was partially rebuilt afterwards. To characterize the dynamic behaviour ofthis structure we developed a Finite Element Model (FEM) and performed a set of ambient vibration tests. In this paper we present a preliminary FEM of the structure and its comparison with experimental data. This work was developed in the framework of project CARAVELA -Instituto Politécnico de Lisboa.

INTRODUCTION

The city ofLagos (SW Portugal) is located in an area ofmoderate seismic risk. In historie times the city was severely damaged by large earthquakes and, for the well known 1755.11.01 Lisbon event, the macro seis­mie intensity in Lagos was estimated as IX-X, similar to what was observed in the city of Lisbon. Histori­cal documents report earthquakes in Lagos at 382 BC, 1356, 1504, 1719, 1722. More recently, the 1969.02.28 event, located south of the Gorringe Bank, with a magnitude near 7.9, caused considerable damage to this city.

Instrumental seismicity close to Lagos is not yet well characterized, in spite of the significant develop­ment of the Portuguese seismic network in Algarve, which is now able to locate a large number of small magnitude events in the southern Portuguese shelf.

Recently the local authorities devoted particular attention to preserve and rebuild the main historical buildings of the town centre.

In this study we present the preliminary results on dynamic characterization of the church of N" Sr" do Carmo or "Igreja das Freiras" ("nuns church", see Figure 1).

The church was built in the 16th century and was partially destroyed by the catastrophic earthquake of 1755.11 .0 I - known as the Lisbon earthquake. It was

Figure I. View ofthe church ofNa sa do Carmo.

Figure 2. Frontal view.

partially rebuilt after that event but since then no significant retrofit was perforrned.

The municipality of Lagos approved recently a major rehabilitation project for this monumento The works will take 18 months, divided in 3 stages: repair

483

and rehabilitation ofstructural elements (walls arches roof, vault); restoration of the interior archi~ectoni~ elements; restoration ofthe glazed ti le panels and guild woodwork.

2 THE STRUCTURE

Stone masonry churches are particularly vulnerable to earthquakes due to their configuration: open spaces, presence of thin walls, poor connect ions among the structural elements, and also due to the fact that the materiais used have low or no tension strength.

The central nave is covered with a barrei vault a structural fo rm quite common in religious Europe'an heri tage; a spherical dome is located over the high­chapel (Figure 4). The externai walls discharge directly on the soi!.

The structure has a rectangular plan and its dimen­sions are approximately 30 m by 7 m; its height varies from 9 m to Ii m. The highest point of the structure is the bell tower with 2 m by 2 m and 12.15 m height.

The interior of the church keeps some remarkable architectonic details: the arch of the High Chapel, the glazed tile panels, the throne, the tabernacJe and, in the fioor of the central nave, a considerable number of grave stones.

In the central nave we may observe a large fissure that runs along the barrei vault (cf. Figure 5). This

1 7,15

1 ~ 7,20 -!. 4,20 -1--1 1 ,70 --,--7,00---l

Figure 3. Plan ofthe church.

Figure 4. View of the dome.

crack may compromise safety of the structure in case of moderate or strong seismic activity.

3 AMBlENT VIBRATION TESTS

Dynamic analysis based on ambient vibration tests is a methodology used to characterize the dynamic behaviour of a structure excited by low amplitude vibrations. The information obtained can be useful to calibrate and update f inite element models ofthe struc­ture, or can be used for the health monitoring of the building.

Modal identification is the determination ofmodal parameters of existing built structures from exper­imentai data. Output-only modal identification of structures is normally used to identify modal parame­ters from the natural responses of many structures. In these cases the loads are unknown, and the identifi­cation process is carried out, based on the responses only. Many Applications on modem civil engineering structures can be found in I iterature (Ventura, 1996; Brincker, 2002). In Portugal several experiments were carried out, e.g. Ferreira (200 I), Baptista (2004). An example of ambient vibration testing on historical structures can be found in Shelley (2000).

The response of the structure must be recorded at specific locations in orderto maximize the information content with respect to the parameter identification. At this stage, the conditions inside the church do not permit a correct sensor location so the ambient vibra­tion tests performed were rather simple. A preliminary test was performed in order to obtain the first natural frequency for FEM updating (Figure 6).

The procedures for ambient vibration tests are described below.

The instruments used in this preliminary experi­ment were Kinemetrics force balance accelerometers (FBA ES-U); the sensors are connected to a 12 channel

Figure 5. Interior of the church - view ofthe vault.

484

~~ il==~~=::;;;;;:' I! ! I'

SENSORS lOCATlON

data acquisition system (DAS) with dynamic range of 114 dB, and output data format with 24 bits.

