The 3rd International Conference on Earthquake Engineering and Disaster Mitigation 2016

(ICEEDM-III 2016)

Development of Numerical Model for Pushover Analysis of

Confined Masonry with and without Opening
M. Suarjana1, D. Kusumastuti1, C. A. Riva’i2 & N. I. Pratiwi2
1
Department of Civil Engineering, Institut Teknologi Bandung (ITB), Indonesia
2
Former Student, Department of Civil Engineering, ITB, Indonesia
Email: [email protected]

Abstract
A numerical study was conducted to develop and calibrate numerical method suitable for nonlinear monotonic
static analysis (pushover analysis) of confined masonry wall with and without opening. In this study, nonlinear
beam-column element is used to model reinforced concrete members. The masonry panel is modeled as
nonlinear strut. The calibration of parameters of the nonlinear strut that can represent the response of the
masonry panel is the main objective of the study. The calibration process is performed by comparing analysis
results with experimental results reported previously. In the previous experimental study, two full scale confined
masonry wall specimens (3 x 3 m2) were constructed following the general construction practice in Indonesia.
The first model was a solid panel without opening, while the second was constructed with an 80 x 120 cm2 to
model a typical window opening. The models were subjected to cyclic in-plane lateral loads.

Keywords: confined masonry, opening, numerical model, pushover analysis

1 Introduction
Masonry walls are commonly used in Indonesia for low story residential building as well as for large
building, including multi stories building. In one story houses, plain masonry walls without
reinforcement or confinement are common. This type of construction is very prone to damage due to
earthquake. Confinement by reinforced concrete frame has proven to improve the performance of the
masonry wall. However, due to unsatisfactory construction quality and inadequate seismic resistant
details, observations after recent earthquakes have shown that confined masonry houses in Indonesia
were also prone to damage, even collapse. The vast numbers of houses damaged by earthquake
greatly affect the number of casualties. Therefore, improving the performance of residential buildings
under seismic load has become main priority to reduce fatalities and economic losses. This study is a
part of our research series to study and improve the performance of typical residential houses in
Indonesia to resist seismic load. Several experimental and numerical studies of confined masonry and
reinforced concrete frame infilled with masonry walls were conducted to better understand the
performance of simple house structure. The studies focused on the parameters of the structural
elements such as masonry properties, mortars and concrete used. The studies also covered details of
masonry wall confined by reinforced concrete frame resistance to the seismic loads. Various studies
also suggested improvements on confining frame and the connection of walls to the frame to increase
the capacity and ductility of masonry wall systems.

In a confined masonry residential house, besides the reinforced frame elements, the masonry walls
contribute very significantly to the strength and stiffness of the system. Methods to predict the
strength and stiffness of a masonry wall are not simple. The problem becomes more complicated
when the masonry wall has opening. Typically openings are required on the masonry wall for
windows and doors, and these openings may significantly reduce the performance of the panels. In
Indonesia, the wall openings are typically rimmed by wooden frame that installed mainly for window
and door frame. Structurally, the wooden frame can act as confining elements that protect the
weakened masonry wall around the opening.

An experimental study was carried out to verify the reduced strength and stiffness of the masonry wall
panels due to openings, and to understand their effects on the overall performance. The experimental
study was presented in Suarjana et. al. (2012). The experimental results were used as the bases for
developing and calibrating numerical model presented in his paper.

2 Summary of Experimental Setup and Results

An experimental study was carried out to verify the reduced strength and stiffness of the masonry wall
panels due to openings, and to understand their effects on the overall performance. The experimental
works are presented in Suarjana et al. (2012). Two full-scale (3m × 3m) confined masonry wall
specimens, where one of them is a solid confined masonry wall panel without opening, while the
second model was constructed with an 80 x 120 cm2 wooden framed opening in the middle of the wall
to model a typical window opening. Both specimens were constructed following the general
construction practice in Indonesia.

2.1 Material Characteristics

The specimens were constructed following the general construction practice in Indonesia, with the
sequence of: reinforcement assembling, concrete foundation pouring, brick laying, and finally
concrete frame pouring. The wall specimens constructed by moderate quality red bricks and average
concrete and mortar quality. The material properties based on averaged material test are presented in
Table 1.

2.2 Experimental Setup

The details of the two wall specimens are shown in Figure 1 and Figure 2. Specimen 1 represents
typical confined masonry wall, where typically the concrete members are made with the same
thickness with the masonry for simple and efficient formwork. The height/width of the beam and
columns are specified such that the area of the concrete elements equal to 15 x 15 cm2, which is the
minimum concrete area according to the house construction guideline The Ministry of Public Work,
Indonesia (2009). Specimen 2 is similar to specimen 1, except for an 80 x 120 cm2 framed window
opening in the center region. The window frame is made from 5 x 10 cm2 timber with average quality
commonly available and used for residential building. Both specimens were not finished by plaster.

The lateral cyclic load was applied at the top beam-column joint. The load came from a hydraulic jack
attached to the reaction wall. Figure 3 shows the test setup for the lateral cyclic load of confined
masonry wall. The response of the wall specimen was measured using strain gauges and LVDT
(Linear Variable Displacement Transducers). Cracks development and other damages were recorded
and marked for each load cycle.
 Table 1 Material properties.
Material Properties
Diameter (deformed bar) 10.0 mm
Longitudinal
Steel yield stress fy 384.9 MPa
rebar diameter (un-deformed bar) 7.8 mm,
Transversal
yield stress fy 350.9 MPa
beam size 100 x 225 mm2
column size 100 x 225 mm2
Concrete mixture by volume proportions
1:2:3:1
(cement : sand : aggregate : water)
compressive strength 15.32 MPa
55 × 100 × 205
size
Burned Clay Brick mm3
compressive strength 3.5 MPa
brick spacing 15 mm
mixture by volume proportions
Mortar 1:5:1
(cement : sand : water)
compressive strength 8.74 MPa
opening size 800 × 1200 mm2
Opening
wooden frame size 50 x 100 mm2
Finishing bare brick wall, not plastered

Figure 4 shows the loading cycles by displacement control that were applied during each experiment
which follows ACI 374.1-05 recommendation. The specimens were subjected to a series of increasing
cyclic lateral load. The largest drift applied on the structure was 3.5 percent or equivalent to top
displacement of 105 mm.

