Rel SERIES NTUA Final · seismic response of reinforced concrete (RC) frames with innovative solutions for masonry infill walls, considering both the in-plane and out-of-plane behaviour

SEVENTH FRAMEWORK PROGRAMME

Capacities Specific Programme

Research Infrastructures

Project No.: 227887

SERIES SEISMIC ENGINEERING RESEARCH INFRASTRUCTURES FOR

EUROPEAN SYNERGIES

Work package [WP9 – TA5 LNEC]

Assessment of innovative solutions for non‐load bearing masonry enclosures

‐ Final Report ‐

User Group Leader: Elizabeth Vintzileou Revision: Final

July, 2013

SERIES 227887 MASONRY ENCLOSURES Project

i

ABSTRACT

This document reports the outcomes of the research project “Masonry Enclosures” developed in

the framework of the transnational access (TA) to LNEC shake table within the FP7 project

SERIES. This TA project addresses the seismic performance of masonry enclosures in European

countries with moderate and high seismicity and consisted on the experimental evaluation of the

seismic response of reinforced concrete (RC) frames with innovative solutions for masonry infill

walls, considering both the in-plane and out-of-plane behaviour of the enclosures.

In order to ensure that in-plane and out-of-plane damage of masonry infill walls due to seismic

actions complies with the performance levels’ requirements, Eurocode 8 imposes the use of

reinforced solutions. Nevertheless, it does not provide any design rules or methodologies for

such reinforced masonry enclosures. An experimental programme was thus defined for assessing

the response of innovative solutions for non-load bearing masonry enclosures using LNEC’s

triaxial shake table. This experimental programme comprised the testing of one RC frame

building and four independent wall panels. Both a horizontal reinforcement in the bedding planes

of the masonry units and a reinforced mortar coating solutions were tested on single leaf clay

brick infill walls. Furthermore, a testing device for masonry infill panels was specifically

conceived for this project. A detailed description of the methods used is given and the

experimental results are shown and interpreted on the basis of the structural response and its

evolution with damage.

Keywords: Non-load bearing masonry enclosures, reinforced concrete frames, bed joint reinforcement, wire mesh coating reinforcement, shaking table test, innovative test setup


ii

ACKNOWLEDGMENTS

The research leading to these results has received funding from the European Union Seventh

Framework Programme [FP7/2007-2013] under grant agreement n° 227887 [SERIES].


iii

REPORT CONTRIBUTORS

LNEC António Araújo Correia

Paulo Xavier Candeias

Alfredo Campos Costa

Ema Coelho

NATIONAL TECHNICAL UNIVERSITY Elizabeth Vintzileou

OF ATHENS Vasiliki Palieraki

UNIVERSITY OF MINHO Paulo B. Lourenço

João Leite


iv


v

CONTENTS

1 Introduction .............................................................................................................................1

2 Description of the models and construction technique ............................................................4

2.1 Previous experimental studies ........................................................................................4

2.2 Previous tests at LNEC ...................................................................................................7

2.2.1 Results of model 1 ..............................................................................................7

2.2.2 Results of model 2 ............................................................................................25

2.3 Design and construction of the models .........................................................................37

2.3.1 Building model .................................................................................................37

2.3.2 Wall panels .......................................................................................................47

3 The LNEC Earthquake Engineering testing facility ..............................................................50

3.1 General information on the laboratory .........................................................................51

3.2 The facility: LNEC-3D Shaking Table .........................................................................51

3.3 General information on the shaking table .....................................................................53

3.4 Shaking table description ..............................................................................................53

3.5 Characteristics of the control system ............................................................................54

3.6 Complementary facilities ..............................................................................................55

4 Sensors technical data ............................................................................................................56

4.1 Displacement transducers .............................................................................................56

4.1.1 LVDT displacement transducers ......................................................................56

4.1.2 Hamamatsu optical system ...............................................................................57

4.1.3 Krypton K600 camera ......................................................................................58

4.2 Accelerometers .............................................................................................................60

4.3 Load cells ......................................................................................................................61

4.4 Acquisition system .......................................................................................................61

5 Test setup ...............................................................................................................................63


vi

5.1 Building model test setup .............................................................................................63

5.2 Wall panels test setup ...................................................................................................69

6 Seismic testing protocol ........................................................................................................76

6.1 Testing procedure .........................................................................................................76

6.2 Shaking table tuning .....................................................................................................76

6.3 Seismic test sequence ...................................................................................................79

7 Signal generation procedure for the shaking table tests ........................................................80

7.1 Building model .............................................................................................................80

7.2 Wall panels ...................................................................................................................83

8 Identification technique .........................................................................................................86

8.1 White noise ...................................................................................................................86

8.2 Impulse signal ...............................................................................................................87

9 Analysis of results .................................................................................................................89

9.1 Building model test results ...........................................................................................89

9.1.1 Initial test results ..............................................................................................89

9.1.2 Complementary test results ............................................................................100

9.1.3 Comparison of test results ..............................................................................109

10 Main conclusions .................................................................................................................119

References ....................................................................................................................................121


vii

List of Figures

Figure 1.1: Reduced scale RC building .......................................................................................... 2

Figure 1.2: Details of reinforced mortar coating ............................................................................ 3

Figure 1.3: Geometry of the model ................................................................................................. 3

Figure 1.4: Wall panels test setup ................................................................................................... 3

Figure 2.1: Test setup in [1]: (a) in-plane test; (b) out-of-plane test ............................................... 4

Figure 2.2: Model detailing in [2] and [4]: (a) infill wall solutions; (b) detail of the mortar

coating reinforcement ..................................................................................................................... 5

Figure 2.3: Testing setup and models [17]: (a) in-plane test; (b) out-of-plane test; (c) model 1; (d)

model 2; (e) model 3 ....................................................................................................................... 6

Figure 2.4: Position and label of the accelerometers in model 1 .................................................... 7

Figure 2.5: Crack patterns of the exterior leaf of model 1 after stage 3 (2475 YRP) (Note: the

drawn lines on the RC frame represent damage on the clay bricks applied to the RC frame to

avoid thermal bridges) .................................................................................................................... 9

Figure 2.6: Crack patterns of the interior leaf of model 1 after stage 3 (2475 YRP) (Note: the

drawn lines on the RC frame represent damage on the clay cricks applied to the RC frame to

avoid thermal bridges) .................................................................................................................... 9

Figure 2.7: Stage 4 (4574 YRP) of the shaking table test of model 1: (a) out-of-plane collapse of

the exterior leaf of the infill wall at the ground floor of the South façade; (b) out-of-plane

collapse of the interior leaf of the infill wall at the ground floor of the South façade; (c) out-of-

plane collapse of the exterior jambs of the infill walls at the first storey of the East façade; (d)

out-of-plane collapse of both leaves of the infill wall at the ground storey of the North façade; (e)

model 1 after the fourth stage, collapsed and without all the infill walls of the ground floor; (f)

ground floor column of the Northwest after collapsing at the top and disintegration up to mid-

height; (g) plastic hinge developed on the top of the ground RC column of the Northeast corner;

(h) barely damaged infill wall at the first storey of the South façade. .......................................... 11

Figure 2.8: Plastic hinge formation at the top of the ground floor columns ................................. 12

Figure 2.9: Mode shapes of the DI 0 of model 1 (initial dynamic identification test) .................. 15


viii

Figure 2.10: Frequency change along loading stages: (a) variation of the FRF’s along the test of

model 1 at the accelerometer BNE – 1T; (b) evolution of the frequencies along the test of model

1 and their final variation in respect to DI 0 ................................................................................. 15

Figure 2.11: Seismic vulnerability curves of model 1 in the transverse and longitudinal

directions, using the PGA and the Input Energy as input. Here, the damage indicator is a measure

of the frequency change ................................................................................................................ 17

Figure 2.12: Evolution of the frequencies of the infill walls in the North and South façades along

the test of model 1 and their final variation in respect to DI 0. .................................................... 19

Figure 2.13: Interstorey displacements and drifts of model 1 ...................................................... 20

Figure 2.14: Recorded PGA and amplifications at the infill walls and at the RC structure for each

test stage of model 1 ..................................................................................................................... 22

Figure 2.15: Out-of-plane deformation of the North and South infill walls along the tests of

model 1.......................................................................................................................................... 24

Figure 2.16: Out-of-plane PGD of the East and West infill walls of model 1 .............................. 25

Figure 2.17: Position and label of the accelerometers in model 2 ................................................ 25

Figure 2.18: Crack patterns of model 2 after stage 3 (2475 YRP) (Note: the drawn lines on the

RC frame represent damage on the rendering applied to the RC frame) ...................................... 27

Figure 2.19: Crack patterns of model 2 after stage 4 (4574 YRP) (Notes: the drawn lines on the

RC frame represent damage on the rendering applied to the RC frame. The blue lines developed

after stage 3) .................................................................................................................................. 27

Figure 2.20: Damage in model 2 after the fourth stage (4574 YRP): (a) North façade; (b) South

façade; (c) West façade from the inside; (d) detail of the left jamb of the door on the North

façade and a horizontal crack at mid-height of the Northeast corner column; (e) horizontal crack

at mid-height of the Southwest corner column; (f) heavily damaged top column-beam connection

of the Southwest corner column with loss of the concrete cover and rebar exposure .................. 28

Figure 2.21: Mode shapes of the DI 0 of model 2 (initial dynamic identification test) ................ 29


model 2 at the accelerometer BNE – 2T; (b) evolution of the frequencies along the test of model


Figure 2.23: Seismic vulnerability curves of model 2 in the transverse and longitudinal

directions, using the PGA and Input Energy as input ................................................................... 31


ix


the test of model 2 and their final variation in respect to DI 0 ..................................................... 32

Figure 2.25: Interstorey displacements and drifts of model 2 ...................................................... 33



Figure 2.27: Out-of-plane deformation of the infill wall at the ground level of the South façade

(mm) .............................................................................................................................................. 35

Figure 2.28: Out-of-plane deformation of the infill wall at the first storey of the South façade

(mm) .............................................................................................................................................. 36

Figure 2.29: Out-of-plane deformation of the infill wall at the first storey of the North façade

(mm) .............................................................................................................................................. 36

Figure 2.30: Out-of-plane PGD of the North, East and West infill walls (mm) ........................... 37

Figure 2.31: Prototype geometry (m)............................................................................................ 37

Figure 2.32: Geometry of the tested model reduced to a scale of 1:1.5 ....................................... 39

Figure 2.33: Response spectra after the application of the similitude law, see Table 2.2,

according to EC8 [29] ................................................................................................................... 41

Figure 2.34: Geometry of the openings in each façade: (a) North façade; (b) West façade; (c)

East façade; (d) South façade ........................................................................................................ 42

Figure 2.35: Additional steel masses for: (a) RC concrete structure, bolted to the slabs of the 1st

floor and roof with 82x82x26 cm and 12KN each; (b) Infill walls, bolted to both sides of the wall

with 15x15x4 cm and 0.072KN each ............................................................................................ 43

Figure 2.36: Construction of the models: (a) horizontally aligned surface on which the models

were constructed; (b) RC ring beam with steel connector with an eye in lift and transport the

model to the shaking table; (c) model 1 on the shaking table before the test ............................... 44

Figure 2.37: Single leaf clay brick infill walls with reinforced plaster from Model 3 already

scaled: (a) spacing of the Hilti X-M8H10-37-P8 connectors along the height of the RC column;

(b) detail at the RC column; (c) detail of the Hilti X-M8H10-37-P8 connectors ......................... 45

Figure 2.38: Construction of the infills of model 3: (a) Bekaert Armanet ϕ1.05mm 12.7x12.7mm;

(b) Hilti X-M8H10-37-P8; (c) application of the grid in the outer surface at a corner column; (d)

application of the grid in the inner surface; (e) additional masses with steel rings attached to the

infill walls ..................................................................................................................................... 46

Figure 2.39: Frame panel of typical RC building ......................................................................... 47


x

Figure 2.40: Reinforcement layout for the RC frames ................................................................. 47

Figure 2.41: Prestressing details ................................................................................................... 48

Figure 2.42: RC frame ready for infill construction ..................................................................... 48

Figure 2.43: Masonry infill construction with bed joint reinforcement ........................................ 48

Figure 2.44: Specimens ready for testing ..................................................................................... 49

Figure 3.1: LNEC Earthquake Engineering testing facility (Ferry Borges building) ................... 50

Figure 3.2: LNEC earthquake engineering testing room .............................................................. 52

Figure 3.3: LNEC-3D shaking table ............................................................................................. 52

Figure 4.1: LVDT displacement transducers (source: RDP) ........................................................ 56

Figure 4.2: HAMAMATSU optical 2D displacement transducer ................................................ 57

Figure 4.3: Hamamatsu system configuration [12] ....................................................................... 58

Figure 4.4: The Krypton K600 camera [19] ................................................................................. 58

Figure 4.5: Representative measurement volume of a Krypton K600 camera [19] ..................... 59

Figure 4.6: Accuracy zones of the Krypton K600 camera system [19] ........................................ 59

Figure 4.7: Endevco accelerometers ............................................................................................. 60

Figure 4.8: PCB Piezotronics accelerometers ............................................................................... 60

Figure 4.9: Instron load cells ........................................................................................................ 61

Figure 5.1: Accelerometers used in the shaking table tests: (a) piezoelectric from PCB [31], [32];

(b) piezoelectric from Wilcoxon [40]; (c) capacitive from Endevco [9] ...................................... 64

Figure 5.2: Accelerometers setup: (a) North and East façades; (b) South and West façades ....... 65

Figure 5.3: PCB Piezotronics accelerometers: (a) at the infill walls; (b) at the corners of the RC

slabs............................................................................................................................................... 65

Figure 5.4: Hamamatsu photonics c5949 [12]: (a) position of the Hamamatsu leds in the first

storey; (b) position of the Hamamatsu leds in the roof; (c) camera and led at the corner of the

structure; (d) infrared led; (e) controller ....................................................................................... 66

Figure 5.5: Acquisition and control room: (a) from top to bottom: NI-SCXI-1001, PCB

Piezotronics 481A02 and NI PXI-1052; (b) control room with the shaking table’s controls and

the model’s acquisition system ..................................................................................................... 67

Figure 5.6: K600 Krypton camera: (a) three CCD cameras; (b) Space Probe used to calibrate the

initial geometrical plan of the LED’s; (c) acquisition control; (d) distribution of the LED’s along

the infill wall on the upper floor of the North façade ................................................................... 68

Figure 5.7: Main steel caisson frames of TIM (construction phase) ............................................ 69


xi

Figure 5.8: Base columns of main steel frame with detail of bolted connection to the shake table

(construction phase) ...................................................................................................................... 69

Figure 5.9: Guiding system of RC frame upper beam and rollers for longitudinal motion

(construction phase) ...................................................................................................................... 70

Figure 5.10: Pyramidal support for strut connection between the RC frame and the reaction wall

....................................................................................................................................................... 71

Figure 5.11: Hinged base supports for the RC frame specimens (construction phase) ................ 71

Figure 5.12: Hinged base and pyramidal supports on their final position .................................... 71

Figure 5.13: Assembly of TIM components on the shaking table ................................................ 72

Figure 5.14: Positioning of TIM over the wall panel setup .......................................................... 72

Figure 5.15: Complete setup for wall panels tests ........................................................................ 72

Figure 5.16: Schematic representation of the finite element models used for the design of TIM,

taking into account (right) or not (left) the wall panels ................................................................ 73

Figure 5.17: Vibration modes of TIM without the wall panel contribution: a) longitudinal

(f = 19.9 Hz); b) transverse (f = 33.8 Hz) ..................................................................................... 73

Figure 5.18: Vibration modes of TIM with the wall panel contribution: a) longitudinal

(f = 18.4 Hz); b) transverse (f = 23.1 Hz); c) torsional (f = 25.5 Hz) ........................................... 73

Figure 5.19: Out-of-plane wall panel deformation monitoring with Krypton K600 camera ........ 74

Figure 5.20: Video camera and target points for data image correlation measurement of in-plane

deformations at one RC frame node ............................................................................................. 74

Figure 5.21: Hamamatsu setup for measuring the horizontal translations of the RC frame nodes

....................................................................................................................................................... 75

Figure 5.22: Accelerometer setup for RC frame out-of-plane vibration measurements ............... 75

Figure 5.23: Load cells for strut reaction measurement ............................................................... 75

Figure 6.1: Shaking table tuning application: definition of parameters ........................................ 77

Figure 6.2: Shaking table tuning application: FRF obtained ........................................................ 77

Figure 6.3: Signal tuning iterative process ................................................................................... 78

Figure 6.4: Calibration of the input signals with masses attached to the shaking table ............... 78

Figure 7.1: Comparison between pseudo-acceleration response spectra of the accelerograms

generated and the response spectra, already scaled following the similitude law of Cauchy-

Froude, obtained from EC8 [30] ................................................................................................... 81


xii

Figure 7.2: Time histories of the input signal of stage 2 (SD 475 YRP) reduced at 1:1.5 scale

using Cauchy-Froude’s similitude law (see Table 2.2) ................................................................ 82

Figure 7.3: Comparison between pseudo-acceleration response spectra of the accelerograms

generated and the response spectra obtained from EC8 [30] ........................................................ 83

Figure 7.4: Representative building model for wall panel input time-history definition ............. 83

Figure 7.5: Longitudinal modes of vibration 1 (1.35Hz) and 2 (4.28Hz) ..................................... 84

Figure 7.6: Transverse modes of vibration 1 (2.78Hz) and 2 (10.94Hz) ...................................... 84

Figure 7.7: Interstorey drift time-history for in-plane motion ...................................................... 85

Figure 7.8: Absolute acceleration time-history for out-of-plane motion ...................................... 85

Figure 8.1: White noise signals for dynamic identification tests .................................................. 87

Figure 9.1: Position and label of the accelerometers in model 3 .................................................. 89

Figure 9.2: Crack patterns of model 3 after stage 2 (475 YRP) (Note: the drawn lines on the RC

frame represent damage on the mortar rendering applied to the RC frame) ................................ 90

Figure 9.3: Crack patterns of model 3 after stage 3 (2475 YRP) (Note: the drawn lines on the RC

frame represent damage on the mortar rendering applied to the RC frame) ................................ 91

Figure 9.4: Damage in model 3 after stage 3 (2475 YRP): (a) infill wall at the ground floor of the

North façade; (b) infill wall at the upper floor in the East façade; (c) damaged mortar rendering

at the Southeast corner; (d) damaged mortar rendering at the Southwest corner ......................... 91

Figure 9.5: Mode shapes of the DI 0 of model 3 (initial dynamic identification test) .................. 93


model 3 at the accelerometer BNE – 1L; (b) evolution of the frequencies along the test of model


Figure 9.7: Seismic vulnerability curves of model 3 in the transverse and longitudinal directions,

using the PGA and Input Energy as input ..................................................................................... 94


the test of model 3 and their final variation in respect to DI 0 ..................................................... 95

Figure 9.9: Interstorey displacements and drifts of model 3 ........................................................ 96



Figure 9.11: Out-of-plane deformation of the infill wall at the ground level of the South façade of

model 3 in mm .............................................................................................................................. 99


xiii

Figure 9.12: Out-of-plane deformation of the infill wall at the first storey of the South façade of

model 3 in mm .............................................................................................................................. 99

Figure 9.13: Out-of-plane deformation of the infill wall at the first storey of the North façade of

model 3 in mm. ........................................................................................................................... 100

Figure 9.14: Out-of-plane PGD of the North, East and West infill walls of model 3 in mm ..... 100

Figure 9.15: Damage in model 3B after stage 3 (2475 YRP): (a) North façade; (b) South façade;

(c) crack and mortar rendering loss at the Northwest corner; (d) crack and mortar rendering loss

at Northeast corner; (e) crack at the a lateral jamb in the infill wall at the ground floor of the East

façade; (f) crack at the interior jambs in the infill wall at the ground floor of the East façade .. 102

Figure 9.16: Damage in the infill walls and RC structure after the reinforced rendering removal

at the ground floor: (a) infill wall of the North façade; (b) South infill wall with a compression

crush at right down corner; (c) gap between one of the West the infill wall and RC frame in the

West wall; (d) infill walls of the West façade; (e) extensive cracking at the upper column-beam

connection in the Northwest corner ............................................................................................ 103

Figure 9.17: Evolution of the frequencies along the test of model 3B, and their final variation in

respect to DI 0 of model 3, at the RC structure and infill walls in South façade and ground level

of the North façade...................................................................................................................... 104

Figure 9.18: Interstorey displacements and drifts of model 3B .................................................. 105


test stage of model 3B ................................................................................................................. 106

Figure 9.20: Out-of-plane deformation of the infill wall at the ground level of the South façade of

model 3B in mm.......................................................................................................................... 107

Figure 9.21: Out-of-plane deformation of the infill wall at the first storey of the South façade of

model 3B in mm.......................................................................................................................... 108

Figure 9.22: Out-of-plane deformation of the infill wall at the first storey of the North façade of

model 3B in mm.......................................................................................................................... 108

Figure 9.23: Out-of-plane PGD of the North, East and West infill walls of model 3B in mm .. 109

Figure 9.24: Longitudinal direction target/acquired comparison (Fourier filter: 1-40Hz): (a)

PGA; (b) Root Mean Square Acceleration; (c) Arias Intensity: (d) Input Energy ..................... 112

Figure 9.25: Transverse direction target/acquired comparison (Fourier filter: 1-40Hz): (a) PGA;

(b) Root Mean Square Acceleration; (c) Arias Intensity: (d) Input Energy ............................... 112


xiv

Figure 9.26: Vulnerability curves of the 1st mode in each main direction of the RC structure of

the three tested models ................................................................................................................ 113

Figure 9.27: Average vulnerability curves of the North and South infill walls of the three tested

models ......................................................................................................................................... 114

Figure 9.28: Interstorey displacements of the three tested models in the transverse and

longitudinal directions ............................................................................................................... 116

Figure 9.29: Interstorey drifts of the three tested models in the transverse and longitudinal

directions ..................................................................................................................................... 116

Figure 9.30: Average recorded PGA and amplifications at the infill walls and at the RC structure

for each test stage of all tested models ........................................................................................ 118


xv

List of Tables

Table 2.1 - Experimental damping ratio of model 1 ..................................................................... 16

Table 2.2 - Cauchy-Froude similitude law ................................................................................... 39

Table 2.3 - Design loads of the models already reduced at scale of 1:1.5 .................................... 40

Table 3.1 – Name and location of the Laboratory ........................................................................ 51

Table 3.2 – Name and location of the parent organization ........................................................... 51

Table 3.3 – Name of the LNEC-3D shaking table ........................................................................ 53

Table 3.4 – Type of shaking table ................................................................................................. 53

Table 3.5 – Characteristics of the Platform .................................................................................. 53

Table 3.6 – Characteristics of the Actuators ................................................................................. 53

Table 3.7 – Shaking table performances ....................................................................................... 54

Table 3.8 – Characteristics of the analogue part ........................................................................... 54

Table 3.9 – Characteristics of the digital part ............................................................................... 54

Table 4.1 – Characteristics of the RDP displacement transducers ............................................... 57

Table 4.2 – Characteristics of the HAMAMATSU displacement transducers ............................. 57

Table 4.3 – Characteristics of the Endevco accelerometers ......................................................... 60

Table 4.4 – Characteristics of the PCB accelerometers ................................................................ 60

Table 4.5 – NI PXI controller ....................................................................................................... 61

Table 4.6 – NI PXI chassis ........................................................................................................... 62

Table 6.1 - Shaking table test procedure for the building model .................................................. 79

Table 9.1 - Experimental damping ratios of model 3 ................................................................... 94


1

1 Introduction

The masonry enclosures project focused on the seismic performance assessment of traditional

and innovative masonry enclosures of European countries with moderate to high seismicity

(Greece, Portugal, Italy and Slovenia). A complete redesign of the experimental program was

undertaken in order to potentiate the goals of the project which were to characterize the

seismic behaviour of different types of traditional and immediate future systems for masonry

enclosures.