To obtain the fundamental frequency ofthe church, a preliminary test was made: acceleration data was recorded for 30 minutes and a sampling frequency of 250 Hz was used.

Figure 7 presents an acceleration record where we can see the amplitude of vibrations inside the church in ambient vibration conditions.

Figure 6, Location of sensors in this preliminary testo

Data processing was performed using Artemis Extractor version 3.4, SVBs (2002). The power spectral densities matrices are obtained using the Frequency Domain Decomposition. This technique decomposes approximately the spectral density matrix of the system response into a set of SDOF systems using Singular Value Decomposition; the singular val­ues are estimates of the spectral density of the SDOF systems, and the singular vectors are estimates of the mode shapes Ventura (2001).

:;; ~

0.00175

0.00075

Analysis of the Figure 8 enables the clear iden­tification of the first frequency (fundamental) of 3.3 Hz.

i 0.00025 4 THE FfNlTE ELEMENT MODEL ~

-0.00025

-0 .00175 L-~_~_~_~~_~,.-~_~_~-' o 1000 1200 1400 1600 1800 2000

TIme[.]

In order to obtain a finite element model ofthe struc­ture it is necessary to use idealisations of material behaviour and geometry. Geometry of ancient histori­cal structures may be rather complex as often there is no clear distinction between decora tive and structural elements (Lourenço, 2001).

Figure 7. Accelerogram recorded on March 2003 , inside the church.

The FEM has been developed using the compute r program SAP 2000 version 7.10 (Computers & Struc­tures, 1998). This program can be used for linear and

dB 1(1 .0 m/s"f 1Hz Frequency Domain Decomposttion - Peak Picking Average 01 lhe Normalized Singular Values 01

Spectral Dens~y Matrices 01 ali Data Sets.

~r----------,,---------r----------r-----------'

20

o

·20

·80

.BOO~--------~3~L--------~6~--------~g~----------i12

Frequency [Hz]

Figure 8. Singular values of spectral densities matrices using FDD.

485

Figure 9. Numerical model.

non-Iinear, static and dynamic analyses of a three­dimensional model of the structure. In our study the program has been used to determine the fundamen­tai frequencies and the corresponding mode shapes of structure, based on its physical and mechanical properties.

In the preliminary phase, it is asswned that ali the materiais of the structure have the following char­acteristics: (i) they are homogeneous and isotropic; (ii) modulus of Elasticity, E = 0.7 GPa (based on a preliminary ambient vibration test); (iii) Poisson's ratio 0.2; iv) linear elastic behaviour is assumed. The geometry of the structure is idealised considering the structure to be made of shell elements.

Figure lO shows the 1st mode shape, obtained by the fini te element modelling; it corresponds to a trans­lation in the transversal direction, with a frequency of 3.3 Hz. The modulus of elasticity has been tuned in order to identify the model from the experimental result.

The second mode shape, obtained from the FEM, involves mainly the bell tower (see Figure li).

5 FINAL CONSlDERATlONS

In order to obtain an accurate estimate of natural fre­quencies and mode shapes the sensors should be placed at the base of the vault see Figure 12. Those tests are scheduled for July 2004 as they require lhe availabil­ity of an auxiliary structure that will be used during rehabilitation works.

(a)

Figure 10. I si mode shape, with a frequency of 3.3 Hz; (a) 3D view; (b) plan view.

Figure 11. 2nd mode shape, with a frequency of 3.99 Hz.

From the structural point ofview, the ma in concern of the rehabilitation process is the presence of longi­tudinal cracking along the vault. In order to study this effect, one possibility is the simulalion of the crack, by considering a reduction of the modulus of Elastic­ity in the vertical direction. This hypothesis will be verified experimentally with a new sensors location (see Figure 12).

Procedures to evaluate dynamic behavior, ofhistor­ical structures, under seismic load are not yet quite well established so any quantitative information on

486

LJ. /r' ",~ / 0 ·\ . A~/'~~I~

~ ~ . SEllSO", LOCAnON I I SéNSO", LOCAnON

(a) Sensors Location

(b)

Figure 12. Future sensors location, in order to obtain mo de shapes: (a) transversal section; (b) plan section.

their dynamic properties is important to understand and to infer their seismic performance.

6 THE SEISMIC BEHAVIOUR

The damage observation to historical monurnents after strong earthquakes has shown that churches are one of the most vulnerable typologies among ancient structures (Lagomarsino and Podestà, 2003a). This is due to the presence of slender walls , wide halls (the nave), vaults oflarge span, etc.

Learning from the past earthquakes, a methodol­ogy has been established for the seismic vulnerability analysis, which is able to forecast the possible collapse mechanisms in churches (Lagomarsino and Podestà, 2003b). However, the method may be used for a risk analysis at territorial leveI, conducted on ali the churches in a Region, in order to obtain a list of vul­nerability and decide on the priorities for mitigation strategies.

In the case of a single church, a statistical approach is not suitable and a structural analysis should be per­formed. To this aim, some simplified approaches can be adopted, which consider the theorems oflimit anal­ysis for masonry structures (Lagomarsino et aI. 2004).

However, in this case the availability of ambient vibration tests and the possibility of building up a detailed finite element model (thanks to a detailed geo­metric survey), suggest us to perform a detailed non linear finite element analysis.

In order to evaluate the seismic performance of a masonry structure, identified by ambient vibrations, a proper non linear constitutive model is necessary for the masonry material. The one developed by Lagomarsino and Calderini (2004) will be adopted; it considers the masonry as a composite material, made by blocks and mortarjoints, in which various microme­chanical fa ilure mechanisms (openings, sliding with friction, crushing, etc.) are homogenized in a contin­uum damage model, suitable for the finite element procedure.

Pushover static non linear incrementaI analyses will be performed, in order to evaluate the capacity of the church in terms of total base shear and displace­ments. The seismic analysis uses the capacity spectrum method (Fajfar, 2000).

ACKNOWLEDGMENTS

The authors wish to thank Instituto Politécnico de Lisboa for financiaI support of project Caravela (50- 2003) and Architect Frederico Paula for his collaboration.

REFERENCES

Baptista, M.A. , Mendes, P., Campos Costa, A., Sousa Oliveira, C , Afilhado, A., Silva, P. 2004. Use of Ambient Vibration Test for Modal Evaluation of a 16 Floor Rein­forced Concrete Building In Lisbon, PortugaL ?roc. of 13th World Conference on Earthquake Engineering, Paper 1985, Vancouver, RC, Canada, August 1- 6.

Brincker, R. , Ventura, CE., Andersen, P. 2002. Why Output­Only Modal Testing is a Desirable Tool for a Wide Range of Practical Applications. online.

Computers & Structures Inc. 1998. SAP 2000 1ntegrated Finite Element Analysis and Design of Structures, Berkeley, California, USA .

Fajfar, P. 2000. A Nonlinear Analysis Method for Performance-Based Seismic Design. Earthquake Spectra, 16 (3), pp. 573- 592.

Ferreira, P. , Proença, JMFM 200 I. Identificação Modal do Corpo 4 do Hospital de Santa Maria, em Lisboa. Infer­ência da Vulnerabilidade Sísmica., SÍSMICA 2001 - jO

Encontro Nacional de Sislllologia e Engenharia Sísmica. Açores, 2001 , pp. 381 - 393 .

Lagomarsino, S., Calderini , C 2004. A Micromechanical Damage Model for Complex Masonry Structures. Proc. 13th 1nternational Brick and Block Masomy Conference, Amsterdam, July 4-7, pp. 1195- 1204.

Lagomarsino, S. , Podestà, S. 2004(a). Seismic vulnerability ofancient churches: I. Damage assessment and emergency planning. Earthquake Spectra, 20 (2), pp. 377-394.

Lagomarsino, S. , Podestà, S. 2004(b). Seismic vulnerability of ancient churches: 11. Statistical analysis of surveyed data and methods for ri sk analysis. Earthquake Spectra, 20 (2), pp. 395-412 .

487

Lagomarsino, S. , Podestà, S., Resemini, S. 2004. Obser­vational and mechanical models for the vulnerability assessment of monumental bu ildings. Prac. of lhe 13th World Conference on Earthquake Engineering, Paper 942, Vancouver, B.c., Canada, August 1-6.

Lourenço, P.B. 200 I. Analysis of historical constructions: From thrust-lines to advanced simulations. Historical Constructions, possibilities of numerical and experimen­ta/ techniques, Guimarães, pp. 91- 116.

Shelley, E.O. , Ordaz, M., Singh, S.K., Lemo, 1. 2000. Esti­mation of Seismic Hazard to Rehabilitate lhe Temple of La Compafíia in Puebla, Mexico. Proc. Inl. Millennium Congress - Archi 2000, Unesco.

SVBs, 2002. Slruclura/ Vibration Solutions ApS: ARTeMIS Exlraclor, Release 3.2, Users Manual. Denmark.

Ventura, C.E. , Schuster, N.D. 1996. Structural dynamic prop­erties of a reinforced concrete high-rise building during construction. Canadian Journal of Civil Engineering, 950- 972.

Ventura, C.E., Brincker, R., Dascotte, E., Andersen, P. 200 I . FEM Updating of the Heritage Courl Building Struc­ture. Proc. 19th Int. Modal Ana/ysis Conference (lMAC), Florida, USA , pp. 324-330.

488