In fact, the recent L’Aquila earthquake of 2009 has reminded seismic engineers that the

current masonry infill solutions are not effective, as illustrated by the considerable in-plane

damage and out-of-plane collapses verified in a common basis throughout the affected areas.

Eurocode 8 addresses this issue by imposing the use of reinforced infill solutions but fails to

give design and detailing methodologies.

With the above in mind, a shake table experimental research programme was devised in order

to investigate the seismic behaviour of reinforced infill walls and how they affect the global

structural response.

Four types of masonry enclosures were intended to be tested within this TA project:

i) Unreinforced masonry

ii) Horizontal layers of reinforcement between masonry units

iii) Reinforced mortar coating

iv) Both a reinforced mortar coating and horizontal layers of reinforcement between

masonry units

The first phase of the research activity involved the seismic testing of a two-storey reinforced

concrete (RC) infilled frame building designed to the Eurocodes and built at a 1:1.5 scale. As

a follow up of previous tests performed at LNEC using RC buildings either with double leaf

unreinforced masonry infills or with single leaf masonry with bed joint reinforcement, this

reduced scale model was built with single leaf clay bricks and reinforced mortar coating, as

shown in Figure 1.1 to Figure 1.3. Wire mesh reinforcement was placed on both sides of the


2

infill walls and anchored to the RC frame and masonry units. Additional masses were

attached to the walls in order to respect Cauchy-Froude similitude requirements. From these

tests it was possible to assess the evolution of the seismic behaviour of infills and their

influence on the RC structure through several acceleration inputs of increasing amplitude,

associated to cumulative damage limit states.

The second phase of this transnational access activity comprised the dynamic testing of four

specimens of a closed RC plane frame with external dimensions of 6.40mx3.25m. This plane

frame was tested simultaneously for in-plane and out-of-plane dynamic actions, representing

the response of a typical upper storey frame panel of a RC building. Both motions match

realistic floor response spectra, of narrow band frequency content. The in-plane motion

enforces an inter-storey drift time-history on the frame by restraining the upper beam and by

imposing the displacement of the shaking table on the lower beam. On the other hand, the out-

of-plane motion consists on a rigid-body vibration of both the upper and lower beams,

reproducing the storey absolute accelerations and thus inducing high-frequency inertia forces

perpendicular to the masonry panel and leading to a local vibration of the infill wall.

This unique testing setup (Figure 1.4) was specifically designed for this test and is mainly

composed of a stiff steel caisson three-dimensional frame which moves rigidly with the

shaking table. It is fixed to the upper beam in the transversal direction, while a system of

rollers allows for an independent motion in the longitudinal direction.

Figure 1.1: Reduced scale RC building


3

Figure 1.2: Details of reinforced mortar coating

Figure 1.3: Geometry of the model

Figure 1.4: Wall panels test setup


4

2 Description of the models and construction

technique

2.1 PREVIOUS EXPERIMENTAL STUDIES

The interaction between the infill wall and the surrounding frame has been studied by several

authors since the 1960’s (see [13]; [15]; [16]; [23]; [24]; [25]; [39]; [40]). Given the

objectives of the present work, more recent studies that relate the in-plane and out-of-plane

damage, or that use dynamic actions are of higher relevance. In [1], the out-of-plane

behaviour of RC frames with infill walls, after damaging the frame in-plane by applying a

horizontal load, was evaluated. Eight RC frames were tested during the experimental

program, at a 1:1 scale, using the test setup in Figure 2.1, and the main conclusions were: (i)

the in-plane stiffness of the system is directly proportional to the compressive strength of the

masonry, and it drops significantly after the first crack; (ii) the out-of-plane capacity depends

on the slenderness of the wall and on the compressive strength of the masonry; (iii) for large

slenderness, the out-of-plane capacity of the infill wall decreases after being damaged in-

plane.

(a) (b)

Figure 2.1: Test setup in [1]: (a) in-plane test; (b) out-of-plane test


5

In the work described in [10], the authors tested masonry infill walls confined in metallic

frames for in-plane, out-of-plane and combined actions with the objective of understanding

the seismic behaviour of damaged infills. The authors also intended to identify the least

favourable combination of actions, for which reason the combined actions test was performed

in different sequences. The in-plane test was carried out by imposing cyclic incremental

displacements. The main conclusions were: (i) damaged infill walls have a large out-of-plane

capacity in spite of presenting larger deformations when compared to undamaged ones; (ii)

the out-of-plane damage in the infill walls does not affect the maximum in-plane compressive

strut capacity.

In [2] and [4], the authors tested RC frames with both reinforced and unreinforced infill walls

loaded initially in-plane and then out-of-plane, with the objective of assessing the seismic

capacity of different reinforcement solutions. The frames were built at 1:1 scale, with in-plane

dimensions of 4.2mx3.0m, according to the solutions described in Figure 2.2.

(a) (b)

Figure 2.2: Model detailing in [2] and [4]: (a) infill wall solutions; (b) detail of the mortar coating reinforcement

The test plan for each frame consisted in the application of a vertical load to the columns, kept

constant during the test in order to simulate the load transferred by the upper storeys, followed

by a cyclic in-plane drift, from 0.1% to 3.6%. The infill wall was then loaded out-of-plane

with a monotonic load applied in four different points. The test results led to the following

conclusions: (i) the reinforcement reduces the in-plane damage but it does not considerably

increase the stiffness, when compared to the unreinforced solution; (ii) damaged unreinforced

infill walls needed an acceleration five times lower to collapse out-of-plane, when compared


6

to the undamaged infill wall; (iii) the presence of reinforcement increases the out-of-plane

stiffness of the infill wall, hence smaller displacements were recorded.

In the experimental campaign presented in [17], the author tested RC frames with infill

masonry walls, at a 1:2 scale, by applying in-plane cyclic horizontal loads to the frame

followed by out-of-plane accelerations imposed by a uniaxial shaking table, as represented in

Figure 2.3. The objective was to understand the combined seismic behaviour of a simple and

slender solution (model 1), a solution with an RC lintel and column at mid-span (model 2)

and a solution with a more robust RC frame (model 3). The conclusions of the work were: (i)

the more slender models presented the highest inertial force at mid-height while the more

robust model presented the highest value at the top; (ii) the additional RC members in model

2 improved its seismic behaviour by reducing the out-of-plane displacements and through a

slower crack spreading process; (iii) the out-of-plane collapse is dependent not only on the

corresponding inertial forces but also on the excessive out-of-plane displacements.

(a) (b)

(c) (d) (e)

Figure 2.3: Testing setup and models [17]: (a) in-plane test; (b) out-of-plane test; (c) model 1; (d) model 2; (e) model 3


7

2.2 PREVIOUS TESTS AT LNEC

Two building models, similar to the one included in this project, were previously tested at the

Earthquake Engineering and Structural Dynamics Division (NESDE) of the National

Laboratory for Civil Engineering (LNEC). Their description and test results are described in

the following sections.

2.2.1 Results of model 1

Model 1 was designed following the pre-Eurocode normative, RSA [36] and REBAP [35],

using the most commonly used concrete and steel for rebars (C20/25 and S400, respectively),

together with double leaf, unreinforced, clay brick infill walls. Hence, model 1 represents the

built heritage in Portugal for the last three decades.

The following results were obtained using the acquisition equipment described in Chapter 4:

data recordings of the tests (quantitative results) and damage maps drawn between each of the

test stages (qualitative results). Figure 2.4 presents the position and label of the

accelerometers in model 1. Since the clay brick infills have two leaves, a set of accelerometers

was placed in the interior leaf at the exact same position of the exterior accelerometers seen in

Figure 2.4, for comparison purposes. The label of the interior accelerometers was obtained by

replacing the E with an I.

North South East West

Figure 2.4: Position and label of the accelerometers in model 1

NE1 1

BNE 2L

NE2 2

NE1 2

INP L

NE1 3

NE2 3

BNE 1L

NE2 1

SE1 3

SE2 2

BSW 1L

SE1 1

SE2 1 SE2 3

SE1 2

BSW 2L

EE1.21

EE1.11

BNE 1T

EE2.2 2

BNE 2T

EE2.1 1

EE2.2 1

WE1.1 1

BSW 2T

BSW 1T

WE1.2 1

INP T


8

Overall damage and crack patterns

The test procedure followed in the present work, which is the input of four seismic actions of

increasing amplitude, leads to damage accumulation. The evolution of the damage can be a

strong indicator of the collapse mechanism developed by the structure, especially in model 1

due to its collapse during the last stage of the shaking table test. Furthermore, the analysis and

relation of the crack patterns with the quantitative results, obtained from the data acquisition

equipment as accelerometers and displacement measurement cameras, will increase the

reliability of conclusions.

Even though the model was transported to the shaking table using a crane, the model did not

present any noticeable damage before the first stage of the shaking table test. After the first

two stages (225 and 475 YRP), model 1 did not present any visible damage, which is not in

agreement with the dynamic data that shows a small decrease in the model frequencies. This

loss of stiffness can be related to two aspects: i) the separation of the infill walls from the

reinforced concrete (RC) frame, a damage that is difficult to detect and is camouflaged by the

mortar rendering of the infill walls; ii) cracks and micro cracks in the RC frame that remain

undetected due to the clay bricks applied externally to avoid thermal bridges. As expected,

after the third stage (2475 YRP), the model presented clear cracks on both leaves of the infill

walls, see Figure 2.5 and Figure 2.6, mainly at the ground storey of the North, East and West

façades. The infill wall at the first storey of the North façade also presented some cracks. The

cracks appeared mainly at the connection between the infill wall and the RC frame, and at the

corners of the openings and moving towards the RC frame. In the infill walls at the ground

floor in the East and West façade, and on both leaves, the crack pattern around several

opening jambs is clear, separating them from the RC frame and the section of the infill wall

below the openings. This damage is related to the in-plane displacements of the RC frames in

the North-South (longitudinal) direction. Associated with this damage, the frequencies of the

first three mode shapes decreased 13.6%, 28.4% and 20.2%, respectively, in comparison to

the undamaged state.


9

North South

East West

Figure 2.5: Crack patterns of the exterior leaf of model 1 after stage 3 (2475 YRP) (Note: the drawn lines on the RC frame represent damage on the clay bricks applied to the RC frame to avoid thermal bridges)

North South

East West

Figure 2.6: Crack patterns of the interior leaf of model 1 after stage 3 (2475 YRP) (Note: the drawn lines on the RC frame represent damage on the clay cricks applied to the RC frame to avoid thermal bridges)

Model 1 collapsed during the fourth and last stage (4574 YRP), after losing the infill walls,

see Figure 2.7 (a) to (d), with subsequent failure of the three RC columns at the ground storey

of the West façade, see Figure 2.7 (e) and (f). The collapse mechanism developed, designated


10

here by soft storey, is characterized by the concentration of plastic hinges at the columns of a

given storey, while the upper part remains rather stiff, Figure 2.7 (g), is very undesirable

during a seismic action since it commonly leads to the partial or total collapse of the structure,

as it was the case in the present test. A beam sway mechanism, where the plastic hinges are

developed at the beams and not at the columns, is more desirable since it dissipates the energy

transferred by the earthquake without compromising its stability [40]. The collapse of the

columns occurred at their top, in the RC node, see Figure 2.7 (g) and Figure 2.8, followed by

disintegration of the concrete and instability of the steel up to mid-height of the column. This

failure further stresses the need to adequately confine concrete in the nodes and the need to

add more stirrups to avoid shear failure. It seems that the concentration of damage and

deformation of the columns in the nodes is also forced by the stiff behaviour of the first storey

and, possibly, the presence of the masonry infills in the ground storey, before collapse.

Before the collapse of the structure, the central jambs at the first storey of the East façade

collapsed out-of-plane, see Figure 2.7 (c), followed by the infill wall at the ground storey of

the North façade, see Figure 2.7 (d). The exterior leaf of the infill wall at the ground floor of

the South façade and the infill walls at the ground floor of the East and West façade collapsed

out-of-plane simultaneously, see Figure 2.7 (a).


11

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)

Figure 2.7: Stage 4 (4574 YRP) of the shaking table test of model 1: (a) out-of-plane collapse of the exterior leaf of the infill wall at the ground floor of the South façade; (b) out-of-plane collapse of the interior leaf of the infill wall at the ground floor of the South façade; (c) out-of-plane collapse of the

exterior jambs of the infill walls at the first storey of the East façade; (d) out-of-plane collapse of both leaves of the infill wall at the ground storey of the North façade; (e) model 1 after the fourth stage,

collapsed and without all the infill walls of the ground floor; (f) ground floor column of the Northwest after collapsing at the top and disintegration up to mid-height; (g) plastic hinge developed on the top of the ground RC column of the Northeast corner; (h) barely damaged infill wall at the first storey of the

South façade.

All these infills collapsed with a rotation mechanism with a hinge line at their bottom support

or at the first masonry joint (as a cantilever). The interior leaf of the infill wall at the ground

floor of the South façade was the last infill to collapse, see Figure 2.7 (b). This infill collapsed


12

with three hinge lines (top, centre and bottom). Immediately after, the structure collapsed. The

jambs around the windows collapsed usually by rotating out-of-plane as a rigid body with a

hinge line close to the connection to the spandrel (either the support or the first masonry

joint), or the rest of the masonry, again as a cantilever.

Figure 2.8: Plastic hinge formation at the top of the ground floor columns


13

Modal frequencies of the RC structure

Model 1 was subjected to four dynamic identification tests, from DI0 (undamaged state) to

DI3 (after stage 3). The model collapsed during stage 4, therefore it was not possible to

perform the last dynamic identification test. The quality of the obtained results can be

measured by the coherence between the input and output signals, which should be close to 1,

and this was the case in all the DI tests. Clear peaks could also be identified in the FRF’s

(Frequency Response Functions).

Five mode shapes were identified in DI0, see Figure 2.9, namely: the first and second

transverse modes; the first and second longitudinal modes; the (first) torsional mode. As

expected, the first mode is transverse (East-West) at a frequency of 7.71Hz, as the RC frames

in that direction are single-bay and the total length of the model is smaller than in the

longitudinal direction. The second mode is longitudinal (North-South) at the frequency of

9.62Hz since the RC frames are double bay and the total length of the model is higher than

the transverse one. Due to influence of the infill walls, and the fact that the percentage of

openings is not the same in all façades, the first transverse and longitudinal modes have a very

small component in the longitudinal and transvers directions, respectively. As it can be seen

in Figure 2.10 (a), the first mode shape was clearly identified in the FRF.

The torsional mode has a frequency of 26.95Hz, considerably higher than the previous

identified modes. This increment in the global torsional stiffness is due to the infill walls,

otherwise it a frequency closer to the previous modes would be expectable. The frequency of

the mode was not as clear as the previous two modes in the FRF but still visible, while the

mode shape presents some incoherence. As it was stated above, the openings in the infills are

not symmetric, which leads to a deviation of the centre of mass from the centre of stiffness

[7], and in the present case it would be expectable that the centre of stiffness would be closer

to the Southeast corner. The mode experimentally detected presents a rotation around a point

closer to the Southwest corner. Similar problems were found in the detected torsional modes

of Models 2 and 3, and, hence, the problem can be associated to an undesirable interaction

between the model and the shaking table, as discussed later. The interaction between the

shaking table and the model is due to the construction process adopted, the transportation

method and the manual bolting of the model to the table. This means that it is impossible to

control possible geometric irregularities of the foundation RC ring beam and the connection

of the building with the shaking table is made with a series of springs (steel bars), with the

adjustment in the bolts allowing for small movements and closing gaps in compression.


14

The fourth and fifth detected modes were, respectively, the second longitudinal at 32.84Hz

and the second transverse at 39.43Hz. These mode shapes are characterized by the movement

of the first floor and roof slabs in the same direction but in opposite ways, with inversion of

curvature. Once again, the influence of the infills can be noted by a small component in the

perpendicular direction of the mode. The change in order of these two second modes, when

compared to the first ones, is possibly due to the large stiffness of infill walls of the South

façade (without openings). The FRF functions also present clear peaks for these last two

modes.

The repetition of the dynamic identification tests after each test stage, DI1 to DI3, allowed for

the detection of the decrement of the frequency of all peaks in the FRF that represent the

above mentioned mode shapes, see Figure 2.10 (a). The increase or decrease of the gain

factor along the dynamic identifications can overlap nearby peaks of other mode shapes,

hence the changes in the FRF need to be tracked in more than one output signal. The damage

in the structure does not only affect the value of the frequency but the shape of the mode as

well, and it is possible, with a considerable amount of damage, for the mode shapes to

disappear, merge or change order. In order to track the evolution of the mode, ensuring a

correct comparison along the loading stages, the Model Assurance Criterion (MAC) [10] was

used:

,

∑ ∅ ∅

∑ ∅ ∑ ∅ (1)

where ∅ and ∅ are the eigenvectors for two different dynamic identification tests and is

the number of degrees of freedom. The MAC was used to compare each mode shape,

identified from DI1 to DI3, with the mode shapes identified in DI0 and it ranges from 0 (no

correlation) to 1 (perfect correlation).


15

1st Transverse Mode

(7.71Hz)

1st Longitudinal Mode

(9.62Hz)

Torsional Mode

(26.95Hz)

2nd Longitudinal Mode

(32.84Hz)

2nd Transverse Mode

(39.43Hz)

Figure 2.9: Mode shapes of the DI 0 of model 1 (initial dynamic identification test)

(a) (b)

Figure 2.10: Frequency change along loading stages: (a) variation of the FRF’s along the test of model 1 at the accelerometer BNE – 1T; (b) evolution of the frequencies along the test of model 1 and their final

variation in respect to DI 0

Figure 2.10 (b) presents the frequency variation of the identified mode shapes along the

dynamic identifications. All five mode shapes were identified from DI0 to DI2, while on DI3

the first two modes, 1st transverse and 1st longitudinal, merged into a single mode shape due to

2 4 6 8 101

2

3

4

5

6

6.41 Hz

7.28 Hz

Gai

n F

acto

r

Frequency (Hz)

DI 0 DI 1 DI 2 DI 3

1st Transversal

7.71 Hz

DI 0 DI 1 DI 2 DI 35

10

15

20

25

30

35

40

45

27.82 Hz(29.5%)

20.88 Hz(36.4%)

39.43 Hz

32.84 Hz

6.43 Hz(16.9% - 33.4%)

9.62 Hz

7.71 Hz

17.42 Hz(35.4%)

Freq

uenc

y (H

z)

Dynamic identification

1st Transversal

1st Longitudinal Torsion

2nd Longitudinal

2nd Transversal

26.95 Hz


16

damage in the RC structure. After the first two stages of the shaking table test, all the

identified mode shapes had and average frequency decrease, in regard to DI0, of 3% and the

first three modes had an average MAC of 0.934. This means that the RC structure was barely

damaged after the 275YRP and 475YRP seismic actions, stages 1 and 2 respectively, and that

the first three mode shapes remained unaltered. This is in agreement with the observed results

since the structure did not present any visible damage after these first two stages. The fourth

and fifth mode presented a lower average MAC of 0.559. Given the small frequency variation

and subsequent lack of damage, this low value can be associated to the difficulties of

capturing more complex mode shapes.

After the third stage, the average frequency decrease of all modes, in respect to DI0, was

30.3% and the average MAC of the last three modes was 0.390. The first two modes merged

into a single mode with a diagonal translation following the Southeast – Northwest direction.

These results seems not in full agreement with the recorded damage after stage 3 (2475 YRP),

see Figure 2.5 and Figure 2.6, which is not enough to assume a loss of almost one third of the

total stiffness of the structure. On the other hand, the collapse of the structure during the last

stage is in agreement with the dynamic data since the structure was already considerably

damaged.

Table 2.1 presents the experimental estimation of the damping ratios along the several

dynamic identifications. None of the identified mode shapes had the expected damping

increment along the tests, confirming the difficulties in the experimental estimation of this

parameter.

Table 2.1 - Experimental damping ratio of model 1

1st

Transverse 1st

Longitudinal Torsion

2nd Longitudinal

2nd Transverse

DI 0 (%) 14.56 3.46 2.15 2.25 0.54

DI 1 (%) 15.44 3.26 2.76 2.05 -

DI 2 (%) 9.98 2.99 2.20 1.81 0.89

DI 3 (%) 4.00 5.33 3.60 4.03 4.00

The seismic vulnerability curves presented in Figure 2.11 relate the damage indicator , see

(2, with the PGA recorded at the base of the model and the computed Input Energy, see (8, for

each mode shape. The damage indicator is computed as:


17

1 (2)

where is the damage indicator of a given mode at stage , is the frequency of the given

mode at stage and is the undamaged, or initial frequency of the given mode. This linearly

proportional ratio between any frequency and the first frequency (DI0), varies from 0,

representing an undamaged state, to 1, representing the collapse of the structure. The damage

indicator assumes isotropic damage [20] between DI0 and stage . The damage indicator of

the torsional mode was associated to the direction with the highest recorded PGA and Input

Energy, hence the longitudinal direction in case of model 1.

The damage indicator is in agreement with the observed damaged, with a very low value after

the first two stages (225 and 475 YRP) and a considerable leap after the third stage (2475

YRP). With the exception of the 1st transverse mode, all other modes have a damage indicator

between 0.30 and 0.36 after the third stage, confirming a generalized loss of stiffness of the

structure and the evenly distributed damage along the four façades of the structure that was

observed. With the collapse of the structure during stage 4 (4574 YRP) along the transverse

direction, the damage indicator of the transverse modes reached the unitary value for the

maximum recorded PGA at that stage.

Figure 2.11: Seismic vulnerability curves of model 1 in the transverse and longitudinal directions, using

the PGA and the Input Energy as input. Here, the damage indicator is a measure of the frequency change

0 1 2 3 4 5 6 7 8 9 100.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Dam

age

indi

cato

r d

PGA (m/s2)

1st Transversal

2nd Transversal

0 1 2 3 4 50.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Dam

age

indi

cato

r d

Input Energy (J)

1st Transversal

2nd Transversal

0 1 2 3 4 5 6 7 8 9 100.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Dam

age

indi

cato

r d

PGA (m/s2)

1st Longitudinal Torsional

2nd Longitudinal

0 1 2 3 4 50.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

Dam

age

indi

cato

r d

Input Energy (J)


2nd Longitudinal


18

Modal frequencies of the infill walls

Following the same procedure used for the global mode shape identification, the peak

identification in the FRF’s, the frequencies of the first mode shape of the infill walls, in the

North and South façades, were identified in the first dynamic identification, DI0, and their

evolution was registered in the subsequent dynamic identifications, see Figure 2.12. Only the

infills of the North and South were identified due to the fact that the used accelerometers

could not read any information above 80Hz, and the preliminary estimations showed that the

infills of the East and West façades, which are considerably smaller in length when compared

to North and South ones, had the first mode shape at higher frequencies.

The infills of model 1 had two leaves, an exterior one with a 9cm thickness and an interior

one with a 7cm thickness. The results showed that the exterior leaves have a slightly higher

frequency when compared to interior ones, which is expected because the stiffness increases

to the third power of the thickness while the mass only increases linearly. The reason for the

small increase is likely to be the boundary conditions, as the exterior leaves are partly

overhanging the slab, thus with lower restriction to rotation. The infills of the South façade

have a higher frequency than the infills at the North façade, in the same position, due to the

lack of openings. The infills at the second floor have a higher frequency than the ones at the

ground floor.

After the first test stage, DI1, the infill walls did not present any considerable frequency

decrease, in accordance with the observed damage and the dynamic information of the global

structure. On the other hand, after the second stage, when no damaged was observed and no

considerable frequency decrease was registered in the global structure, the infill walls of the

ground floor of the south façade and the exterior leaf of the ground floor of the North façade

presented a frequency decrease of 16.4%, 7.7% and 4.2%, respectively. This frequency loss,

since the walls did not present any visible damage, is likely to be due to the loss of connection

between the infill and RC frame, which makes the wall more flexible. The in-plane damage of

the infills is associated to the interstorey drifts, and in stage 2 a 5.9 mm displacement,

corresponding to 0.30% drift, was recorded at the ground RC frames in the transverse

direction, see Figure 2.13, hence the loss of connection between the infill wall and the RC

frame.

After stage 3, the infill walls of the South façade had an average frequency loss of 16.4%

while the walls on the North façade had an average frequency loss of 15.0%. In the South

façade, the exterior and interior leaves, both in the ground and first floors, converged to the

same frequency after in DI3, which indicates larger damage in the exterior walls. The same


19

situation was not registered in the North façade, where stiffness reduction was proportional,

with the exception of the P1 external leaf.

The infill walls of the ground floor presented a higher frequency loss when compared to the

ones on the first floor, which is in agreement with the observed damage but not with the

interstorey displacements and drifts, since the first storey registered similar or higher values.

Also, none of the infill walls on the first storey collapsed during stage 4, apart from an

exterior jamb on the East façade, while all the infill walls on the ground floor collapsed. The

exterior leaf of the infill wall of the ground floor at the North façade presented a frequency

loss of 43.1%, which is in agreement with the observed damaged since this infill was more

damaged than any other in the transverse direction, and it was one of the first walls to

completely collapse out-of-plane.

North façade infill walls South façade infill walls

Figure 2.12: Evolution of the frequencies of the infill walls in the North and South façades along the test of model 1 and their final variation in respect to DI 0.

Interstorey displacements and drifts

Figure 2.13 presents the interstorey displacements and drifts in each main direction,

transverse and longitudinal, for the three first test stages. Increasingly higher displacements

were recorded for each test stage, as expected, with the exception of the displacements in the

transverse direction in stage 3 (2475 YRP). In the first stage (225 YRP), both directions

presented a similar behaviour and similar maximum displacement values, while on the second

stage the transverse direction was considerably more flexible, with three times larger

displacements than the longitudinal direction. In the third stage, again, both directions have a

similar shape and the maximum displacements are similar. These results are in agreement

with the dynamic identification, since until the second stage (475 YRP) the first mode shape

DI0 DI1 DI2 DI330

35

40

45

50

55

60

65

7064.8 Hz(3.0%)

56.2 Hz(5.5%)

52.7 Hz(8.4%)

34.4 Hz(43.1%)

66.7 Hz

60.5 Hz59.5 Hz

P1 exterior leaf P1 interior leaf P2 exterior leaf P2 interior leaf

Fre

quen

cy (

Hz)


57.5 Hz

DI0 DI1 DI2 DI350

55

60

65

70

75

62.3 Hz(7.2%-11.8%)

50.9 Hz(20.0%-26.7%)

70.7 Hz69.4 Hz

67.1 Hz

P1 exterior leaf P1 interior leaf P2 exterior leaf P2 interior leaf

F

requ

ency

(H

z)


62.8 Hz


20

is in the transverse direction while the second one is in the longitudinal direction. After the

third stage these two modes merged into a single mode shape which has a diagonal

translation, due to a similar stiffness in both main directions.

In the transverse direction, the first storey recorded increasingly higher drifts when compared

to the ground storey, while on the longitudinal direction the first storey recorded increasingly

lower drifts than the ground floor. In the first stage, the model presented similar drifts in both

transverse and longitudinal directions, while on the second stage, just as in the maximum

displacements, the transverse direction presented significantly higher drifts. The decrement of

the modal frequencies of the infills in the North and South façades only correlates with the

0.30% drift recorded at the ground level but not with the 0.54% drift recorded at the second

floor in the transverse direction during stage 2. After the third stage the drift values of both

main directions are very similar with a 15.6% difference, on average, between them.

Transverse direction Longitudinal direction

Figure 2.13: Interstorey displacements and drifts of model 1

PGA of the infill walls and RC structure

As damage increases along the test stages, the RC structure and the infill walls lose stiffness

but there seems to be no clear trend with respect to amplifications of the base accelerations.

Figure 2.14 presents the maximum recorded acceleration at the infills, in any of the

0 1 2 3 4 5 6 7

stage 1 stage 2 stage 3

storey 1(2 meters)

Displacement (mm)

roof(4 meters)

00 1 2 3 4 5 6 7 8


storey 1(2 meters)

Displacement (mm)

roof(4 meters)

0

0.0 0.1 0.2 0.3 0.4 0.5 0.6


0

storey 1(2 meters)

Drift (%)

0.00 0.05 0.10 0.15 0.20 0.25 0.30


storey 1(2 meters)

Drift (%)

roof(4 meters)

0


21

accelerometers placed at the infill wall, and at the slab levels of the first storey and roof for

each test stage, as well as the maximum amplification, obtained by dividing the maximum

acceleration by the PGA of that same direction. Analysing the RC structure, as expected, the

measured accelerations increased along the test stages in all directions and floors. As for the

amplifications in the longitudinal direction, apart from the roof level from stage 1 (225 YRP)

to stage 2 (475 YRP), the recorded values increased slightly along the three test stages, while

on the transverse direction the amplifications decreased from stage 2 to stage 3 (2475 YRP) at

both levels. The lower stiffness of the transverse direction is in agreement with the observed

collapse mode during stage 4 (4574 YRP), see Figure 2.7. On average, it can be said that the

amplification in the RC structure is not significant.

The infill walls, on the first and second stage, presented similar maximum values for the same

façade, with the exception of the outer leaf at the ground level in the North façade, which

exhibited higher values than the other walls. During the third stage, the maximum recorded

accelerations were no longer similar between the walls of the same façade, but no particular

pattern regarding the position or leaf was found. Very similar values were also recorded in the

infill walls of the same direction, North-South and East-West, during stages 1 and 2. In stage

3, higher maximum accelerations were recorded on the North and East infill walls when

compared to the South and West ones, respectively.

The infill walls of the North and South façades presented a small amplification decrease from

the stage 1 to stage 2, except for the interior leaf at the ground floor at the North façade which

presented a 24.0% decrease and another 15.6% decrease from stage 2 to stage 3. All other

infill walls in the North façade had an amplification increment from stage 2 to stage 3, while

on the South façade the exterior leaf at the ground level and the interior leaf in the first storey

presented a small decrease in the amplification while the outer leaves presented an increment.

In the East and West façades all the infill walls presented a small amplification increment

from stage 1 to stage 2, while on stage 3 half presented a small increment and the other half a

small decrement in the recorded amplification, without any particular pattern as far as the

level or leaf are concerned. Therefore, no clear conclusion can be made regarding the

decrease or increase of amplification with damage, with different trends found.


22

North façade South façade

East façade West façade

RC structure

Figure 2.14: Recorded PGA and amplifications at the infill walls and at the RC structure for each test stage of model 1

Out-of-plane PGD and deformation of the infill walls

Figure 2.15 presents the maximum out-of-plane displacement (PGD), based on double

integration of the measured maximum acceleration, of each infill wall at the North and South

façades and the recorded displacement on the other two accelerometers at the same instant.

Therefore, the line in the graphs of Figure 2.15 represents the out-of-plane deformation of the

infill wall along its length seen from above, and the opposite curvatures of the lines are

associated to interior or exterior bending. It is stressed that these are the displacements of the

5 4 3 2 1 00 5 10 15 20 25 30

1

2

3

NE P2 NE P1 NI P2 NI P1

Amplification Acceleration (m/s2)

Stage

2.0 1.5 1.0 0.5 0.00.0 2.5 5.0 7.5 10.0 12.5 15.0

1

2

3

SE P2 SE P1 SI P2 SI P1


Stage

2.0 1.5 1.0 0.5 0.00.0 2.5 5.0 7.5 10.0 12.5 15.0

1

2

3

EE 2.1 EE 2.2 EE 1.1 EE 1.2 EI 2.1 EI 2.2 EI 1.1 EI 1.2

Amplification

Stage

Acceleration (m/s2)

1.5 1.0 0.5 0.00.0 2.5 5.0 7.5 10.0

1

2

3

WE 1.1 WE 1.2 WI 1.1 WI 1.2

Amplification

Stage

Acceleration (m/s2)

2.0 1.5 1.0 0.5 0.00.0 2.5 5.0 7.5 10.0 12.5 15.0

1

2

3

Storey 1 Trans Dir Storey 1 Long Dir Roof Trans Dir Roof Long Dir


Stage


23

infill with respect to the frame, being therefore a relative displacement and not a full

displacement.

As expected, the deformation increases along the seismic tests and the infill walls of the first

storey have a higher deformation than the ones at the ground floor. With the exception of the

infill walls on the first storey at the North façade, the exterior leaf presents a higher

deformation when compared to the interior one. The exterior leaf has a higher thickness than

the interior one, 9cm and 7cm respectively, but the exterior leaf is partially out of the plane of

the RC frame. As noted before, this confirms the fact that the less restraint to rotation is found

on the on the exterior walls, when compared to the interior walls. Even though the infill walls

in the South façade do not have openings and the ones on the North façade have openings, the

range of displacements along the wall is the same for both façades, indicating a small

influence of the openings.

The deformation shape of the infill wall at the ground floor of the North façade was not

altered with the damage along the seismic tests, and both leaves presented the same shape

with PGD recorded at the middle of the wall. Only on stage 3 (2475 YRP) did the

accelerometer next to the door of the exterior leaf, one of the most damaged infills of

model 1, presented a displacement closer to the middle one, confirming the important damage

found. The interior and exterior leaves of the infill wall at the first storey presented the same

deformation shape, but with maximum values at opposite sides under the windows, along the

three recorded test stages.

The interior leaf of the infill wall at the ground floor of the South façade did not present any

visual damage, even though its modal frequency decreased 26.7% from DI0 to DI3, and its

deformation shape did not change along the tests with all three accelerometers recording

similar displacements, hence confirming that the loss of stiffness is related to the loss of

connection between the infill wall and the RC frame. A similar situation can be found in the

exterior leaf until the second stage (475 YRP), since in the third stage (2475 YRP) some

damage was recorded near the RC columns and the displacements at the side and centre

increased. At the first storey, both leaves maintained their deformed shape along the test, and

once again no visual damage was recorded along the test but a decrement in the modal

frequencies was, and nearly the same PGD at the centre and East side of the wall, while on the

West side the exterior leaf had increasingly higher values.


24

North walls South walls S

tage

1

(225

YR

P)

Sta

ge 2

(475

YR

P)

Sta

ge 3

(247

5 Y

RP

)

Figure 2.15: Out-of-plane deformation of the North and South infill walls along the tests of model 1

Figure 2.16 presents the PGD recorded in the infill walls of the East and West façade for the

first three test stages of model 1. As expected, the PGD increases along the test stages and

higher values were recorded in the infill walls of the first storey. The interior leaves of the

infills at the West façade presented the lowest PGD, while all the other infill walls presented

similar values, which is in agreement with the observed damage since both façades presented

a similar crack pattern.

Acc 1 Acc 2 Acc 32.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

Out

-of-

plan

e di

spla

cem

ent (

mm

) N-E-P2 N-E-P1 N-I-P2 N-I-P1

Acc 1 Acc 2 Acc 32.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

Out

-of-

plan

e di

spla

cem

ent (

mm

)

S-E-P2 S-E-P1 S-I-P2 S-I-P1

Acc 1 Acc 2 Acc 34.04.55.05.56.06.57.07.58.08.59.09.5

10.0

Out

-of-

plan

e di

spla

cem

ent (

mm

)

N-E-P2 N-E-P1 N-I-P2 N-I-P1

Acc 1 Acc 2 Acc 34.04.55.05.56.06.57.07.58.08.59.09.5

10.0

Out

-of-

plan

e di

spla

cem

ent (

mm

)


Acc 1 Acc 2 Acc 312

14

16

18

20

22

24

26

28

30

Out

-of-

plan

e di

spla

cem

ents

(m

m)

N-E-P2 N-E-P1 N-I-P2 N-I-P1

Acc 1 Acc 2 Acc 312

14

16

18

20

22

24

26

28

30

Out

-of-

plan

e di

spla

cem

ents

(m

m)



25

Figure 2.16: Out-of-plane PGD of the East and West infill walls of model 1

2.2.2 Results of model 2

Model 2 was designed following the Eurocode normative, EC2 [29] and EC8 [30], using

concrete and steel rebars of higher classes (C30/37 and S500, respectively), together with

single leaf clay brick infill walls with bed joint reinforcement, connected to the RC frame

with steel dowels (or connectors), every second bed joints. Therefore, model 2 represents

likely future solutions for RC frames and masonry infills, where the reason to use a single leaf

is to place an external thermal insulation system. Figure 2.17 presents the position, and label,

of the accelerometers in model 2, noting that the accelerometers were placed on the exterior

only, since walls are single leaf.




Model 2 was tested following the same test procedure as model 1 in four stages with

increasing seismic amplitude and, as before, no relevant damage due to transportation

occurred. After the first two stages (225 and 475 YRP), and again as the previous model,

0 5 10 15 20 25 30 35 40

1

2

3

EE 2.1 EE 2.2 EE 1.1 EE 1.2 EI 2.1 EI 2.2 EI 1.1 EI 1.2

Stag

e

Out-of-plane PGD (mm)

0 5 10 15 20 25

1

2

3

WE 1.1 WE 1.2 WI 1.1 WI 1.2

Stag

e

Out-of-plane PGd (mm)

BNE 2L

BNE 1L

N1 4

N1 1N1 3

N1 2

INP L

N1 5

BSW 2L

S2 6

S2 9

S2 3

S2 7

S2 1S2 4

BSW 1L

S1 4

S1 7

S1 1S1 6S1 3

S1 9

S2 2

S2 8

S2 5

S1 8

S1 2

S1 5E1.1 2

E1.1 3

E1.2 1

E2.1 3

E1.1 1

E2.1 1

E2.1 2

BNE 2T

BNE 1T

E2.2 2E2.2 1

INP T

BSW 2T

BSW 1T

W1.2 1W1.1 1


26

model 2 did not present any visible damage but, on the contrary to model 1, model 2

presented negligible frequency decrease, hence negligible stiffness loss. After stage 3 (2475

YRP), the model presented the crack pattern shown in Figure 2.18, with all damage

concentrated at the ground floor. The concentration of lines around the RC columns

represents mortar rendering expulsion, leaving nearly half of the RC column visible, although

no cracks were visible in the RC. Cracks starting from the corners of the openings and

progressing towards the RC frame were also visible after stage 3 in most of the openings. In

the East and West façade, the crack pattern around several jambs is clear, separating them

from the RC frame and the section of the infill wall below the opening, just as in model 1 but

not as clear. The inside face of the model also presented expulsion of the rendering at the

intermediate columns of the East and West façades, leaving the RC columns visible, and a

crack pattern similar to the outside one. After stage 3, the RC structure and the infill walls of

the North and South façades presented an average frequency loss in the identified modes of

28.1% and 17.8%, respectively.

Model 2, contrary to model 1, did not collapse during the fourth and last stage of the test

(4574 YRP), but it was heavily damaged as shown in Figure 2.19 and Figure 2.20. The South

façade, see Figure 2.20 (b), presented the lowest amount of damage, with the first level infill

presenting no cracks within the wall and the ground level infill presenting cracks mainly at

the connection between the infill wall and the RC frame. All the mortar applied to the first

floor columns and part of the mortar of the second floor columns was expelled. On the North

façade, see Figure 2.20 (a), the first storey infill presented cracks at the lateral and upper

connections of the infill to the RC frame, at the intermediate jamb and cracks starting at the

corners of the openings moving towards the RC frame. The ground infill was the most

damaged in the model, as it became completely detached from the surrounding RC frame and

was prevented from falling out-of-plane only by the bed joint reinforcement and the

connectors. The intermediate jamb was completely loose and could be hand pushed out-of-

plane. The East and West façades, see Figure 2.20 (c), presented similar damage, with all the

jambs completely detached from the RC frame and the lower part of the infill wall sustained

only by the bed joint reinforcement and connectors to the RC frame.

All RC columns at the ground level, and part of the RC beams and columns of the second

floor, were visible since the mortar rendering was expelled, and heavy damage was visible.

Columns had mid-height horizontal cracks, see Figure 2.20 (d) and (e), due to the influence of

the infill openings on the horizontal load transfer, aligned with the lower part of the window

openings. One of the columns, see Figure 2.20 (f), presented severe cracking at the upper


27

connection to the beam with rebar exposure, meaning that model 2 could be developing a very

undesirable soft storey collapse mode [40], just as the previous model.

North South

East West

Figure 2.18: Crack patterns of model 2 after stage 3 (2475 YRP) (Note: the drawn lines on the RC frame represent damage on the rendering applied to the RC frame)

North South

East West

Figure 2.19: Crack patterns of model 2 after stage 4 (4574 YRP) (Notes: the drawn lines on the RC frame represent damage on the rendering applied to the RC frame. The blue lines developed after stage 3)


28

(a) (b)

(c) (d)

(e) (f)

Figure 2.20: Damage in model 2 after the fourth stage (4574 YRP): (a) North façade; (b) South façade; (c) West façade from the inside; (d) detail of the left jamb of the door on the North façade and a

horizontal crack at mid-height of the Northeast corner column; (e) horizontal crack at mid-height of the Southwest corner column; (f) heavily damaged top column-beam connection of the Southwest corner

column with loss of the concrete cover and rebar exposure


During the first dynamic identification, DI0, five mode shapes were found, see Figure 2.21,

namely: first and second order transverse; first and second order longitudinal; (first) torsional.

The modes and the order of the modes were the same as in model 1. The first transverse and

the first longitudinal had close frequencies, 7.32Hz and 8.37Hz respectively, being the

longitudinal stiffness slightly higher than the transverse one. The frequency leap to the

torsional mode, at 26.77Hz, is associated to the contribution of the infill walls to the global

stiffness of the RC structure. The last two modes, second order longitudinal and second order

transverse, were identified at 30.33Hz and 36.40Hz, respectively.


29

As in mode 1, the torsional mode shape was not perfectly identified, as far as the shape is

concerned, since the North façade did not keep a 90º angle with the other two façades, which

is impossible due to the presence of the RC slab. The second order transverse mode also

presented some distortion at the corners, which is again impossible due to the RC slab. These

problems are associated to the limitations of the used equipment and reading errors. The

dynamic identifications had high coherences, close to the unit value, ensuring the good

quality of the results, and the peaks in the FRF’s were also perfectly clear, see Figure 2.22 (a),

along the test. The MAC [10]values were used to better understand the changes in the mode

shapes.

1st Transverse Mode (7.32Hz)

1st Longitudinal Mode (8.37Hz)

Torsional Mode (26.77Hz)

2nd Longitudinal Mode (30.33Hz)

2nd Transverse Mode (36.40Hz)


Figure 2.22 (b) presents the frequency variation of the model along the test. After the first two

stages (225 and 475 YRP) the model did not present any significant frequency decrease in the

mode shapes, when compared to DI0. This is in agreement with the observed results as the

model did not present any visible damage after these two stages, and the reinforcement

connectors prevents the masonry infills to separate from the RC frame. During stage 3 (2475

YRP) the model endured considerable damage in the RC structure, with an average frequency


30

loss of 13.0% in the transverse direction (first and second modes) and 38.2% in the

longitudinal direction (first and second modes) when compared to DI0. The first transverse

and longitudinal modes changed positions and the torsion mode also had a frequency decrease

of 36.7%, with an average MAC of 0.81 which indicates that the mode shapes are only being

slightly altered by the damage. These results are due to the damage observed, possibly due to

damage concealed in the RC structure by the mortar rendering and separation between the

infill walls and the RC frame.

The last stage (4574 YRP) left the model at a near collapse state and the first transverse and

the first longitudinal modes merged, with an average loss of frequency of 78%, with the new

mode being a torsion with the centre of rotation very close to the South face, therefore with a

high movement amplitude of the North façade. These results are in clear agreement with the

observed damage, as the RC structure and the infill walls presented heavy damage, and the

North façade presented the most damaged infill wall at the ground floor and a RC column

with heavy damage and exposed rebar at the connection with the beam. Only the second

longitudinal mode was also identified, and it presented a 56.6% frequency decrease.

The seismic vulnerability curves presented in Figure 2.23 confirm the observed damage and

dynamic data, as until stage 2 (475 YRP) none of the mode shapes present significant damage

and after stage 3 (2475 YRP) the longitudinal modes presented an average damage of 0.38

while the transverse modes presents an average damage of 0.14. The crack pattern observed is

more associated to the transverse damage value than the longitudinal one. After stage four

(4574 YRP), the first and second mode presented a damage around 0.8, indicating the already

mentioned near collapse state of the RC structure and of some of the infill walls.

(a) (b)

Figure 2.22: Frequency change along loading stages: (a) variation of the FRF’s along the test of model 2 at the accelerometer BNE – 2T; (b) evolution of the frequencies along the test of model 2 and their final


0 2 4 6 8 101

2

3

4

5

6

7

8

9

10

11

7.03 Hz 7.32 Hz

DI 0 DI 1 DI 2 DI 3 DI 4

1.73 Hz

6.35 Hz

Gai

n F

acto

r

Frequency (Hz)

1st Transversal

7.32 Hz

DI 0 DI 1 DI 2 DI 3 DI 40

5

10

15

20

25

30

35

40

20.41 Hz(32.7%)

16.95 Hz(36.7%)

1st Transversal

1st Longitudinal Torsion

2nd Longitudinal

2nd Transversal

36.40 Hz

30.33 Hz

6.35 Hz(13.2%)

4.72 Hz(43.7%)

31.29 Hz(14.0%)

1.73 Hz(76.3%)

1.64 Hz(80.5%)

8.38 Hz

7.32 Hz

15.79 Hz(56.6%)

Fre

quen

cy (

Hz)


26.77 Hz


31

Figure 2.23: Seismic vulnerability curves of model 2 in the transverse and longitudinal directions, using the PGA and Input Energy as input


Figure 2.24 presents the frequency decrease of the infill walls on South façade and at the

ground floor of the North façade. As expected, the infills of the South façade present a higher

frequency as they have no openings. The infill walls present an initial small frequency

decrease, after stage 2 (475 YRP), even without any visible damage, possibly associated to

some loss of connection between the infill wall and the RC frame. After stage 3 (2475 YRP),

the infill walls at the ground floor presented a frequency decrease of around 20%, while the

infill at the upper floor presented a decrease of 12%. These results are in agreement with the

crack pattern, since the upper floor presented no visual damage but the ground floor did.

After the last stage (4574 YRP) the infill wall at the ground floor of the North façade was so

damaged and detached from the RC frame that it was not possible to identify its first modal

frequency. The infill wall at ground floor of the South façade presented a higher frequency

decrease when compared to upper floor infill wall, which is in agreement with the observed

crack patterns, although its damaged was mainly at the connection with the RC frame.

0 1 2 3 4 5 6 7 8 9 10 11 120.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Dam

age

indi

cato

r d

PGA (m/s2)

1st Transversal

2nd Transversal

0 1 2 3 4 5 6 7 80.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Dam

age

indi

cato

r d

Input Energy (J)

1st Transversal

2nd Transversal

0 1 2 3 4 5 6 7 8 9 10 11 120.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Dam

age

indi

cato

r d

PGA (m/s2)


2nd Longitudinal

0 1 2 3 4 5 6 7 80.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Dam

age

indi

cato

r d

Input Energy (J)


2nd Longitudinal


32

Figure 2.24: Evolution of the frequencies of the infill walls in the North and South façades along the test of

model 2 and their final variation in respect to DI 0


The interstorey displacements increased with the seismic amplitude, see Figure 2.25, and low

and similar values were recorded for stage 1 and 2 (225 and 475 YRP) in both transverse and

longitudinal directions. During stage 3 (2475 YRP), the longitudinal direction recorded on

average 40.9% higher displacements, confirming the higher loss of stiffness expressed in the

dynamic identification when compared to the transverse direction. On the last stage of the test

(4574 YRP), the highest displacement was recorded in the transverse direction at the first

floor level, while the roof level recorded higher displacements in the transverse direction. The

displacements recorded during the last stage were, on average for both directions, 85.4%

higher when compared to stage 3.

Model 2 recorded maximum interstorey drifts, see Figure 2.25, below 0.05% and 0.08% in the

first two stages (225 and 475 YRP), respectively, and the longitudinal direction recorded

higher values when compared to the transverse direction. Given the higher stiffness of the

longitudinal direction until the third stage, it would be expectable for the transverse direction

to have higher drifts. On stage 3 (2475 YRP), the recorded values on the longitudinal

direction were 38.7% higher, which is in agreement with the dynamic data, as the first

transverse and first longitudinal modes changed order. During the last stage (4574 YRP), the

amplitude of the maximum recorded drifts was considerably higher than the stage 3 values, on

average 88.6%. Except for the transverse direction during stage 1, the interstorey drift values

were always higher at the ground level in comparison to the upper level, which is in perfect

agreement with the crack patterns, as the damage in the RC structure and the infill walls was

DI0 DI1 DI2 DI3 DI420

25

30

35

40

45

50

55

60

65

70

52.1 Hz(22.3%)

57.4 Hz(11.7%)

48.7 Hz(25.1%)

45.9 Hz(19.3%)

23.6 Hz(64.7%)

67.0 Hz65.0 Hz

56.9 Hz

P1 North facade P1 South facade P2 South facade

Fre

quen

cy (

Hz)



33

concentrated at the ground floor. It is noted that values up to 4% were found, which are much

larger than values normally accepted in static tests.




As expected, the RC structure exhibited increasingly higher accelerations along the test, with

the longitudinal direction presenting larger values than the transverse direction, and the roof

recorded higher values in comparison to the first floor RC slab, see Figure 2.26. The

amplification increased from stage 1 to stage 2 and decreased from stage 2 to stage 3 in the

transverse direction, while the opposite was registered in the longitudinal direction, being kept

constant in the last stage. This confirms the rather complex behaviour of the RC frame /

masonry infill system.

The infill walls on the transverse direction recorded lower accelerations along the test when

compared to the infill walls on the longitudinal direction. On the longitudinal direction, the

infill walls at the first floor always had a PGA higher than the ones at the ground floor, but on

the transverse direction, in stage 3 and 4, the three highest PGA’s were recorded at the ground

infill walls. As for the amplifications, on the transverse direction, the changes were in general

small, with the exception of two walls for the last stage (W 1.1 and E 1.2). For the

0 10 20 30 40 50 60 70 80

stage 1 stage 2 stage 3 stage 4

storey 1(2 meters)

Displacement (mm)

roof(4 meters)

0

0 10 20 30 40 50 60

storey 1(2 meters)


Displacement (mm)

roof(4 meters)

0

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0


0

storey 1(2 meters)

roof(4 meters)

Drift (%)

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0


storey 1(2 meters)

Drift (%)

roof(4 meters)

0


34

longitudinal direction, the amplifications were larger, with a clear increase from stage 2, and

an enormous amplification for S2 at stage 4 (about 3.5).


RC structure


Out-of plane PGD and deformation of the infill walls

The out-of-plane deformation of the infill walls was computed as the value of the

displacement of all accelerometers or LED’s, depending on the wall, at the instant when the

maximum PGD was measured in each of the test stages. The results of the last test stage

(4574 YRP) are not presented in the South façade because the violence of the seismic action

led to incoherent and unreliable results in a great number of accelerometers. As expected, all

infill walls presented increasing displacements along the seismic test due to the higher seismic

amplitude and higher flexibility, as a result of the accumulated damage, and in the South

façade the infill wall of the first storey presented higher displacements than the ground level

one, see Figure 2.27 and Figure 2.28. Until the third stage (2475 YRP), the infill wall on the

first storey of the South façade did not present any change in the deformed shape, with the

highest values recorded at upper part of the infill, which is in agreement with the observed

crack pattern since the infill did not present any visible damage and the frequency loss is

associated to damage in the connection to the RC frame. The infill wall at the ground floor

2 1 00 5 10 15 20

1

2

3

4

E 1.1 E 1.2 E 2.1 E 2.2 W 1.1 W 1.2

Amplification

Stage

Acceleration (m/s2)

5 4 3 2 1 00 5 10 15 20 25 30 35 40

1

2

3

4

N1 S1 S2


Stage

2 1 00 5 10 15 20

1

2

3

4



Stage


35

presented the highest displacements, until the second stage (475 YRP), at the right side of the

wall where more extensive cracks were observed and in the third stage (2475 YRP) also at the

top. This change is associated to the observed damage.

The infill wall at the first floor of the North façade, see Figure 2.29, presented its highest

displacements below the left window until the third stage (2475 YRP) of the test, and the lack

of change in the deformed shape is associated to the absence of localized damage, as the

increment in the flexibility is associated to the loss of connection to the RC frame. On the last

stage (4574 YRP), the maximum values were recorded at both sides of the wall. On the South

façade the results were obtained from the double integration of the data recorded by the

accelerometers, while on the North façade the results were obtained using the KRYPTON

camera, and given the difference characteristics, precision and acquisition frequency of the

equipment, the obtained displacements are different and should not be directly compared.

The PGD of all other infill walls on model 2 are presented in Figure 2.30, but only for the

three first stages of the test because some of the results of the last stage were unreliable. All

the infills recorded higher PGD’s along the test and the infills at the first floor recorded higher

PDG in all three stages, when compared to ground floor ones. At the ground floor, the infill in

the North façade recorded the highest PGD in stages 2 and 3, as expected being the lengthiest

in this comparison, and the most damaged in the model.

Stage 1 (225 YRP) Stage 2 (475 YRP)

Stage 3 (2425 YRP)

Figure 2.27: Out-of-plane deformation of the infill wall at the ground level of the South façade (mm)

0.7665 0.7665

0.7010

0.8320

0.6355

0.89750.9630

0.8320

1.029

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

1.3991.399 1.274

1.5231.647

1.7721.896 2.021

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

3.295

3.8134.330

4.848

3.295

5.365

2.777

5.883

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)


36


Stage 3 (2475 YRP)

Figure 2.28: Out-of-plane deformation of the infill wall at the first storey of the South façade (mm)



Figure 2.29: Out-of-plane deformation of the infill wall at the first storey of the North façade (mm)

4.919

4.991 5.064

5.136 5.209

4.846

5.281

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

8.178

8.289

8.4008.511

8.622

8.067

8.733

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)25.32

25.71 26.1026.49

24.9324.54

26.8826.88

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

-0.018 0.120.26

0.12

0.39 0.12

-0.018

0.530.53

-0.018

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

openingopeningopening

Infill length (m)

Infi

ll h

eigh

t (m

) window window

0.480.39

0.29

0.58

0.19 0.098

0.580.67

0.29

0.67

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window

0.97

0.44

1.52.0

-0.090

2.63.1

0.97

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window

-13-13

-6.30.40 7.1 14 20

-20

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window


37

Figure 2.30: Out-of-plane PGD of the North, East and West infill walls (mm)

2.3 DESIGN AND CONSTRUCTION OF THE MODELS

2.3.1 Building model

The first step in the experimental program was the definition of the geometry and the building

solutions of the prototypes. The geometry survey was done elsewhere to define the average

height and length of the RC frames [34], and the resulting geometry, a building with a two

storey single bay frame in one direction and a two storey double bay frame in the other

direction, can be seen in Figure 2.31. Given that one of the objectives of the present work is to

assess the performance of modern RC frame structures, the RC frame was designed according

to the most recent standards, EC2 [29] and EC8 [29], including reinforced solutions for the

infill walls.

Figure 2.31: Prototype geometry (m)

0 5 10 15 20 25 30 35

1

2

3

E 1.1 E 1.2 E 2.1 E 2.2 W 1.1 W 1.2 N 1

Stag

e


Facade A Facade B

B

A5,7

33

3,23 3,23

Plan view

6,45

5,7


38

Testing the complete structure and not only the RC frames with infills seems a more

reasonable option as the interplay of the response of all components can be captured. Ideally,

the test should be performed at full scale but the physical limitations (maximum dimensions

and payload capacity) of the testing equipment and laboratories impose, in most cases, the use

of scaled models in shaking table tests. This is an important drawback of this type of facility.

Scaled models are obtained using similitude laws, which add complexity to the construction

and test setup in order to fulfil their requirements. In order to test a RC model to its ultimate

capacity the following need to be correctly simulated: i) the geometry; ii) the stress-strain

relationship of the materials; iii) the mass and gravity forces; iv) the initial conditions and the

boundary conditions [5].

The first condition is easily fulfilled by direct application of a geometric scale, although some

pre-fabricated construction elements may have a limited range of dimensions, as infill clay

units. Very small scales may also represent higher construction challenges. Obtaining

adequate stress-strain relationships of the materials can be a much more complex task since

such a relationship has to be fulfilled throughout different stress or strain levels, rates,

gradients, etc. [2]. For very small scales it is not uncommon to use different materials in the

models. The mass and gravity forces are addressed, respectively, by the Cauchy and Froude

similitude laws [5]. The first one is adequate for phenomena in which the restoring forces are

derived from the stress-strain constitutive relationships and the elastic restoring forces, see

Eq. (3). Froude similitude is adequate for phenomena in which the gravity forces are

important, being the Froude value the ratio between inertia and gravity forces, see Eq. (4).

The use of both laws simultaneously is the obvious choice in order to more accurately

replicate the dynamic behaviour of structures, particularly when strongly non-linear behaviour

is expected. In the present work both laws were taken into consideration in the model

definition, following the relations described in Table 2.2. As for the boundary conditions, the

soil-structure interaction is not considered as the model is fixed to the shaking table using

bolts, meaning that the input signal is directly transferred to the structure.

(3)

(4)


39

Table 2.2 - Cauchy-Froude similitude law

Parameter Scale Factor Parameter Scale Factor

Length (L)

Mass (m)

Modulus of elasticity ( E ) 1

Weight (w)

Specific mass (ρ) Force (F)

Area (A) Moment (M)

Volume (V) Stress (τ) 1

Displacement (d) Strain (ε) 1

Velocity (v) Time (t)

Acceleration (a) 1 Frequency (f)

High resistance class materials were chosen for the concrete and steel rebars (C30/37 and

S500, respectively), while the scale factor was chosen taking into account the physical

limitations of the shaking table of LNEC, see Chapter 3. The models were designed at a

reduced scale of 1:1.5, meaning that reduced loads were applied to a model with the geometry

also reduced, see Figure 2.32, using the similitude law, see Table 2.2 ( 1.5 . The design

loads used in all three models, reduced using the similitude law relations described in

Table 2.2, can be seen in Table 2.3.

Figure 2.32: Geometry of the tested model reduced to a scale of 1:1.5


40

Table 2.3 - Design loads of the models already reduced at scale of 1:1.5

Load Description Load value

Self-weight

Partition walls mortar (1.5 cm) + clay masonry units (11cm) + gypsum (1.5cm)

1.11 KN/m2

Floor slab storey 1 reinforced concrete slab (thickness = 0.12 m) 1.07 KN/m2

Roof slab reinforced concrete slab (thickness = 0.12 m)- 1.07 KN/m2

Infill walls mortar (1.5cm) + clay masonry unit (9cm) + clay masonry unit (7cm) + gypsum (1.0cm)

3.74 KN/m

Infill wall mortar (1.5cm) + clay masonry unit (15 cm) + mortar (1.5cm)

3.00 KN/m

Imposed Load

Storey 1 domestic and residential 1.33 KN/m2

Roof accessible 0.67 KN/m2

The EC8 [29] contemplates seismic design using response spectra for near-field and far-field

earthquakes. In order to define these spectra, one must know the geographical location of the

building and the soil type of the building. The EC8 [29], as an international standard, refers

specific parameters to the National Annex. Therefore, in article NA-3.2.1(2) of the National

Annex of EC8 [29], continental Portugal is divided in a local council-based zoning,

considering six magnitude levels for Type 1 far-field seismic actions (1 to 6) and five

magnitude levels for Type 2 near-field seismic actions (1 to 5).

As for the soil type, EC8 [29] defines in its Table 3.1 seven types of soil, from rock (type A)

to soils with liquefaction characteristics (type S2). The specific parameters associated to each

soil type needed to compute the response spectra can be found in the National Annex of EC8

[29], more specifically in its Tables NA-3.2 and NA-3.3. Here, it is assumed that the models

would be built in Lisbon, zones 1.3 and 2.3, and in a type A rocky soil. The obtained spectra,

reduced following the similitude law, can be seen in Figure 2.33.


41

Figure 2.33: Response spectra after the application of the similitude law, see Table 2.2, according to EC8 [29]

All four façades of the model had infill walls. The South façade was blind while the others

had openings in around 20% of the surface area, see Figure 2.34. The clay brick units used

were only scaled in the thickness, while height and length were the same for all three models

and kept at a 1:1 scale.

0.0 0.5 1.0 1.5 2.0 2.5

0.25

0.50

0.75

1.00

1.25

1.50 Type 1 Type 2

a (m

/s2 )

T (s)


42

(a) (b)

(c) (d)

Figure 2.34: Geometry of the openings in each façade: (a) North façade; (b) West façade; (c) East façade; (d) South façade

The chosen similitude law implies that the specific mass of the prototype and the model are

different, see Table 2.2, as . This problem was solved using two types of

additional steel masses: one applied to the RC structure; and another to the infill walls. The

masses of the RC structure had 82x82x26 cm, weighted around 12 KN each and were bolted

to the slab of the first floor and the slab of the roof. A total number of twelve masses were

used, six in each slab, see Figure 2.35 (a). The masses applied to the infill walls had 15x15x4

cm and weighted around 0.072 KN each. These masses were applied to both sides of the wall,

evenly distributed and bolted in two edges of the plate, see Figure 2.35 (b). Each mass was

bolted to the surface of a single unit, in order not to increase the strength of the masonry joints

or influence the crack pattern. A total number of three hundred and thirty four masses were

used in each model.


43

(a)

(b)

Figure 2.35: Additional steel masses for: (a) RC concrete structure, bolted to the slabs of the 1st floor and roof with 82x82x26 cm and 12KN each; (b) Infill walls, bolted to both sides of the wall with 15x15x4 cm

and 0.072KN each

The model to be tested in the shaking table was built inside the NESDE building and then was

transported to the shaking table using the existing crane. In the present work, the model was

built by a construction company hired specifically for this project using techniques and

workers accustomed to RC and masonry construction. Given these conditions two aspects had

to be considered: the foundation of the model had to be plane, otherwise the model could be

damaged when bolted to the shaking table; lifting eye bolts had to be provided for the

transportation of the model. The first issue was solved by constructing the model on top of a

horizontally aligned platform, see Figure 2.36 (a). The second issue was solved by designing

the models with a RC ring beam with four steel plates with a lifting eye at its corners, see

Figure 2.36 (b). This ring beam was also perforated in order to bolt the model to the shaking

table. Figure 2.36 (c) shows the model already on top of the shaking table, ready for testing.


44

(a)

(b)

(c)

Figure 2.36: Construction of the models: (a) horizontally aligned surface on which the models were constructed; (b) RC ring beam with steel connector with an eye in lift and transport the model to the

shaking table; (c) model 1 on the shaking table before the test

The infills of the model represent another future possibility when the design follows EC2 [29]

and EC8 [29], a single leaf clay brick wall with reinforced rendering nailed to the RC frame

and infill wall on both sides, see Figure 2.37. The leaf was completely within the RC from

plane, mortar rendering was used on both sides of the infill and the units were horizontally

perforated. The mortar used for the bed joints and plaster was pre-batched and with a M5

class.

The reinforcement grid chosen, Bekaert Armanet ϕ1.05mm 12.7x12.7mm, see Figure 2.38

(a), was nailed to the RC frame using a Hilti X-M8H10-37-P8, see Figure 2.38 (b) and (c),

using a gun and Hilti shot powder actuated tools. Similar nails should have been used to nail

the grid to the infill wall but were substituted by the additional masses above mentioned, see

Figure 2.35 (b), as these had to be used already due to the similitude law chosen. Figure 2.38

(d) shows the application of the grid in the inner surface of an infill. In order to simulate the

attachment of the grid to the infill wall with nails, a ring was installed between the mass and


45

the wall, in each bolt, with a contact area similar to the Hilti X-M8H10-37-P8 nail applied in

the RC frame, see Figure 2.38 (e).

(b)

(a) (c)

Figure 2.37: Single leaf clay brick infill walls with reinforced plaster from Model 3 already scaled: (a) spacing of the Hilti X-M8H10-37-P8 connectors along the height of the RC column; (b) detail at the RC

column; (c) detail of the Hilti X-M8H10-37-P8 connectors

20.0

40.0

40.0

40.0

40.0

RC column

Bekaert ArmanetØ1.05mm 12.7x12.7mm

1.5cm M5 mortar

Hilti X-M8H10-37-P8

1.5cm M5 mortar

RC column

Bekaert ArmanetØ1.05mm 12.7x12.7mm 1.5cm M5 mortar

1.5cm M5 mortar Hilti X-M8H10-37-P8

30x20x15cm

Hilti X-M8H10-37-P8

Bekaert ArmanetØ1.05mm 12.7x12.7mm

1.5cm M5 mortar


46

(a)

(b)

(c)

(d)

(e)

Figure 2.38: Construction of the infills of model 3: (a) Bekaert Armanet ϕ1.05mm 12.7x12.7mm; (b) Hilti X-M8H10-37-P8; (c) application of the grid in the outer surface at a corner column; (d) application of the

grid in the inner surface; (e) additional masses with steel rings attached to the infill walls


47

2.3.2 Wall panels

In order to test the four types of masonry enclosures’ solutions described in Chapter 1, the

second part of this transnational access activity comprised the dynamic testing of a closed RC

plane frame with external dimensions of 6.40m by 3.25m. The structural elements of this

frame have cross-sectional dimensions of 0.50m by 0.40m (beams) and 0.40m by 0.40m

(columns). These plane frame specimens were tested simultaneously for in-plane and out-of-

plane dynamic actions, representing the response of a frame panel in the 4th floor of an eight

storey RC building (see Figure 2.39). The columns have a centred prestress which represents

the vertical load from the floors above.

Figure 2.39: Frame panel of typical RC building

The first two masonry enclosure solutions tested were the unreinforced masonry and the one

with horizontal reinforcement between masonry units (Bekaert Murfor RND/Z-5-200).

Afterwards, both masonry infills were demolished and rebuilt using a reinforced mortar

coating. The following pictures (Figure 2.40 to Figure 2.44) show the construction process of

the models.

Figure 2.40: Reinforcement layout for the RC frames


48

Figure 2.41: Prestressing details

Figure 2.42: RC frame ready for infill construction

Figure 2.43: Masonry infill construction with bed joint reinforcement


49

Figure 2.44: Specimens ready for testing


50

3 The LNEC Earthquake Engineering testing

facility

LNEC owns a large scale experimental facility for seismic testing of structures which is part

of the European Seismic Engineering Research Infrastructures and whose construction was

partially financed by the European Union. The Earthquake Engineering and Structural

Dynamics Division operates this facility, pictured in Figure 3.1, and develops R&D activity in

the fields of Earthquake Engineering and Structural Dynamics.

Figure 3.1: LNEC Earthquake Engineering testing facility (Ferry Borges building)

The experimental activity carried out in the LNEC earthquake engineering testing facility, and

related research, aims at assessing the performance of structures subjected to dynamic and

seismic loadings. The tests are carefully setup in order to simulate on the models the same

conditions as in the real prototypes and measure all the relevant effects necessary for the


51

performance assessment. The instrumentation is calibrated and a record is made of all the

maintenance and interventions made. The preparation and operation of the shaking tables

follows standard protocols to achieve the targets proposed. The analysis process uses the most

advanced techniques in the fields of signal processing, dynamics of structures and earthquake

engineering to achieve optimum results. All these aspects contribute to guarantee that the tests

carried out meet the highest quality standards.

3.1 GENERAL INFORMATION ON THE LABORATORY

Table 3.1 – Name and location of the Laboratory Full Name of the Laboratory Núcleo de Engenharia Sísmica e Dinâmica de Estruturas

Abbreviated Name NESDE/LNEC

Address Av. Brasil, 101 1700-066 Lisbon

Location Lisbon

Country Portugal

Telephone +351218443824/3307

Telefax +351218443035

E-mail/www http://www.lnec.pt/LNEC/DE/NESDE/

Table 3.2 – Name and location of the parent organization Full Name of the Parent Organization Laboratório Nacional de Engenharia Civil

Address Av. Brasil, 101 1700-066 Lisbon

Location Lisbon

Country Portugal

Telephone +351218443000

Telefax

E-mail [email protected]

3.2 THE FACILITY: LNEC-3D SHAKING TABLE

The facility has a large testing room, shown in Figure 3.2, and includes two shaking tables,

one large triaxial and another one smaller uniaxial, and various other equipment for seismic

testing of structures. The triaxial shaking table, pictured in Figure 3.3, is capable of testing

large civil engineering structures subjected to earthquake motions up to collapse.


52

Figure 3.2: LNEC earthquake engineering testing room

Figure 3.3: LNEC-3D shaking table


53

3.3 GENERAL INFORMATION ON THE SHAKING TABLE

Table 3.3 – Name of the LNEC-3D shaking table Full Name of the Shaking Table LNEC TRIAXIAL SHAKING TABLE

Abbreviated Name LNEC-3D

Designer/Manufacturer LNEC and INSTRON

Year of Installation 1995

3.4 SHAKING TABLE DESCRIPTION

Table 3.4 – Type of shaking table

Longitud. X Transverse Y Vertical Z Pitch Roll Yaw

Uniaxial - - - - - -

Biaxial - - - - - -

Multiaxial Y Y Y N/A N/A N/A

Table 3.5 – Characteristics of the Platform

Size (m×m) 4.6 x 5.6 Weight (kN) 392 Material Steel

Type of Actuation Hydraulic

Table 3.6 – Characteristics of the Actuators

Manufacturer Total Force (kN) Number of units/axis

Longitudinal INSTRON 1250 1

Transverse INSTRON 750 2

Vertical INSTRON 375 1


54

Table 3.7 – Shaking table performances

Frequency Range Hz 0.1 – 40.0

Stroke (effective/maximum) Horizontal mmpp 290/400

Vertical mmpp 290/400

Max Velocity (nominal/limit) Horizontal

Transverse cm/s 70.1/121.5

Longitudinal cm/s 42.4/73.5

Vertical cm/s 41.9/72.6

Max Acceleration at bare table Horizontal

Transverse m/s² 18.75

Longitudinal m/s² 31.25

Vertical m/s² 9.38

Yaw Rotation degrees ° N/A

Velocity rad/s N/A

Pitch/Roll Rotation degrees ° N/A

Velocity rad/s N/A

Max Overturning Moment kN×m N/A

Max Specimen Dead Weight kN 392

Max Compensated Dead Weight kN 392

3.5 CHARACTERISTICS OF THE CONTROL SYSTEM

Type of Control Analogue Digital Mixed

Table 3.8 – Characteristics of the analogue part

Manufacturer LNEC

Type LNEC-CTL

Table 3.9 – Characteristics of the digital part

Hardware

Computer Host PC+NI PXI Real Time Controller+4 RIO FPGA Virtex-5

D/A Channels 8 ADC channels, 16 bit

96 configurable digital channels A/D Channels

Software Designer LNEC

Controlled motions Sinusoidal Random Shock Seismic

No. of Controlled Channels 3 3 3 3

No. of Acquisition Channels 6 6 6 6


55

3.6 COMPLEMENTARY FACILITIES

Floating Foundation Hydraulic System Bridge Cranes Capacity

Dimensions (m×m) - Electric Power (kW) 330 No. of Cranes 2

Weight (kN) - Flow Rate (l/min) 690 Max Load (kN) 392

Natural Freq. (Hz) - Pressure (MPa) 20.7 Useful Height (m) 8


56

4 Sensors technical data

The experimental assessment of a model performance under seismic actions imposed on the

shaking table requires the measurement of several different types of physical quantities like

displacements, accelerations, strains and forces. The sensors necessary for this includes

displacement transducers, accelerometers, strain gauges and load cells, all of them available at

the LNEC Earthquake Engineering and Structural Dynamics Division. Below is listed only

the subset of the existing instrumentation that is relevant for the tests carried out in the scope

of the present study.

4.1 DISPLACEMENT TRANSDUCERS

4.1.1 LVDT displacement transducers

RDP Electronics ACT2000C, ACT4000C and ACT6000 inductive displacement transducers,

having work strokes of +/-50mm, +/-100mm and +/-150mm, respectively, can be used for

measuring displacements. Figure 4.1 shows the general aspect of the inductive displacement

transducers available at LNEC, while Table 4.1 shows some of its main characteristics.

a) ACT captive guided

b) ACT series unguided

Figure 4.1: LVDT displacement transducers (source: RDP)


57

Table 4.1 – Characteristics of the RDP displacement transducers

Manufacturer RDP ELECTRONICS (www.rdpe.com)

Models available ACT2000, ACT4000, ACT6000

Stroke +/-50mm (ACT2000), +/-100mm (ACT4000), +/-150mm (ACT6000)

Sensitivity 15mV/V/mm (ACT6000) to 30mV/V/mm (ACT2000)

Energising supply 5Vrms, 5kHz

Linearity deviation 0.08% (ACT2000) to 0.3% (ACT6000)

4.1.2 Hamamatsu optical system

Optical displacement transducers HAMAMATSU C5949 (comprising F50mm lens, sensor

head and LED target) and HAMAMATSU conditioning PSH controllers C2399 (see

Figure 4.2) can be used for measuring 2D displacements on a plane perpendicular to the line

of sight (typically either on a horizontal or vertical plane). Table 4.2 shows some of the main

characteristics of the HAMAMATSU displacement transducers.

Figure 4.2: HAMAMATSU optical 2D displacement transducer

Table 4.2 – Characteristics of the HAMAMATSU displacement transducers

Type Spectral

Response [nm]

Measurement Points [-]

Sampling Frequency

[Hz]

Position Detecting Error [%]

Resolution [-] Error due to light

[%]

C2399-00 1 300

C5949 700 to 1150 1 to 7 300 ±1 1/5000 ±1


58

The cameras used detect the positions of the points concerned along the x- and y- axes with

respect to their absolute reference system [12]. Figure 4.3 shows a typical Hamamatsu system

configuration.

Figure 4.3: Hamamatsu system configuration [12]

4.1.3 Krypton K600 camera

The K600 camera system takes measurements in 3D and consists of three linear CCD

cameras. While two external cameras, present inside the instrument, detect the position y- and

z- coordinates of the point concerned, the central one calculates the x-coordinate. Thus the

position of an infrared LED is calculated by triangulation [19]. Figure 4.4 shows the Krypton

K600 camera mounted on a tripod.

Figure 4.4: The Krypton K600 camera [19]

The overlap area of the three linear CCD-cameras in the camera unit, results in an overlapping

pyramidal volume as shown in Figure 4.5. The top angle of the pyramid is 34° (+17° / -17°).


59

The K600 Krypton Camera offers a 17m3 field of view volume with accuracies up to ±0.1mm

for static and dynamic LED position measurements. According to the rule-of-thumb, the

lateral visibility limit (measured from the symmetry plane of the camera) is half the distance

from the camera [19].

Figure 4.5: Representative measurement volume of a Krypton K600 camera [19]

The field of view is divided into three accuracy zones which are determined based on the

distance from the camera as seen in Figure 4.6. Accuracy of LED dynamic measurements has

been observed to be higher for motions in the X-Y plane of the viewing volume.

Figure 4.6: Accuracy zones of the Krypton K600 camera system [19]

The acquisition laptop communicates with the camera controller to acquire the 3D position

data of the individual LEDs. The camera control unit synchronizes the LEDs with the

acquisition of the camera and serves as an interface between the laptop and the camera unit.

The strober distributes the control signals from the controller to individual LEDs and

translates them into pulse trains causing the infrared LEDs to flash. The strober can be

connected to the strober port of the controller or can be daisy-chained to another strober. The

LED’s are plugged into a strober and then attached to the object that the user wants to


60

measure. The best results are achieved if they are placed at large distances forming large

triangles as seen by the camera.

4.2 ACCELEROMETERS

Two main types of accelerometers are used in the LNEC Earthquake Engineering and

Structural Dynamics Division: Endevco, model 7290-A with variable capacitance

(Figure 4.7), and PCB Piezotronics, model 337A26 (Figure 4.8). The first type is mainly used

on the shaking table while the second one is mainly used on the mock ups. Both are high

frequency accelerometers adequate for the measurement of accelerations in dynamic and

seismic tests; their main characteristics are summarised in Table 4.3 and Table 4.4.

Figure 4.7: Endevco accelerometers

Figure 4.8: PCB Piezotronics accelerometers

Table 4.3 – Characteristics of the Endevco accelerometers

Manufacturer MEGGITT SENSORS (www.endevco.com)

Model 7290A-2 and 7290A-10

Sensitivity (at 100Hz) [mV/g] 1000+/-20 and 200+/-10

Measurement range [g peak] +/-2 and +/-10

Amplitude response at +/-5% [Hz] 0 to 15 and 0 to 500

Transverse sensitivity [% max] 2

Table 4.4 – Characteristics of the PCB accelerometers

Manufacturer PCB Piezotronics (www.pcb.com)

Model 337A26

Sensitivity [mV/g] 100

Measurement range [g peak] 100

Broadband resolution [g rms] 0.0001

Frequency range [Hz] 0.5 to 5000


61

4.3 LOAD CELLS

Two load cells from Instron were used to measure the forces applied to the wall panels. The

load cell is an electric transducer used to measure a force applied on a mechanical or

structural component, through the measurement of an electrical signal that varies due to the

deformation that produces such a force on the component itself. The relief of the mechanical

deformation takes place in an indirect manner, via a reading in mV or V, and subsequently

transformed into the correct unit of measurement.

An individual load cell is constituted of a hollow steel cylinder. Six strain gauges are

connected to the body of the load cell – four in the longitudinal direction and two in the

transverse direction relative to the axis of the cylinder. The strain gauges are all connected in

a full-bridge configuration to amplify the magnitude of the signal. Figure 4.9 shows the load

cells.

3

Figure 4.9: Instron load cells

4.4 ACQUISITION SYSTEM

The acquisition system available at the LNEC Earthquake Engineering and Structural

Dynamics Division comprises the following components:

Table 4.5 – NI PXI controller

NI PXI-8106 CONTROLLER

2.16 GHz Intel Core 2 Duo T7400 dual-core processor

Up to 46% higher performance than the PXI-8105 512 MB (1 x 512 MB DIMM) dual-channel 667 MHz DDR2 RAM standard, 4 GB maximum 10/100/1000 BaseTX (Gigabit) Ethernet, ExpressCard/34 slot, and 4 Hi-Speed USB ports

Integrated hard drive, GPIB, serial, and other peripheral I/O

Windows OS and drivers already installed; hard-drive-based recovery


62

Table 4.6 – NI PXI chassis

NI PXI-1052 CHASSIS

• 4 slots for 3U PXI modules and 8 slots for SCXI modules • Latest chassis technology • AUTO/HIGH fan-speed selector to optimize cooling and acoustics • 0 to 55 °C temperature range • 42 dBA acoustic emissions • Multiplexed operating mode for SCXI

• SCXI high-voltage analog backplane integrated internally

Form Factor PXI Platform, SCXI

PXI Bus Type PXI Hybrid Compatible Operating System / Target Windows, Real-Time

LabVIEW RT Support Yes

Power Supply AC

Number of Slots12

Number of PXI Peripheral Slots4

Maximum System Bandwidth132 MB/s Accepts both 3U PXI and CompactPCI ModulesYes Optional Front or Rear Rack Mountable Yes

Integrated ControllerNo

Remote Power-inhibit Control and Voltage Monitoring Yes Total Available Power 450 W Input Voltage Range 100..240 V Input Frequency Range 50/60 Hz Field-replaceable Power Supply No


63

5 Test setup

5.1 BUILDING MODEL TEST SETUP

The test setup for the building model was already shown in Figure 2.36. After the base slab

was attached to the shaking table the model was thoroughly instrumented in order to

characterize its behaviour during the tests.

The results of a shaking table test can be obtained via: (a) visual record of the damage,

immediately after the test or at a later stage through video and photographic data; (b) data

acquisition equipment attached to specific points of the model. The model was filmed with a

high frequency video recording system and photographed before the first stage and after each

different stage, in order to produce damage maps and their evolution along the experimental

test. All cracks were painted along the test using different colours to clearly record their

evolution.

Different instruments were installed for recording displacements, both absolute and relative,

and accelerations. The acquisition equipment uses sensors to transform physical quantities

(displacements, velocities and accelerations) in electric signals and the most commonly used

sensor in shaking table tests are accelerometers (ACC). In the present case, these are SDOF

(single degree of freedom) systems having an inertial mass that moves proportionally to the

amplitude of the acceleration of a moving body, which is converted into an electrical signal in

the form of voltage [13]. Displacement transducers (LVDT and infrared cameras) are

frequently used as well, although velocities and displacements can be obtained from the

integration and double integration, respectively, of the acceleration signals. There are several

types of ACC and two different ACC were used here: piezoelectric and capacitive, see

Figure 5.1. The main differences between both systems are the power supply, since

piezoelectric ACC need an external power source, and a limited range of 1000Hz in the

capacitive ACC. Piezoelectric ACC are also capable of measuring uniform acceleration

signals.


64

(a) (b) (c)

Figure 5.1: Accelerometers used in the shaking table tests: (a) piezoelectric from PCB [31], [32]; (b) piezoelectric from Wilcoxon [40]; (c) capacitive from Endevco [9]

Regarding the piezoelectric type, three different models from two different manufacturers

were attached to the structure, all sharing the same sensitivity (1000 mV/g±5%) and

measurement range (±5g). The NESDE has also ACC with a lower measurement range and

higher sensitivity that could provide more accurate results during the dynamic identification

tests, see Table 6.1, but due to their limited range these ACC could not be used during the

seismic tests. Switching ACC between dynamic identifications and seismic tests is not

recommended in such a complex test that involves several kilometres of cables to connect all

the ACC to the acquisition equipment, over forty ACC, all the laboratory technicians and at

least one full day. Switching potentiates mistakes and malfunctions of the highly sensitive

equipment in use, and increases exponentially the time needed to perform the test. The

capacitive ACC were pre-installed in the shaking table in each direction (longitudinal,

transverse and vertical). The definition of the instrumentation setup is based on the expected

response of the model to the input, obtained from preliminary studies [20], and the objectives

of the test, meaning that the instrumentation can be divided in two groups: (i) setup to acquire

the out-of-plane behaviour of the infill walls; (ii) setup to acquire the global behaviour of the

RC concrete structure. The out-of-plane behaviour of the infill walls was captured by a set of

ACC distributed in the surface of the wall, see Figure 5.2. The blind walls received nine ACC

each, with one line at mid-height and the other two at half distance to the RC beams and

columns and the infill walls at the East and west façades received ACC at a third of the height

below the opening. The North infill wall at the ground floor received ACC below the opening,

at a third of the height also and at the centre lines between the openings. A total number of

thirty-four ACC were used.


65

(a) (b) Figure 5.2: Accelerometers setup: (a) North and East façades; (b) South and West façades

To avoid damaging the infill wall, all the ACC were bolted to a wooden surface and then

glued to surface of the infill wall, see Figure 5.3 (a). The global behaviour of the RC structure

was captured using: (i) two Piezotronics ACC orthogonally placed in the Northeast and

Southwest corners of each RC slab, Figure 5.2 and Figure 5.3 (b); (ii) motion detecting

cameras placed in the Northwest and Southeast corners of the RC slab of the first storey and

Northeast and Southwest corners of the RC slab of the roof, see Figure 5.4 (a) and (b). Four

motion detection cameras were used, Hamamatsu Photonics C5949 [12], see Figure 5.4 (c),

capable of determining the position of an infrared led each, see Figure 5.4 (d). Both

components, camera and infrared led, are connected to a controller that conditions the

acquired electrical signal, see Figure 5.4 (e). The camera only detects the planar movement of

the infrared led, and each infrared led was connected to the above mentioned corners of the

slabs through steel lever supports, see Figure 5.4 (c).

(a)

(b)

Figure 5.3: PCB Piezotronics accelerometers: (a) at the infill walls; (b) at the corners of the RC slabs


66

It may seem redundant to use ACC and motion detection cameras at the corners of the tested

models since the ACC can provide both accelerations and displacements, but the acquired

data is used for different purposes due to the different sensitivity and resolution of the sensors.

As an example, the ACC are used during the dynamic identifications to obtain dynamic

properties of the model and their evolution during the test, while the motion detection

cameras define the inter-storey drifts more accurately during the seismic tests.

Two extra piezotronics ACC were placed in the RC foundation ring beam, one in each main

direction, in order to compare the acquisitions recorded by the ACC of pre-installed in the

shaking table and the base of the model. These two sets of recordings have to be the same,

excluding possible differences due to intrinsic characteristics of the acquisition equipment,

since the foundation of the model cannot have relative displacements with respect to the

shaking table.

(a)

(b)

(c)

(d)

(e)

Figure 5.4: Hamamatsu photonics c5949 [12]: (a) position of the Hamamatsu leds in the first storey; (b) position of the Hamamatsu leds in the roof; (c) camera and led at the corner of the structure; (d) infrared

led; (e) controller

North

Eas

t

West

South

Hamamatsuposition

North

Eas

t

West

South

Hamamatsuposition


67

The acquisition of the above mentioned equipment (model 1: 44 ACC on the model, 2 ACC

and 2 LVDT on the shaking table and 4 motion displacement cameras; models 2 and 3: 40

ACC on the model, 2 ACC and 2 LVDT on the shaking table and 4 motion displacement

cameras) was carried out using one SCXI chassis connected to a PXI-1001, both from

National Instruments. The ACC channels were conditioned using SCXI-1500 series from

National Instruments and a 481A02 module from PCB Piezotronics, see Figure 5.5 (a). The

acquisition equipment is located in the control room of the laboratory, along with the control

system of the shaking table, see Figure 5.5 (b).

(a)

(b)

Figure 5.5: Acquisition and control room: (a) from top to bottom: NI-SCXI-1001, PCB Piezotronics 481A02 and NI PXI-1052; (b) control room with the shaking table’s controls and the model’s acquisition

system

The K600 Krypton camera [19], see Figure 5.6 (a) to (c), was used on the upper infill wall at

the North façade, see Figure 5.6 (d). It is composed by three CCD cameras that are able to

track the 3D displacement of a set of infrared LED’s, which in this case were placed

throughout the infill wall. The objective was to obtain the out-of-plane deformed shape of the

infill wall, hence two LED’s were placed in the RC beam while the other twenty-two LED’s

were placed on the infill wall. The initial geometrical plan of the LED’s has to be defined,

using the Space Probe, while the acquisition is done using the manufacturers own software, at

an acquisition frequency of 200Hz.


68

(a)

(b)

(c)

(d)

Figure 5.6: K600 Krypton camera: (a) three CCD cameras; (b) Space Probe used to calibrate the initial geometrical plan of the LED’s; (c) acquisition control; (d) distribution of the LED’s along the infill wall on

the upper floor of the North façade


69

5.2 WALL PANELS TEST SETUP

The testing setup for the wall panels simultaneously uses the shaking table, one reaction wall

and the specially designed Testing device for Innovative Masonry infills (TIM), as

represented in Figure 1.4. Such unique testing setup was specifically designed for this test and

is mainly composed of a stiff steel caisson three-dimensional frame, shown in Figure 5.7 and

Figure 5.8, which moves rigidly with the shaking table. It is fixed to the upper beam in the

transverse direction, while a system of rollers allows for an independent motion in the

longitudinal direction (Figure 5.9).

Figure 5.7: Main steel caisson frames of TIM (construction phase)

Figure 5.8: Base columns of main steel frame with detail of bolted connection to the shake table

(construction phase)


70

Figure 5.9: Guiding system of RC frame upper beam and rollers for longitudinal motion (construction

phase)

Both the in-plane and out-of-plane motions should match a given floor response spectra, of

narrow band frequency content. The in-plane motion will enforce an inter-storey drift time-

history on the frame by restraining the upper beam, which is prestressed for withstanding

push-pull actions, and by imposing the displacement of the shaking table on the lower beam.

The upper beam is prevented from moving in the longitudinal direction through a strut

connection to the reaction wall. This connection between the strut and the reaction wall is

performed by a pyramidal support, as depicted in Figure 5.10, which distributes the strut

reaction on the wall. A long rod links the pyramidal support to the RC frame upper beam

through hinged connectors.

All beam-column joints are free to rotate in the plane of the infill, through special hinged base

supports (Figure 5.11 and Figure 5.12). On the other hand, the out-of-plane motion will

consist on a rigid-body vibration of both the upper and lower beams, reproducing the storey

absolute accelerations and thus inducing high-frequency inertia forces perpendicular to the

masonry panel and leading to a local vibration of the infill wall. Note that this shaking table

motion is transmitted to the top beam through the rigid steel caisson. The assembly process of

the test setup is shown in Figure 5.13 to Figure 5.15.

The design of TIM was controlled by the requirement of a very stiff behaviour in the

transverse direction, which was ensured by a vibration frequency above 20 Hz. A parametric

study was conducted using a finite element model with and without the wall panel, as

represented in Figure 5.16, resulting in the modes and frequencies of vibration shown in

Figure 5.17 and Figure 5.18.


71

Figure 5.10: Pyramidal support for strut connection between the RC frame and the reaction wall

Figure 5.11: Hinged base supports for the RC frame specimens (construction phase)

Figure 5.12: Hinged base and pyramidal supports on their final position


72

Figure 5.13: Assembly of TIM components on the shaking table

Figure 5.14: Positioning of TIM over the wall panel setup

Figure 5.15: Complete setup for wall panels tests


73

Figure 5.16: Schematic representation of the finite element models used for the design of TIM, taking into

account (right) or not (left) the wall panels

a) b)

Figure 5.17: Vibration modes of TIM without the wall panel contribution: a) longitudinal (f = 19.9 Hz); b) transverse (f = 33.8 Hz)

a) b) c)

Figure 5.18: Vibration modes of TIM with the wall panel contribution: a) longitudinal (f = 18.4 Hz); b) transverse (f = 23.1 Hz); c) torsional (f = 25.5 Hz)


74

The instrumentation of the wall panels comprises several different sensors:

i) Krypton K600 camera for measuring the out-of-plane deformations of the wall panel

using 16 leds (Figure 5.19)

ii) Video camera for measuring the RC node deformation using data image correlation

methods (Figure 5.20)

iii) Hamamatsu optical system for measuring the horizontal translation of the bottom and

top corners of the RC frame (Figure 5.21)

iv) 36 accelerometers for vibration monitoring of the shaking table, TIM, RC frame and

masonry wall panel (Figure 5.22)

v) Load cells for measuring the dynamic reaction on the strut connecting the reaction

wall and the top beam of the RC frame (Figure 5.23)

Figure 5.19: Out-of-plane wall panel deformation monitoring with Krypton K600 camera

Figure 5.20: Video camera and target points for data image correlation measurement of in-plane

deformations at one RC frame node


75

Figure 5.21: Hamamatsu setup for measuring the horizontal translations of the RC frame nodes

Figure 5.22: Accelerometer setup for RC frame out-of-plane vibration measurements

Figure 5.23: Load cells for strut reaction measurement


76

6 Seismic testing protocol

6.1 TESTING PROCEDURE

The test procedure consisted in a combination of dynamic and seismic tests, the former to

identify the dynamic properties of the buildings and the latter to assess their seismic

performance.

6.2 SHAKING TABLE TUNING

Before conducting a seismic test the shaking table has to be tuned. The tuning procedure is

carried out in an iterative adaptive process starting with a low level intensity and progressing

in several iterations to the final required signals. The procedure follows a sequence of steps:

Dynamic identification of the whole system (ST+model) using a low amplitude “pink”

noise in acceleration to obtain a frequency response function (FRF);

Start with a low amplitude displacement signal input;

Measure displacements and accelerations as output;

Obtain a feedback synthesized displacement signal using a selected cross-over

frequency between measured displacements and accelerations;

Deconvolution of the feedback signal through the system FRF;

Obtain the new “error” signal to use as input signal to the shaking table;

Iterate until input signal is tuned for shaking table;

Increase the input signal and repeat the process for the next stage.

Figure 6.1 shows an example of the definition of the parameters to compute the system

(ST+model) FRF for the shaking table tuning process for the hands on application.


77

Figure 6.1: Shaking table tuning application: definition of parameters

Figure 6.2 shows an example of the system (ST+model) FRF obtained with the “pink” noise

excitation. An example of the signal tuning process is shown in Figure 6.3.

Figure 6.2: Shaking table tuning application: FRF obtained

Input signal (displacement)

Output signals (displacement

and acceleration)

ST-model interaction

Oil column resonance

Delay between input and output


78

Figure 6.3: Signal tuning iterative process

The previous sequence of steps is performed when the model is already placed on the shaking

table. However, the scale and destructive nature of a shaking table experimental program

implies the non-repeatability of any test stage. Therefore, shaking table tests are usually pre-

calibrated with masses representing the model to be tested, see Figure 6.4. This initial

calibration is performed in order to define an adequate input based on the expected behaviour

of the ST+model system. The main issue of using these masses is that they should represent

as closely as possible the inertia distribution of the model.

Figure 6.4: Calibration of the input signals with masses attached to the shaking table

Previous drive

Next drive

Target signal

Acquired signal

Error correction factors

Error signal

Inverse FRF deconvolution


79

6.3 SEISMIC TEST SEQUENCE

The test sequence is obtained applying different scaling factors to the selected accelerograms

(section 7.1), while performing dynamic identification tests in between the different stages.

Table 6.1 describes the seismic test sequence adopted.

Table 6.1 - Shaking table test procedure for the building model

Stage Identification Description

1

DI 0 Initial dynamic identification test

DL Seismic test based on Damage Limitation Limit State

DI1 Dynamic identification test after first stage

2 SD Seismic test based on Significant Damage Limit State

DI2 Dynamic identification test after second stage

3 NC Seismic test based on Near Collapse Limit State

DI 3 Dynamic identification test after third stage

4 1.5xNC

Seismic test with an amplitude of 1.5 times the Near Collapse Limit State stage

DI 4 Dynamic identification test after fourth stage


80

7 Signal generation procedure for the shaking table

tests

7.1 BUILDING MODEL

Shaking table tests can be performed by inputting an earthquake record of a past event,

usually scaled, or an artificial accelerogram. Given the unique and randomness character of a

seismic event, it is difficult to find a suitable earthquake record as far as duration,

accelerations and frequency content are concerned. Using the stochastic tools available in

LNEC-SPA [26], it is possible to generate artificial accelerograms adapted to the code

spectra. In the present experimental work, six artificial accelerograms were generated, whilst

two other were scaled based on the generated ones, in order to obtain four stages of the

shaking table tests, with increasing amplitude, see Table 6.1. The accelerograms of the first

three stages were adapted to the response spectra (damping ratio equal to 5%) of each damage

state described in section 3 of EC8 [30], see Figure 7.1: Damage Limitation (DL - 225 YRP);

Significant Damage (SD - 475 YRP); Near Collapse (NC - 2475 YRP). Here, YRP is the

return period in years. The response spectra for each damage state is obtained by multiplying

the accelerations of the elastic response spectra, which corresponds to the SD state, by the

factor described in EC8 [29].

The last stage was defined as the maximum capacity of the table in terms of velocity, given

the size and mass of the model, and its YRP computed assuming 1.5 and a reference

YRP of 2475. As described in Chapter 3, the maximum weight supported by the table is 392

KN in order to achieve the maximum velocity and acceleration. The weight of the tested

model in the present experimental work (model, foundation RC ring beam and additional

masses) was nearly 434 KN. But this excess of mass did not influence the capacity of the

shaking table to fulfil the input of any of the stages.


81

Transverse direction (East-West)

Longitudinal direction (North-South)

Figure 7.1: Comparison between pseudo-acceleration response spectra of the accelerograms generated and the response spectra, already scaled following the similitude law of Cauchy-Froude, obtained from

EC8 [30]

Given the geographical situation of continental Portugal, in any design situation one obtains

two response spectra as seen in Figure 2.33: (i) type one corresponding to a scenario of a far-

field seismic action; (ii) type two corresponding to a scenario of a near-field seismic action. In

design, the envelope of the response using both spectra is used in order to obtain the seismic

design internal forces. Here, the input signals generated for each stage were adapted only to

type one (far-field seismic action) response spectrum since it provides higher accelerations in

the expected natural frequencies of the models.

One signal was input in each horizontal direction (N-S or longitudinal and E-W or transverse,

see Figure 3.2). The signals were uncorrelated, with approximately the same PGA (peak

ground acceleration), PGV (peak ground velocity) and PGD (peak ground displacement) and

duration of around 30 seconds in the intense phase, see Figure 7.2. Given the Cauchy-Froude

similitude law, see Table 2.2, the acceleration was not scaled ( 1.5 / for seismic

area 1.3) and the frequency, or time, was. Therefore, the generated accelerograms were

adapted to a response spectra with the time reduced by 1.5 .

0.01 0.1 10

2

4

6

8

10

12

14

16

18

20 Generated

input EC8

Acc

eler

atio

n (m

/s2 )

Period (s)

DL 225yrp

SD 475 yrp

NC 2475yrp

1.5xNC 4574yrp

0.01 0.1 10

2

4

6

8

10

12

14

16

18

20 Generated

input EC8

Acc

eler

atio

n (m

/s2 )

Period (s)

DL 225yrp

SD 475 yrp

NC 2475yrp

1.5xNC 4574yrp


82



Figure 7.2: Time histories of the input signal of stage 2 (SD 475 YRP) reduced at 1:1.5 scale using Cauchy-Froude’s similitude law (see Table 2.2)

0 5 10 15 20 25 30 35-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

Acc

eler

etio

n (m

/s2 )

Time (s)

PGA=1.82 m/s2

0 5 10 15 20 25 30 35-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

Acc

eler

atio

n (m

/s2 )

Time (s)

PGA=1.68 m/s2

0 5 10 15 20 25 30 35-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

Vel

ocity

(m

/s)

Time (s)

PGV=0.13 m/s

0 5 10 15 20 25 30 35-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

Vel

ocity

(m

/s)

Time (s)

PGV=0.11 m/s

0 5 10 15 20 25 30 35-25

-20

-15

-10

-5

0

5

10

15

20

25

Dis

plac

emen

t (m

m)

Time (s)

PGD=22.09 mm

0 5 10 15 20 25 30 35-30-25-20-15-10-505

1015202530

Dis

plac

emen

t (m

m)

Time (s)

PGD=26.73 mm


83

7.2 WALL PANELS

The same seismic inputs were used for the wall panels tests as in the building model, except

that these were at full scale. The comparison of the response spectra of both components with

the code one corresponding to the Significant Damage Limit State is represented in

Figure 7.3. These accelerograms were then applied at the base of a representative building

model, see Figure 7.4, in order to obtain the time-history of the wall panel at the 4th floor. The

most important vibration modes for this building are represented in Figure 7.5 for the

longitudinal direction and in Figure 7.6 for the transverse direction.

Figure 7.3: Comparison between pseudo-acceleration response spectra of the accelerograms generated

and the response spectra obtained from EC8 [30]

Figure 7.4: Representative building model for wall panel input time-history definition


84

Figure 7.5: Longitudinal modes of vibration 1 (1.35Hz) and 2 (4.28Hz)

Figure 7.6: Transverse modes of vibration 1 (2.78Hz) and 2 (10.94Hz)

The response time-history in terms of interstorey drift at the 4th floor for the seismic input

considered is shown in Figure 7.7 and corresponds to the shaking table motion to be applied

in the longitudinal direction. On the other hand, the absolute acceleration observed in the out-

of-plane direction is represented in Figure 7.8 and corresponds to the shaking table motion

transmitted by TIM to the masonry panel inside the RC frame.


85

Figure 7.7: Interstorey drift time-history for in-plane motion

Figure 7.8: Absolute acceleration time-history for out-of-plane motion


86

8 Identification technique

The dynamic parameters of the structure (modal frequencies, modal damping and modal

configurations) in the seismic tests are usually obtained by the modal analysis. For this

purpose white noise time-histories with low amplitude or impulse signals are required.

8.1 WHITE NOISE

Before the first stage and after each test stage, the building model is subjected to two inputs,

again orthogonally horizontal and uncorrelated, specially generated with the purpose of

obtaining the dynamic properties of the model (natural frequencies, mode shapes and critical

damping ratios) and their evolution along the experimental test, see Figure 8.1. As these

properties are directly related to the stiffness, the damage state of the structure can be

characterized. Dynamic identification tests under these conditions are normally referred to as

forced vibration tests, in opposition to the usual ambient vibration tests. When compared to

the seismic tests, see Figure 7.2, these accelerograms have lower accelerations and higher

frequency range and duration, and are not adapted to a particular response spectrum but

generated as a low amplitude white noise. Obviously, the dynamic identifications should not

introduce additional damage to the structure, and the maximum amplitude remains relatively

low. There is also a difference in the amplitude of the transverse and longitudinal signals,

with a PGA of 0.44 m/s2 and 0.80 m/s2 respectively, due to the different stiffness of the RC

structure in each direction, see Figure 2.32, which has single bays in the transverse direction

and double bays in the longitudinal direction. Hence, the longitudinal direction has a higher

stiffness and requires larger amplitudes of the input signal.


87



Figure 8.1: White noise signals for dynamic identification tests

8.2 IMPULSE SIGNAL

On the other hand, the wall panels’ dynamic identification tests were carried out using

impulsive signals instead. The frequency response gain H fk and phase f k spectra, as

well as the coherence function 2 f k were calculated taking into account the following

approach.

Sequences of data sampled in a time resolution h, are divided into q frames of n points with

the total duration of the sequence given by n h q 1 is then applied.

With the FFT procedure, estimates of one-sided power spectral density functions (psdf) for

the input, output and cross-spectral can be evaluated using the following expressions:

~, ,G

h

NXk i

xk i

2 2 for the input psdf

~, ,G

h

NYk i

yk i

2 2 for the output psdf

~,,

,,G

h

NX Yk i

x yk ik i

2

for the cross psdf between input and output

where X k i,

2 and Yk i,

2 is the squared modulus of FFT components, at frequency f k , for the

input and output respectively and for frame i. The complex conjugate of X k i, is expressed by

X k i, .

0 20 40 60 80 100 120 140 160 180-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

0.2

0.3

0.4

0.5

Acc

eler

atio

n (m

/s2 )

Time (s)

0 20 40 60 80 100 120 140 160 180-0.6-0.5-0.4-0.3-0.2-0.10.00.10.20.30.40.50.6

Acc

eler

atio

n (m

/s2 )

Time (s)


88

Taking into account the averaging process over all the q frames, the expected power spectra

estimates, at frequency f k , for the total duration of the sequences are evaluated by:

~ ~ ~

~ ~ ~

~ ~ ~

, , ,

, , ,

, , ,

, , , ,

Gq

G G G

Gq

G G G

Gq

G G G

kx x x x

y y y y

kx y x y x y x y

k k k q

k k k k q

k k k q

1

1

1

1 2

1 2

1 2

The gain factor and the phase factor of the FRF between input x and output y can now be

carried out using:

H fG

Gk

kx y

kx

,

fG

Gk

kx y

kx y

tanIm

Re

,

,

1

Finally, coherence estimates are evaluated by:

2

2

fG

G Gk

kx y

kx

ky

,


89

9 Analysis of results

9.1 BUILDING MODEL TEST RESULTS

9.1.1 Initial test results

The building model (hereafter termed model 3) is equal to model 2, with the difference that

masonry walls have no bed joint reinforcement but have a reinforced rendering nailed to the

RC frame on both sides. Hence, model 3 represents another possibly future solution for RC

frames with masonry infills. Figure 9.1 presents the position, and label, of the accelerometers

in model 3. In this model, again the accelerometers were placed on the exterior only since the

infill walls have only one leaf.




Model 3 was subjected to the four test stages as models 1 and 2, but the fourth was not

successful due to technical problems. The transportation, done using a crane, and the first

testing stage (225 YRP) did not visually damage the mode. After stage 2 (475 YRP) the

model presented cracks in the mortar rendering in all four corners, starting at the base of the

RC column, and between the jambs on the intermediate columns of the East and West

façades, see Figure 9.2. Small cracks starting at the corners and moving towards the RC frame

BNE 2L

BNE 1L

N1 4

N1 1N1 3

N1 2

INP L

N1 5

BSW 2L

S2 6

S2 9

S2 3

S2 7

S2 1S2 4

BSW 1L

S1 4

S1 7

S1 1S1 6S1 3

S1 9

S2 2

S2 8

S2 5

S1 8

S1 2

S1 5E1.1 2

E1.1 3

E1.2 1

E2.1 3

E1.1 1

E2.1 1

E2.1 2

BNE 2T

BNE 1T

E2.2 2E2.2 1

INP T

BSW 2T

BSW 1T

W1.2 1W1.1 1


90

of some of the openings at the ground floor were also visible, while the first floor presented

no visual damage.

After stage 3 (2475 YRP), see Figure 9.3, the cracks in the mortar at the corners of the models

extended and small pieces of mortar rendering fell, see Figure 9.4 (c) and (d). The cracks in

the jambs of the East and West façade also were further extended. New cracks surrounding

the ground floor infills of the North and South façades appeared, along with some cracks in

the infill wall at the first storey of the North façade, mainly between the openings. Overall,

the model presented light damage, see Figure 9.4 (a) and (b), and the cracks, mainly at the

corners, seemed to affect only the mortar.

North South

East West

Figure 9.2: Crack patterns of model 3 after stage 2 (475 YRP) (Note: the drawn lines on the RC frame represent damage on the mortar rendering applied to the RC frame)


91

North South

East West

Figure 9.3: Crack patterns of model 3 after stage 3 (2475 YRP) (Note: the drawn lines on the RC frame represent damage on the mortar rendering applied to the RC frame)

(a) (b)

(c) (d)

Figure 9.4: Damage in model 3 after stage 3 (2475 YRP): (a) infill wall at the ground floor of the North façade; (b) infill wall at the upper floor in the East façade; (c) damaged mortar rendering at the Southeast

corner; (d) damaged mortar rendering at the Southwest corner


92


The dynamic identifications performed during the tests of model 3 presented the same quality

as the previous models, with very high coherences between the input and output signals,

leading to the identification of the same five mode shapes as in the previous models, see

Figure 9.5, although the first transverse and first longitudinal changed positions. Since model

3 shares the same geometry with the previous models, and the same structural materials as

model 2, this change is most likely associated to an undesirable interaction between the model

and the shaking table due to different stress levels applied on the connecting bolts or

geometrical imperfections in the RC foundation ring beam. Another reason can be the

interaction between masonry infill and frame, which depends on workmanship and the actual

execution of each infill. As for the other three modes, (first) torsional and second longitudinal

and transverse, the order is the same as in the previous models. Also, as in the previous

models, the torsional mode shape was not possibly correctly captured, see Figure 9.5, as the

roof RC slab is not rotating, while the first floor RC slab is rotating. In the second transverse

mode the South façade also presented a small longitudinal movement.

The variation of the peaks in the FRF’s was followed along the several dynamic

identifications, see Figure 9.6 (a), confirming the good quality of the results and allowing for

damage detection through the frequency decrease in all five modes along the three test stages,

see Figure 9.6 (b). Until the second stage, the longitudinal direction presented no frequency

decrease, the transverse direction presented an average 5.1% frequency decrease and the

torsional mode presented a 5.5% frequency decrease, in comparison to DI0. After stage 3

(2475 YRP), the average decrease in the longitudinal direction was 15.75% and the average

frequency loss in the transverse direction was 24.0%, which is not in agreement with model 2.

A possible reason for this higher decrease in the transverse direction is the slightly higher

recorded PGA in that direction when compared to longitudinal recorded PGA, see Figure 9.24

and Figure 9.25, which is the opposite situation of the previous tests. Another possible reason

is the influence of the interface between the masonry and the frame in the response, as

addressed before. The torsional mode presented a 31.1% frequency decrease, possibly

associated to a loss of connection between the infill walls and the RC frame since the high

frequency of the torsional mode is very dependent on the infill walls. The dynamic data is in

accordance with the observed slight damage.


93

1st Longitudinal Mode (6.26Hz)

1st Transverse Mode (10.01Hz)

Torsional Mode (28.02Hz)

2nd Longitudinal Mode (32.55Hz)

2nd Transverse Mode (35.63Hz)


(a) (b)

Figure 9.6: Frequency change along loading stages: (a) variation of the FRF’s along the test of model 3 at the accelerometer BNE – 1L; (b) evolution of the frequencies along the test of model 3 and their final


Table 9.1 presents the experimental estimation of the damping ratios along the several

dynamic identifications. As in the previous models, none of the identified mode shapes had

3 4 5 6 71

2

3

4

5

6.26 Hz6.26 Hz

DI 0 DI 1 DI 2 DI 3

5.49 Hz

Gai

n F

acto

r

Frequency (Hz)

1st Longitudinal

6.26 Hz

DI 0 DI 1 DI 2 DI 30

5

10

15

20

25

30

35

40

27.8 Hz(21.9%)

27.2 Hz(16.3%)

35.6 Hz

32.5 Hz

7.4 Hz(26.0%)5.3 Hz

(15.2%)

19.3 Hz(31.1%)

10.0 Hz

6.3 Hz

Freq

uenc

y (H

z)


1st Longitudinal

1st Transversal Torsion

2nd Longitudinal

2nd Transversal

28.0 Hz


94

the expected damping increment along the tests, confirming the difficulties in the

experimental estimation of this parameter. The vulnerability curves presented in Figure 9.7

confirm that the longitudinal direction presented no considerable damage until the second

stage, although cracks between the jambs at the ground level of the East and West façades

were visible. After stage 3 (2475 YRP), the transverse direction presented considerably more

damage when compared to the longitudinal one, which is clearly associated to the different

Energy Input, more than 30% higher. Overall, the results are in agreement with the crack

patterns. It is also noted that the damaged indicator reached a level far lower than the other

models, which reached 1.0 (collapse) for model 1 and 0.8 for model 2. This seems to indicate

that the capacity reserve of this model is still much higher than the model 2.

Table 9.1 - Experimental damping ratios of model 3

1st Longitudinal 1st Transverse Torsion 2nd Longitudinal 2nd Transverse

DI 0 (%) 6.12 4.80 1.93 2.14 0.99

DI 1 (%) 5.55 2.62 3.00 1.79 0.87

DI 2 (%) 4.80 16.75 2.99 2.06 2.63

DI 3 (%) 8.15 5.97 4.87 3.86 2.43

Figure 9.7: Seismic vulnerability curves of model 3 in the transverse and longitudinal directions, using the PGA and Input Energy as input

0 1 2 3 4 5 6 7 8 90.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35 1st Transversal Torsional

2nd Transversal

Dam

age

indi

cato

r d

PGA (m/s2)

0 1 2 3 4 50.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35 1st Transversal Torsional

2nd Transversal

Dam

age

indi

cato

r d

Input Energy (J)

0 1 2 3 4 5 6 7 8 90.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Dam

age

indi

cato

r d

PGA (m/s2)

1st Longitudinal

2nd Longitudinal

0 1 2 3 4 50.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

Dam

age

indi

cato

r d

Input Energy (J)

1st Longitudinal

2nd Longitudinal


95


Figure 9.8 presents the frequency decrease of the infills on South façade and the infill at the

ground floor of the North façade. The blind infill walls have a higher frequency than the one

with openings and the infill wall at the upper level has a higher frequency than the one at the

ground level. Even though the model presented no damage after the first stage (225 YRP), the

infill walls at the ground floor immediately lost stiffness, presumably due to the loss of

connection to the RC frame or some reparation of the stiff rendering from the RC frame. After

stage 2 (475 YRP) the model presented cracks, mainly at the corners and jambs but some at

the openings, and the infill walls of the ground floor presented a frequency decrease, on

average, of 5%. The results are in agreement with the crack patterns, since most of the cracks

seemed to be only at the mortar rendering. Until this point the infill wall at the upper floor of

the South façade presented no frequency decrease.

After stage 3 (2475 YRP), the South façade presented a higher frequency loss when compared

to North one, which is not in total agreement with the crack patterns as the infill wall at the

ground floor of the North façade presented more cracks. The upper infill wall in South façade

presented no cracks and the ground level one presented cracks mainly at the bottom, while the

cracks on the corners of the façade were on the mortar rendering of the RC columns. Overall,

the infill walls did not present extensive cracking which is in accordance with the loss of

stiffness. Again, the frequency loss in this model did not reach the level of model 1 (25% at

stage 3 and then collapse) and model 2 (up to 65%).

Figure 9.8: Evolution of the frequencies of the infill walls in the North and South façades along the test of

model 3 and their final variation in respect to DI 0

DI0 DI1 DI2 DI3

50

55

60

65

70

61.6 Hz(7.6%)

56.5 Hz(11.8%)

53.2 Hz(5.5%)

66.7 Hz

64.1 Hz

56.3 Hz


Fre

quen

cy (

Hz)



96


Figure 9.9 presents the interstorey displacements and drifts, in the longitudinal and transverse

directions, and in the three test stages. As expected, the displacements and drifts increased

with the seismic amplitude. The ground level recorded similar displacement and drift values

in all three stages, while the roof level recorded slightly higher displacements in the

longitudinal direction, in agreement with the first mode shape, even though the second mode

presented a higher frequency decrease. Both directions recorded increasingly smaller

differences between the ground and first levels, and on the last stage the transverse direction

recorded a higher drift at the upper level. The results are in agreement with the observed

damage which was evenly distributed through the ground level. Note, again, that the drifts are

much lower than in model 1 (0.5% at stage 3 and then collapse) and model 2 (up to 4%).

Longitudinal direction Transverse direction



Figure 9.10 presents the recorded PGA in the RC structure at the slab levels and in all infill

walls, for each direction and test stage. The PGA recorded at the RC structure increased with

the seismic amplitude, the roof level recorded higher PGA’s when compared to the first floor

0 2 4 6 8 10 12

storey 1(2 meters)


Displacement (mm)

roof(4 meters)

00 1 2 3 4 5 6 7 8 9 10


storey 1(2 meters)

Displacement (mm)

roof(4 meters)

0

0.0 0.1 0.2 0.3 0.4


storey 1(2 meters)

Drift (%)

roof(4 meters)

00.0 0.1 0.2 0.3 0.4


0

storey 1(2 meters)

roof(4 meters)

Drift (%)


97

RC slab, and the highest values in the second and third stages (225 and 475 YRP) were

recorded in the longitudinal and transverse directions, respectively. The amplifications in the

RC frame present a moderate change, with the exception of the roof in the longitudinal

direction that almost was halved. For the masonry infills, the longitudinal direction presenting

a slight increase from stage 1 (225 YRP) to stage 2 (475 YRP) and a 42.2% decrease from

stage 2 to stage 3 (2475 YRP), even if not consistently for all walls, while the transverse

direction presented a nearly constant amplification. Again, this confirms the observation in all

models that amplifications do not follow a clear trend, and the initial amplification provides a

reasonable estimate of the dynamic response in the non-linear range.

The PGA recorded at the infill walls also increased with the seismic amplitude. On the

transverse direction, the infill walls at the first storey recorded higher values in all stages but

on the longitudinal direction the maximum PGA’s were recorded in the infill wall at the

ground floor of the North façade. The longitudinal direction recorded higher values when

compared with the transverse direction, even though the input PGA was higher in the

transverse direction, because of the considerably higher amplifications.


RC structure


7 6 5 4 3 2 1 00 5 10 15 20 25

1

2

3

N1 S1 S2


Stage

3.0 2.5 2.0 1.5 1.0 0.5 0.00.0 2.5 5.0 7.5 10.0 12.5 15.0

1

2

3

E 1.1 E 1.2 E 2.1 E 2.2 W1.1 W 1.2

Amplification

Stage

Acceleration (m/s2)

5 4 3 2 1 00 5 10 15 20

1

2

3



Stage


98

Out-of-plane PGD and deformation of the infill walls

Figure 9.11 and Figure 9.12 present the out-of-plane deformation of the infill walls of the

South façade in all test stages, with the first storey infill wall presenting a similar the same

deformed shape, with the maximum deformation at the centre and lower part of the wall, and

PGD in the first (225 YRP) and second stages (475 YRP). In the third stage (2475 YRP), the

maximum deformation area was slightly smaller and the PGD remained the same as in the

previous stages. This unaltered state along the test was expected as the infill wall did not

present any cracks and only a 5.6% frequency decrease, in respect to DI0, was recorded in

stage 3. As for the infill wall at the ground floor, in the first and second stages the wall

presented the same deformed shape and similar PGD, with the maximum value recorded at

the top of the wall, but on the third stage the wall presented a higher flexibility, with at least

twice the displacements of the previous stages, and an increment in the PGD area. This infill

wall presented the highest frequency decrease, in the longitudinal direction, and cracks

surrounding the frame, which confirms the agreement between the results.

Figure 9.13 presents the out-of-plane deformation of the infill wall at the upper floor of the

North façade, with the PGD increasing in all stages and with the area of higher displacements

concentrated between the openings. These results are in agreement with the observed damage,

as the cracks in this infill wall were also concentrated between the openings and started in

stage 2. The PGD of the rest of the infill walls of model 3 are presented in Figure 9.14, where

all infills present an increment in the displacement with the seismic amplitude and the infill

walls at the upper level present higher displacements than infill walls at the ground floor.


99


Stage 3 (2475 YRP)

Figure 9.11: Out-of-plane deformation of the infill wall at the ground level of the South façade of model 3 in mm


Stage 3 (2475 YRP)

Figure 9.12: Out-of-plane deformation of the infill wall at the first storey of the South façade of model 3 in mm

13.3411.44

9.5317.625 5.719

3.8131.906

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

13.2311.35

9.475 7.600 5.725 3.8501.975

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

13.29

13.2911.33

9.3637.4005.438

3.4751.513

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

1.3253.300

5.275 7.2509.22511.2013.18

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

-0.33131.638

3.6065.5757.5449.513

11.48

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

24.98

26.34

23.61

27.7029.0630.4331.79

26.34

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)


100


Stage 3 (2475 YRP)

Figure 9.13: Out-of-plane deformation of the infill wall at the first storey of the North façade of model 3 in mm.

Figure 9.14: Out-of-plane PGD of the North, East and West infill walls of model 3 in mm

9.1.2 Complementary test results

As stated before, due to technical problems in the shaking table, model 3 was only tested until

stage 3 (2475 YRP) but given the light overall damage presented by the structure and infill

walls, both in absolute value and when compared with the other models, the model was

submitted to the first three stages again. The model was not removed from the shaking table,

as the retest was performed the day after the first test, so no damage was introduced in the

transportation and no changes were made to the boundary conditions. The results of these new

three stages, in the model hereafter denominated as model 3B, cannot be directly compared to

0.8540

0.76650.85400.6790

0.9415

0.5915

1.0291.1170.7665

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window

0.300.50

0.69

0.89

1.11.3 1.5

0.50

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window

1.11.5

1.9

2.2

2.63.0

3.4

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window

0 5 10 15 20 25 30 35

1

2

3

E 1.1 E 1.2 E 2.1 E 2.2 W 1.1 W 1.2 N 1

Sta

ge



101

the previous three shaking table tests as far as the values of the parameters are concerned, but

the damage pattern of the RC structure and infill walls and the collapse mode developed after

the structure has been severely damaged can be compared.

Overall damage and rendering removal

After the three test stages, model 3B still did not present as much damage as the previous

models, see Figure 9.15 (a) and (b), as no new cracks appeared but the ones at the corners, see

Figure 9.15 (c) and (d), and jambs, see Figure 9.15 (e) and (f), only at the ground floor,

widened considerably and parts of mortar rendering were expelled. The reinforced plaster

became loose, as if completely disconnected from the infill walls, and as if it was standing

only because of the connection provided by the additional masses, which is unreasonable and

not a true structural feature. This also confirms the importance of the connections of

reinforced plaster to the walls, which should cross the wall, and not only use nails, as done

here. The upper level presented no significant damage with only small cracks at the corners of

the openings.


102

(a) (b)

(c) (d)

(e) (f) Figure 9.15: Damage in model 3B after stage 3 (2475 YRP): (a) North façade; (b) South façade; (c) crack

and mortar rendering loss at the Northwest corner; (d) crack and mortar rendering loss at Northeast corner; (e) crack at the a lateral jamb in the infill wall at the ground floor of the East façade; (f) crack at

the interior jambs in the infill wall at the ground floor of the East façade

After testing, the additional masses bolted to the infill walls were removed, and as these were

working simultaneously as an attachment of the reinforced plaster to the infill wall, it was

possible to simply remove the reinforced rendering as a whole on both sides of the infill

walls, without the use of any specific equipment. This confirmed that the reinforced plaster

was completely detached from the infill walls and that the fixings of the additional masses

worked as connectors, preventing the rendering from collapse. Moreover, careful analysis of

the un-plastered infill walls showed that these presented limited damage, see Figure 9.16 (a)

and (b), but were mostly disconnected from the RC frame, see Figure 9.16 (c) and (d). Hence,

there was a major contribution reinforced plaster was preventing the out-of-plane collapse of

the infill walls. As for the RC structure, no cracks were detected at mid-height of the RC


103

columns, but only at the upper connection to the beams, see Figure 9.16 (e). The RC was very

flexible under these conditions, meaning that the reinforced plaster was also rather important

in preventing the collapse of the entire system.

(a) (b)

(c) (d)

(e) Figure 9.16: Damage in the infill walls and RC structure after the reinforced rendering removal at the ground floor: (a) infill wall of the North façade; (b) South infill wall with a compression crush at right

down corner; (c) gap between one of the West the infill wall and RC frame in the West wall; (d) infill walls of the West façade; (e) extensive cracking at the upper column-beam connection in the Northwest corner

Modal frequencies of the RC structure and infill walls

Figure 9.17 presents the frequency decrease of the RC structure and of three infill walls in

model 3B along the three test stages. The initial dynamic identification, DI0, corresponds to

DI3 in model 3, and the percentage at the last frequency is the total loss since DI0 of model 3,


104

which is the undamaged state of the model. In the RC structure, the identification of the (first)

torsional and second longitudinal and transverse modes was not possible after stage 3 (2475

YRP). The first order transverse mode presented a higher stiffness loss when compared to the

longitudinal mode. After stage 2 (475 YRP) the modes presented 0.2Hz of difference between

them and after stage 3 the transverse direction presented a higher stiffness loss. At the end of

the test, the transverse direction presented a 73.5% frequency loss, while the longitudinal

direction presented a 48.6% decrease, when compared to the undamaged state. This difference

can be associated to the infill walls at the ground floor, which on the transverse direction

presented a clear disconnection to the RC frame at the jambs, while on the transverse

direction the infill walls also presented corner crushing due to compression and extensive

damage around the door.

The infill wall at the upper level of the South façade presented an extra 12.2% of frequency

decrease during the three stages, which add to the previous damage totalizing 19.8%, when

compared to the undamaged state. This infill wall did not present any visible damage, and the

loss of stiffness is associated to the loss of connection between the infill wall and the RC

frame. The infill walls at the ground floor presented a similar and considerably higher

frequency loss at the end of stage 3, when compared to the upper level one, although the

South infill wall presented the highest loss at the first stage (225 YRP) and the North one

presented the highest loss in the last stage. The stiffness loss presented is in agreement with

the observed crack pattern after the reinforced plaster removal.

Figure 9.17: Evolution of the frequencies along the test of model 3B, and their final variation in respect to DI 0 of model 3, at the RC structure and infill walls in South façade and ground level of the North façade

DI 0 DI 1 DI 2 DI 30.02.55.07.5

10.012.515.017.520.022.525.027.530.0

27.5 Hz(22.7%)

26.3 Hz(19.3%)

27.8 Hz27.2 Hz

3.2 Hz(48.6%)

2.7 Hz(73.5%)

17.6 Hz(37.3%)

7.4 Hz

5.5 Hz

Freq

uenc

y (H

z)


1st Longitudinal

1st Transversal Torsion

2nd Longitudinal

2nd Transversal19.3 Hz

DI0 DI1 DI2 DI320

25

30

35

40

45

50

55

60

65

53.4 Hz(19.8%)

26.1 Hz(53.7%)

22.1 Hz(65.6%)

61.6 Hz

56.5 Hz

53.2 Hz


Fre

quen

cy (

Hz)



105


The interstorey displacements and drifts increased with the seismic amplitude, see

Figure 9.18, until the second stage (475 YRP) the longitudinal and transverse directions

presented similar values and the ground level values had a tendency to increase when

compared to the upper level ones. On stage 3 (2475 YRP) the ground level recorded higher

displacements and drifts than the first level. While the first level values were similar in both

direction, the ground floor of the transverse direction presented higher displacements and

drifts. These results are in agreement with the observed crack patterns and dynamic data, as

the transverse direction presented the highest stiffness loss and the corner crushing due to

compression at the ground infill wall of the South façade is associated to in-plane movement

of the RC frame, in this case in the transverse direction. Again, note that the maximum drift at

this stage is in the range of 2%.


Figure 9.18: Interstorey displacements and drifts of model 3B

0 2 4 6 8 10 12 14 16 18 20 22

storey 1(2 meters)


Displacement (mm)

roof(4 meters)

00 5 10 15 20 25 30 35 40


storey 1(2 meters)

Displacement (mm)

roof(4 meters)

0

0.0 0.5 1.0 1.5 2.0


storey 1(2 meters)

Drift (%)

roof(4 meters)

00.0 0.5 1.0 1.5 2.0


0

storey 1(2 meters)

roof(4 meters)

Drift (%)


106


The PGA values recorded on the RC structure and infill walls, see Figure 9.19, presented no

surprises since there was an increment with the seismic amplitude and the upper levels of the

infill walls recorded higher values when compared to the ground floor ones. As for the

amplifications, there is always a small increment in the recorded values from stage 1 (225

YRP) to stage 2 (475 YRP) and a larger decrement from stage 2 to stage 3 (2475 YRP). In the

RC structure, the transverse direction did not present a higher amplification loss due to the

stiffness loss. The infills at the first level of the transverse direction presented the highest

amplification loss, while on the longitudinal direction it were the infills on the South façade

that presented the highest amplification loss.


RC structure

Figure 9.19: Recorded PGA and amplifications at the infill walls and at the RC structure for each test stage of model 3B

Out-of plane PGD and deformation of the infill walls

Contrary to what happened during the test of model 3, in model 3B the damage presented by

the infill walls along the 3 test stages altered their out-of-plane deformed shape. The infill

wall at the ground level of the South façade, see Figure 9.20, presented very similar values in

7 6 5 4 3 2 1 00 5 10 15 20 25

1

2

3

N1 S1 S2


Stage

4 3 2 1 00 2 4 6 8 10 12 14 16

1

2

3

E 1.1 E 1.2 E 2.1 E 2.2 W1.1 W 1.2

Amplification

Stage

Acceleration (m/s2)

5 4 3 2 1 00 5 10 15 20

1

2

3



Stage


107

the first and second stage (225 and 475 YRP) with the maximum values at the lower part of

the wall in the first stage and then concentrating at lower lateral sides and mostly and the

corner with the crushed corner, see Figure 9.16 (b). During stage 3 (2475 YRP) the

displacements increased diagonally from the lower left corner to the upper right corner. The

infill wall at the upper level of the South façade, see Figure 9.21, which did not present any

visible damage at the end of the test, presented the highest deformation at the top in the first

stage and on stage 2 and 3 at the sides and top. The infill wall at the upper level of the North

façade, see Figure 9.22, still presented the highest deformation between the jambs, as during

the tests of model 3, in the stage 1, and on stage 2 and stage 3 moved to the lower and left

side, respectively, even though the only visible damage on the infill were very small cracks

connecting both openings.

Figure 9.23 presents the PGD recorded at all other infill walls of model 3B, with the infill

walls of the upper level recording the highest value in all stages and the infill wall at the

ground floor of the North façade presenting the highest PGD of the ground floor infill walls,

within the same values of the South infill wall.


Stage 3 (2475 YRP)

Figure 9.20: Out-of-plane deformation of the infill wall at the ground level of the South façade of model 3B in mm

9.1007.800

6.500 5.200 3.9002.600 1.300

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

7.4256.163

4.900

8.688

3.6382.3751.113

8.688

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

8.605

9.600

7.610

10.59

6.615

11.5912.58

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)


108


Stage 3 (2475 YRP)

Figure 9.21: Out-of-plane deformation of the infill wall at the first storey of the South façade of model 3B in mm


Stage 3 (2475 YRP)

Figure 9.22: Out-of-plane deformation of the infill wall at the first storey of the North façade of model 3B

in mm

2.2803.360

4.440

5.5206.6007.6808.760

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

8.980

8.980

9.630

8.330

10.2810.9311.5812.23

8.980

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)31.23

30.51

30.51

31.23

31.9432.66

29.80

33.3734.09

31.94

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50

Infill length (m)

Infi

ll h

eigt

h (m

)

2.72.9

3.1

2.7 2.4

2.7

2.2 2.01.7

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window

3.1

1.1

1.1

3.1

5.27.39.3

1.1

11 135.2

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window

169.28

168.82

168.35

167.89

167.42

166.95166.49

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.50.00

0.25

0.50

0.75

1.00

1.25

1.50


Infill length (m)

Infi

ll h

eigh

t (m

) window window


109

Figure 9.23: Out-of-plane PGD of the North, East and West infill walls of model 3B in mm

9.1.3 Comparison of test results

Following, a comparison of the experimental results of all the models is presented but given

that not all were subjected to the last stage or collapsed during it, the comparison is done

considering the first three stages only.

Target/acquired comparison and shaking table performance

The comparison between the target (input), presented in the previous chapter, and the data

acquired (output) by the accelerometers placed in the shaking table, also described in the

previous chapter, was done using the Peak Ground Acceleration (PGA) and a set of five

integral parameters (Root Mean Square Acceleration (RMSA), Arias Intensity (AI) and Input

Energy (IE) [18], [7] in each of the two main horizontal directions (longitudinal or North-

South, and transverse or East-West), as follows:

(5)

1 (6)

2

(7)

0 5 10 15 20 25 30 35 40 45 50

1

2

3

E 1.1 E 1.2 E 2.1 E 2.2 W 1.1 W 1.2 N 1

Sta

ge



110

(8)

Here, is the time history acceleration, is the duration of the signal or earthquake, is

the gravity acceleration, is the mass of the model and is the time history of velocity.

This comparison is needed despite the calibration tests performed with inert masses since the

models are not inert and, therefore, they influence the behaviour of the shaking table. The

parameters used for the comparison were chosen due to their importance in seismic

engineering and structural dynamics. The PGA is commonly used in standards for design

purposes, even if associated to a response spectrum. A single peak value cannot accurately

represent a seismic action, as seismic actions with the same PGA can result in different

damage scenarios, hence the use of integral parameters. All integral computations depend on

the duration of the seismic action, a parameter with more importance regarding the level of

destruction than the amplitude of the accelerations, particularly for unreinforced masonry

structures. A seismic action with shorter duration and higher accelerations (e.g., Ancona, Italy

in 1972) is likely to be less destructive than a seismic action with longer duration and lower

accelerations (e.g., Mexico City, Mexico in 1985).

As far as the PGA is concerned, see Figure 9.24 (a) and Figure 9.25 (a), and analysing all

models simultaneously, the transverse direction of the shaking table outperformed the

longitudinal direction, since smaller differences were recorded with respect to the target. In

the longitudinal direction the acquired PGA of the models was, on average, 42% higher in

respect to the target, and on the transverse direction the acquired PGA was only 20% higher

than the target value. These computations were done considering all four stages for models 1

and 2 and only the first three stages for models 3 and 3B. Model 3 was subjected to the fourth

stage, but as it can be seen in Figure 9.24 (a) and Figure 9.25 (a), the acquired PGA was 62%

lower in respect to the target in the longitudinal direction and 4% higher in the transverse

direction. This was due to a technical problem in the shaking table and therefore this stage

was not considered. Model 3B was not subjected to the fourth stage due its extensive damage

and possible collapse. But technical problems remained since the recorded PGA in the

transverse direction of stage 3 was twice the value of the target, thus larger than the other

models for this direction.


111

If the comparison between the target and the acquired response is done using the RMSA, see

Figure 9.24 (b) and Figure 9.25 (b), the deviations are considerably lower as the RMSA is not

so dependent on peak values. The average deviation of all models was the same for the

longitudinal and transverse directions and equal to 2%. The AI, see Figure 9.24(c) and

Figure 9.25 (c), is a measure of the earthquake destructiveness based on the RMSA but more

dependent on peak values therefore the deviations increase to 34% and 32% in the

longitudinal and transverse directions, respectively. The IE, see Figure 9.24 (d) and

Figure 9.25 (d), is the only parameter in which the deviation is negative, meaning the acquired

response was lower than the target. The average deviation for the models was 10% and 2% in

the longitudinal and transverse directions, respectively. The integral parameters confirmed the

technical problems above referred during the fourth stage of model 3, excluding it from the

present analysis.

In conclusion, the differences between the acquired and target data is within acceptable limits

in parameters dependant on peak values and good considering parameters more dependent on

the duration of the motion. The better results are in the transverse direction, when compared

to the longitudinal one, due to the characteristics of the shaking table, already presented in the

previous chapter, which has two actuators in the transverse direction and only one actuator in

the longitudinal direction. Therefore, the transverse direction is more sensitive and able to

better replicate the target.


112

(a)

(b)

(c)

(d)

Figure 9.24: Longitudinal direction target/acquired comparison (Fourier filter: 1-40Hz): (a) PGA; (b) Root Mean Square Acceleration; (c) Arias Intensity: (d) Input Energy

(a)

(b)

(c)

(d)

Figure 9.25: Transverse direction target/acquired comparison (Fourier filter: 1-40Hz): (a) PGA; (b) Root Mean Square Acceleration; (c) Arias Intensity: (d) Input Energy

Stage 1 Stage 2 Stage 3 Stage 40

2

4

6

8

10

12

Target Model 1 Model 2 Model 3 Model 3B

PGA

(m

/s2 )

EC8 Design


5

10

15

20

25

30

Ari

as I

nten

sity

(m

/s)



5

10

15

20

25

30

Ari

as I

nten

sity

(m

/s)



2

4

6

8

Inpu

t Ene

rgy

(J)



2

4

6

8

10

12

PGA

(m

/s2 )

Target Model 1 Model 2 Model 3A Model 3B

EC8 Design

Satge 1 Stage 2 Stage 3 Stage 40.0

0.5

1.0

1.5

2.0

2.5

Roo

t Mea

n S

quar

e (m

/s2 )



5

10

15

20

25

30

35

Ari

as I

nten

sity

(m

/s)



2

4

6

8

Inpu

t Ene

rgy

(J)



113

Modal frequencies of the structure and infill walls

For the sake of simplicity, and given the amount of data already provided for each model

individually, the comparison of the modal frequencies and their variation is done through the

vulnerability curves that relate the damage indicator, , with the input energy applied at the

base and in the case of the RC structure only for the first mode in each main direction.

Regarding the seismic vulnerability curves of the RC structure, see Figure 9.26, and on the

transverse direction, model 3 presented a higher damage indicator when compared to models

1 and 2 in stage 2, 27% and 49% respectively, and in stage 3, 35% and 49% respectively,

which is unexpected due its higher resistance concrete and rebar and reinforced plaster infills.

On stage 2 all three models were subjected to the same input energy at the base, but on the

third stage model 3 was subjected to an input energy 16% and 14% higher than models 1 and

2, respectively, which can justify the considerably higher damage indicator of model 3. Model

1 presented a damage factor higher than model 2 in all three stages, as expected given the

models’ lower strength materials and weaker unreinforced infills.

In the longitudinal direction, model 3 presents the lowest damage indicator in all three stages,

but on stage 3 the model was subjected to an input energy 28% lower than models 1 and 2,

hence the 64% and 72% lower damage indicator, respectively, due to mechanical issues

already addressed. Model 1 presented a damage indicator 80% higher in stage 2, when

compared to model 2, but 25% lower on stage 3 which is unexpected. Both models were

subjected to stage 4, and while the model 1 collapsed, which means a unitary damage

indicator, model 2 did not collapse even though it was heavily damaged. One of the factors

that contributed to this improved performance of model 2 is the continuous connection of the

reinforced infills to the RC frame provided by the metallic connecters, while on model 1 all

infills collapsed out-of-plane and the RC structure failed shortly after.


Figure 9.26: Vulnerability curves of the 1st mode in each main direction of the RC structure of the three tested models

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.50.00

0.05

0.10

0.15

0.20

0.25

0.30Stage 3 Model 1

Model 2 Model 3

Dam

age

indi

cato

r d

Input Energy (J)

Stage 2

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.0

0.1

0.2

0.3

0.4

0.5 Model 1 Model 2 Model 3

Dam

age

indi

cato

r d

Input Energy (J)

Stage 3

Stage 2


114

Figure 9.27 presents the average vulnerability curves of the infill walls on the North façade,

with openings, and South façade, without openings, of all three tested models and in the first

three stages. The same issues with the input energies are patent here, with model 3 being

subjected a 28% lower input in stage 3 when compared to models 1 and 2. As far as the North

infills are concerned, on the second stage model 3 presented the highest damage indicator and

models 1 and 2 similar values, while on stage 3 model 3 presented a damage indicator 71%

and 63% lower than models 1 and 3, respectively. From stage 2 to stage 3, the bed joint

reinforced infill walls presented a damage indicator 21% higher when compared to the

unreinforced ones, even though the unreinforced ones collapsed during the last stage and the

reinforced ones did not, meaning that the bed joint reinforcement does not contribute to the

decrease of damage but rather prevent the out-of-plane collapse of the infill. In the South

façade, and in stage 2, the unreinforced infill walls presented a higher damage coefficient

when compared to the reinforced ones, while the bed joint reinforced infills presented a

higher damage indicator than the plaster reinforced ones. This order was kept unaltered from

stage 2 to stage 3, as well as the percentage difference between the models.

North infill walls South infill walls

Figure 9.27: Average vulnerability curves of the North and South infill walls of the three tested models


Figure 9.28 presents the interstorey displacements of the first three stages of the tested models

in both main directions. In the second stage, corresponding to the design earthquake, and in

the transverse direction, model 1 presented displacements 70.4% and 13.8% higher, on

average, than models 2 and 3, respectively, as expected since it has a lower strength concrete

and rebar and it was designed with the older design standard [36] instead of the Eurocodes

[29] and [30]. On the third stage, model 1 presented a displacement 20.6% and 51.9% lower,

on average, than models 2 and 3, respectively. Model 3 presented displacements 56.6%

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.00

0.05

0.10

0.15

0.20

Dam

age

indi

cato

r d

Input Energy (J)

Model 1 Model 2 Model 3

Stage 2

Stage 3

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.00.00

0.05

0.10

0.15

0.20

Dam

age

indi

cato

r d

Input Energy (J)


Stage 2

Stage 3


115

higher, on average, during stage 2 and 31.3% higher, on average, during stage 3. In the

longitudinal direction, model 1 did not present the highest displacement on stage 2, but it also

presented the lowest displacements on stage 3. As for modes 2 and 3, in the second stage the

difference between both is 36.4%, with model 3 presenting higher displacements, and in the

third stage the difference is 20.6%, with model 2 presenting higher displacements. This

inversion is mostly associated to the lower input energy in model 3 during the third stage,

which led to lower damage in model 3. The use of different design standards between model 1

and models 2 and 3 is very clear in both directions during stage three, with model 1 presenting

lower displacement capacity or ductility than the other two models.

Figure 9.29 presents the interstorey drifts of the first three stages of the tested models in both

main directions. In the second stage all models present a similar drift on both storeys, or the

second floor has a higher drift, while on stage 3 the ground floor tends to have higher drifts,

introducing considerably more in-plane damage in the infill walls and reducing their out-of-

plane capacity. Model 2 presents lower drifts, when compared to model 3, along the two

stages which indicate that the bed joint reinforcement might contribute to higher in-plane

stiffness when compared to the reinforced plaster. The absence of a clear trend of model 1

regarding the other two models, confirms that the strength of the RC structure and the

presence of reinforcement on the infill walls has a low influence on the in-plane behaviour of

the frames, aggravated by the fact that the most influential parameters, which are the

geometry of the frame and the geometry, presence of openings and strength of the infill, are

the same in all three tested models.


116

Transverse direction Longitudinal direction S

tage

2

Sta

ge 3

Figure 9.28: Interstorey displacements of the three tested models in the transverse and longitudinal directions


Sta

ge 2

S

tage

3

Figure 9.29: Interstorey drifts of the three tested models in the transverse and longitudinal directions

0 1 2 3 4 5 6 7


storey 1(2 meters)

Displacement (mm)

roof(4 meters)

00 1 2 3 4 5


storey 1(2 meters)

Displacement (mm)

roof(4 meters)

0

0 1 2 3 4 5 6 7 8 9


storey 1(2 meters)

Displacement (mm)

roof(4 meters)

00 1 2 3 4 5 6 7 8 9 10 11 12


storey 1(2 meters)

Displacement (mm)

roof(4 meters)

0

0.0 0.1 0.2 0.3 0.4 0.5 0.6

roof(4 meters)


0

storey 1(2 meters)

Drift (%)

0.0 0.1 0.2


storey 1(2 meters)

Drift (%)

roof(4 meters)

0

0.0 0.1 0.2 0.3 0.4

roof(4 meters)


0

storey 1(2 meters)

Drift (%)

0.0 0.1 0.2 0.3 0.4


storey 1(2 meters)

Drift (%)

roof(4 meters)

0


117

PGA of the RC structure and infill walls

Figure 9.30 presents the maximum recorded acceleration and amplification regarding the

input acceleration, on average, for the infills walls in each façade and the RC structure in each

main direction for the three tested models. Model 3 presents higher accelerations and

amplifications along the test, when compared to models 1 and 2, which is unexpected and

possibly unrelated to the characteristics of the model but rather associated to the mechanical

problems above mentioned. As far as the infills are concerned, another justification is

associated to the bolting of the accelerometers to the infills, which is actually only bolted to

the reinforced plaster, and as seen in Figure 9.15 and Figure 9.16 detached from the infills and

remained attached to the RC structure. This means that at later stages along the test some

accelerometers are measuring only the accelerations of a stiff piece of reinforced mortar

plaster and not the actual infill wall, leading to higher accelerations and amplifications.

Analysing models 1 and 2, in the longitudinal direction, the infill walls with openings at the

North façade the unreinforced infill walls of model 1 presented higher amplification when

compared to the reinforced ones of model 2, but lower maximum acceleration, while the blind

infill walls of the South façade present similar results. The infill walls with openings in the

East and West façade follow the ones in the North façade, especially in stage 3, with the

unreinforced infills of model 1 presenting higher amplifications and lower maximum

accelerations when compared to the bed joint reinforced infill walls of model 2.

As for the RC structure, and analysing again models 1 and 2, no trend can be found and both

models present similar results, and the justification is associated to the similarities of both

models regarding the most influential parameters.


118

North façade South façade

East façade West façade

RC structure Storey 1 RC structure roof

Figure 9.30: Average recorded PGA and amplifications at the infill walls and at the RC structure for each test stage of all tested models

6 5 4 3 2 1 00 5 10 15 20 25 30

1

2

3



Stage

5 4 3 2 1 00 5 10 15 20 25

1

2

3



Stage

2 1 00 2 4 6 8 10 12

1

2

3



Stage

2 1 00 2 4 6 8 10

1

2

3



Stage

2 1 00 5 10 15

1

2

3



Stage

4 3 2 1 00 5 10 15 20

1

2

3



Stage


119

10 Main conclusions

Model 1 had a good performance during the seismic standard PGA of stage 2 (475 YRP) with

no visible damage, even though the dynamic data presented frequency loss, on the RC

structure and infill walls. The soft-storey collapse mechanism developed during the fourth

stage is highly undesirable, with a low energy dissipation capacity and brittle collapse [40]. It

was not clear if the collapse mechanism developed was due to the detailing of the RC

structure imposed by the design standard [35] or the influence of the infill walls, since RC

columns also presented hinges at mid-height just before the collapse of the structure. Some of

the thin blocs applied in the RC columns and beams to avoid thermal bridges, a very common

solution in the Portuguese built patrimony, cracked and fell during stage 3 (2475 YRP). The

double leaf unreinforced infill walls underperformed during the last stage (4574 YRP),

collapsing out-of-plane by rotating as a rigid body around the base line of the model. The

interior and exterior leaves presented a similar seismic behaviour.

Model 2 presented a good seismic performance when subjected to the seismic standard PGA

during stage 2 (475 YRP), although the RC structure registered a small decrease in the modal

frequencies. The model did not collapse during the last stage (4574 YRP) but presented

severe, and most likely irreparable, damage and the RC structure developed a soft-storey

mechanism. The seismic standard used in the design of model 2 [29], [30] clearly details the

structure in order for the development of a beam-sway mechanism [40] by forcing the hinges

to appear at the beams and not the columns. Attending also the mid-height cracks found in all

the RC columns after stage 4, it is possible to assume that the infill walls and their openings

influenced negatively the seismic behaviour of the RC structure. The mortar rendering applied

to the RC frame was severely damaged after stage 3 (2475 YRP), specially at the corners of

the model. The single leaf infill walls with bed joint reinforcement connected to the RC frame

had a very good seismic performance, with no visible damage for the seismic standard

accelerations during stage 2 (475 YRP). After the last stage (4574 YRP) none of the infill

walls collapsed out-of-plane, even though the ones with openings at the ground floor


120

presented damage beyond repair. It was clear that the bed joint reinforcement prevented the

out-of-plane collapse due to its connection to the RC frame, otherwise the infill walls would

collapse as a rigid body.

Model 3 presented no considerable damage in the RC structure due to the accelerations

prescribed in the seismic standard during stage 2 (475 YRP), and after stage 3 (2475 YRP) the

damage was also light. After the retesting using the first three stages, with considerable loss

of stiffness, the model still presented visual light damage. After the reinforced rendering

removal, the RC columns did not present hinges at the extremities neither cracks at mid-

height, hence no undesirable collapse mechanism was developed. Given that model 3 was

designed following the EC2 and EC8 [29], [30], it is safe to say that the infill walls did not

influence undesirably the seismic behaviour of the RC structure. The infill walls presented

light damage after all the stages, even though the dynamic data presented a clear stiffness

loss. This was due to detachment of the infill wall from the reinforced rendering, allowing the

wall to move and not influence the behaviour of the RC structure, fact confirmed after the

rendering removal as the infill walls presented barely any cracks but were detached from the

RC structure. Although the reinforced rendering concealed the damage from the infill wall, it

also prevented the out-of-plane collapse because it was applied on both sides of the infill wall

and nailed to the RC frame and infill wall.

As for the comparison between models, some parameters are able to present differences

between the chosen reinforcement solutions of the infill walls, but a global trend that

highlights the better performance of one regarding others does not exist. It is clear that

model 1, designed using an older generation of standards [] and unreinforced solutions for the

infill walls had an undesirable seismic performance by collapsing, and the models designed

the Eurocodes [29], [30] did not, but along the test the amount of damage and global

behaviour were similar. This is due to the same geometry of the RC structure and infill walls

which are the parameters that most influence the structural behaviour, given that the strength

of the RC used was not significantly different. The detailing of the RC using [29] led to a

higher ductility in models 2 and 3 when compared to the detailing of model 1 using [35], and

the use of reinforcement in the infill walls attached to the RC frame prevented the

out-of-plane collapse, and possibly the collapse of the RC structure as well, but did not led to

less damage along the tests when compared to the unreinforced solution.


121

References

[1] Angel, R. Behavior of reinforced concrete frames with masonry infills. PhD Thesis, Dept. of Civil

Engineering, University of Illinois, Urbana-Champaign, USA, 1994.

[2] Bedell, R. J. and D. P. Abrams. Scale relationships of concrete columns. Research Report, Dept. of Civil

Engineering, Environment and Architecture, University of Colorado, Boulder, USA, 1983.

[3] Calvi, G. M., Bolognini, D. and A. Penna. Seismic performance of masonry-infilled r.c. frames: benefits

of slight reinforcements. Proc. of the 6th Portuguese Conference on Seismology and Earthquake

Engineering, Guimarães, Portugal, 2004.

[4] Calvi, G. M., Bolognini, D. and A. Penna. Design of masonry structures with bed joint reinforcement.

Proc. of the Seminar on masonry walls, Guimarães, Portugal, 2007.

[5] Carvalho, E. C. Seismic testing of structures. Proc. of the 11th European Conference on Earthquake

Engineering, Paris, France, 1998.

[6] Coelho, E. and F. Carvalho. Ensaios sísmicos – Instalações do LNEC. Journal Engenharia e Vida, No. 10,

2005, pp. 51-55.

[7] Cosenza, E. and G. Manfredi. Damage indices and damage measures. Progress in Structural Engineering

and Materials, Vol. 2, No. 2, 2000, pp. 50-59.

[8] Drysdale, R. G., A. A. Hamid, et al. Masonry Structures: Behavior and Design. Boulder, Colorado,

2008.

[9] Endevco. 7290A-10 characteristics. At https://www.endevco.com/7290a-10/ (accessed on 30/6/2013).

[10] Ewins, D. Modal testing: theory, practice and application. Research Studies Press, Hertfordshire, UK,

2000.

[11] Flanagan, R. D. and R. M. Bennett. Bidirectional behaviour of structural clay tile infilled frames.

Journal of Structural Engineering, Vol. 125, No. 3, 1999, pp. 236-244.

[12] Hamamatsu. C5949 characteristics. At http://pdf.datasheetcatalog.com/datasheet/hamamatsu/C5949.pdf

(accessed on 30/6/2013).

[13] He, J. and Z. F. Fu. Modal Analysis. Linacre House, Jordan Hill, Oxford, 2001.

[14] Jorquera, L. G. Estudio experimental sobre la resistencia de muros de albañileria sometidos a cargas

horizontales. Journal of IDIEM No. 3, 1964.

[15] Klingner, R. E. and V. V. Bertero. Infilled frames in earthquake-resistant construction. Research

Report, University of California, Berkeley, USA, 1976.

[16] Klingner, R. E. and V. V. Bertero. Earthquake resistance of infilled frames. Proc. of the American

Society of Civil Engineers, EUA, 1978.

[17] Komaraneni, S. Out-of-plane seismic behaviour of brick masonry infilled panels with prior in-plane

damage. PhD Thesis, Dept. of Civil Engineering, IIT Kanpur, India, 2009.

[18] Kramer, S. L. Geotechnical earthquake engineering. Prentice Hall, New Jersey, USA, 1996.


122

[19] Krypton Electronic Engineering n.v., K400/K600 - Hardware & Software Guide (Krypton Help Pages),

Leuven, 2003.

[20] Leite, J. Avaliação da Segurança Baseada em Deslocamentos: Aplicação a Edifícios de Betão Armado

sem e com Paredes de Preenchimento. MSc Dissertation, Dept. of Civil Engineering, University of

Minho, Guimarães, Portugal, 2009.

[21] Lemaitre, J. and R. Desmorat. Engineering damage mechanics: Ductile, creep, fatigue and brittle

failures. Springer, 2005.

[22] LNEC. Main characteristics of triaxial shaking table. At

http://www.lnec.pt/organizacao/de/nesde/ptriaxialcaracteristicas (accessed on 30/6/2013).

[23] Mainstone, R. J. On the stiffness and strength of infilled frames. Proc. of the Institution of Civil

Engineers, No. 49, London, UK, 1971.

[24] Mallick, D. V. and R. P. Garg. Effect of openings on the lateral stiffness of infilled frames. Proc. of the

Institution of Civil Engineers, No. 49, London, UK, 1971.

[25] Mehrabi, A. B. and P. B. Shing. Performance of masonry-infilled r/c frames under in-plane lateral

loads. Research Report, Dept. of Civil Engineering, Environment and Architecture, University of Colorado,

Boulder, USA, 1994.

[26] Mendes, L. LNEC-SPA: Signal Processing and Analysis Tools for Civil Engineers. Earthquake

Engineering and Structural Dynamics Division, Dept. of Structures, National Laboratory for Civil

Engineering (LNEC), Lisbon, Portugal, 2008.

[27] Miranda, E. and S. Taghavi. Approximate floor acceleration demands in multistory buildings. I:

Formulation. Journal of Structural Engineering, Vol. 131, No. 2, 2005, pp. 203-211.

[28] Miranda, E. and S. D. Akkar. Generalized interstory drift spectrum. Journal of Structural Engineering,

Vol. 132, No. 6, 2006, pp. 840-852.

[29] NP EN 1992-1-1. Eurocódigo 2 - Projecto de estruturas de betão - Parte 1-1: Regras gerais e regras

para edifícios. Portuguese Institute for Quality (IPQ), Lisbon, Portugal, 2010.

[30] NP EN 1998-1. Eurocódigo 8 - Projecto de estruturas para resistência aos sismos - Parte 1: Regras

gerais, acções sísmicas e regras para edifícios. Portuguese Institute for Quality (IPQ), Lisbon, Portugal,

2010.

[31] PCB Piezotronics. 393A03 characteristics. At http://www.pcb.com/products.aspx?s=393A03 (accessed on

30/6/2013).

[32] PCB Piezotronics. 626B13 characteristics. At http://www.pcb.com/products.aspx?s=626B13 (accessed on

30/6/2013).

[33] Penna, A., Lagomarsino, S. and A. Galasco. A nonlinear macroelement model for the seismic analysis

of masonry buildings. Earthquake Engineering and Structural Dynamics, Vol. 43, No. 2, 2013, pp. 159-

179.

[34] Pereira, M. F. P. Avaliação do desempenho das envolventes dos edifícios face à acção dos sismos. PhD

Thesis, Dept. of Civil Engineering, University of Minho, Guimarães, Portugal, 2013.

[35] REBAP. Regulamento de Estruturas de Betão Armado e Pré-Esforçado. Diário da República, Portugal,

1983.

[36] RSA. Regulamento de Segurança e Acções para Estruturas de Edifícios e Pontes. Diário da República,

Portugal, 1983.

[37] SAP2000. Structural Analysis Program, Version 15.0.0. Computers and Structures, Inc. Berkeley, EUA,

2013.


123

[38] Seismosoft. SeismoStruct: A computer program for static and dynamic nonlinear analysis of framed

structures, Version 6.9.5. Available at http://www.seismosoft.com, Pavia, Italy, 2013.

[39] Stafford Smith, B. Model test results of vertical and horizontal loading of infilled frames. ACI Journal,

No. 65, 1968.

[40] Wilcoxon. 799M characteristics. At

http://www.wilcoxon.com/vi_index.cfm?PD_ID=14&P_ID=14&search=y (accessed on 30/6/2013).

[41] Zarnic, R. and M. Tomazevic. Study of the behaviour of masonry infilled reinforced concrete frames

subjected to seismic loading - Part II. Research Report, University of Ljubljana, Slovenia, 1985.

Rel SERIES NTUA Final · seismic response of reinforced concrete (RC) frames with innovative solutions for masonry infill walls, considering both the in-plane and out-of-plane behaviour

Documents