40 d

3150
225 225
2700
225

Beam 100 x 225

Longitudinal : 4 D 10
Transversal Ø 8
3000
2776

Column 100 x 225

Longitudinal : 4 D 10
Transversal : Ø 8
Ø 12 - 125
332

2 D13

D13 - 125

3900

Figure 1 Details of wall specimen 1.

Figure 2 Details of wall specimen 2.

 Figure 3 Experimental setup.
Loading Schema

150 5

100
3

2
50
Deformation (mm)

Drift (%)
0 0

-1

-50
-2

-3
-100

-4

-150 -5
Step Simpangan
Figure 4 Deformation cycles for displacement controlled loading scheme.

2.3 Hysteretic Behaviour

Figure 4 and Figure 6 present the hysteretic diagram of load-displacement for Model 1 and 2 during
the tests. Figure 7 shows the comparison of envelope curves of the hysteretic diagrams. The two
specimens appear to have similar hysteretic behavior, with slight difference in the maximum load.
The peak load of specimen 1 occurred at drift 1.75% with peak load 5.18 ton, while the peak load of
specimen 2 occurred at drift 1.0 % with peak load 5.83 ton. The slight difference in maximum load
probably caused by material and workmanship variations. There is not much difference observed in
stiffness degradation and ductility of both models.
 6

0
Load (tf)

-150 -100 -50 0 50 100 150

-2

-4

-6

-8
Displacement (mm)

Figure 5 Hysteretic load-displacement of Model 1.

2
Load (tf)

0
-100 -50 0 50 100 150
-2

-4

-6

-8
Displacement (mm)

Figure 6 Hysteretic load-displacement of Model 2.

70
60
50
Load (kN)

40
30
Model 1
20
Model 2
10
0
-10 0 25 50 75 100 125

-20
Displacement (mm)

Figure 7 Comparison of hysteretic load-displacement envelopes.

3 Structural Element Analysis Model
In this analysis we include linear and nonlinear models for both reinforced concrete members and
masonry panel.

3.1 Reinforced Concrete Element Model

The reinforced concrete elements are modeled as beam-column element with possible plastic hinge
formation on element ends. The element stiffness is based on actual concrete dimensions and material
properties. Plastic hinge properties are based on actual concrete and reinforcement properties and
dimensions.

3.2 Masonry Element Model

At low lateral load and small lateral deformations, the masonry panel and reinforced concrete frame
act as monolithic composite structural element. As the lateral deformation increase, the masonry panel
crack and form various possible failure mechanism. There are several failure modes for infill masonry
wall (Tomaževič 1999):

 Sliding shear failure of masonry walls

 Compression failure of diagonal strut
 Diagonal tensile cracking. This is not a general failure. Higher lateral forces can be supported
by the above failure modes.
 Tension failure mode (flexural), which is not usually a critical failure mode for infill walls

On the basis of comprehensive experimental research, various models have proposed for masonry
panel. Classical finite element model based on theory of elasticity can be used for prediction of linear
behavior of the masonry panel before crack. After crack formation, nonlinear finite element with
crack model can be used, however the analysis become very complicated. Based on experimental
observation of confined masonry response after extensive crack formation, diagonally braced frame
element has been proposed to model the confined masonry panel. The masonry panel is modeled as
bracing or strut element. Some variations are proposed on how to assign the properties of the strut
based on the actual dimension and material properties of the masonry. The analysis model for
masonry panel that we apply is based on Madan et al. (1997) and Mostafaei and Kabeyasawa (2004).
In this analysis the masonry is modeled as strut with nonlinear force-displacement curve as shown in
Figure 8. The actual dimension and material properties of the masonry is used to determine the
parameters of the force-displacement curves as presented in the following sections.
V

Vm

Vy

Vp
St

Up Um Uy
U
ru

Uy Um Up
tE
lem

Vp
en
t

Vy 
Vm

Figure 8 Masonry infill panel modeled as strut element.

3.3 Shear strength of infill masonry wall (Vm)

The shear strength of masonry wall (Vm) is the minimum strength based on various possible failure
modes of the masonry infill. Two failure modes are the most common, that is, diagonal compression
failures and sliding shear. Vm is determined as the minimum strength according to these failure
modes.
a. Shear strength based on diagonal compression failure
The compression strength of masonry wall can be calculated based empirical equation or base on
compression test of masonry units. In this analysis, empirical equation by Eurocode 6 as presented in
Eqn. 1 is used.

f ' m  K  f m0.65  f cm
0.25
(MPa) (1)
where,

f’m = compression strength of masonry wall

fm = compression strength of brick
fcm = compression strength of mortar
K = constant for masonry wall, taken as 0.6 MPa0.1

Provided that f cm  20 MPa and f cm  2f m

Shear strength is computed from horizontal component of the diagonal compression strut

Vc  z t f ' m cos (2)

where,

z is equivalent strut width estimated by the following equation given in FEMA 306 (1998):
1
 E m t sin 2  4
z  0.175 h  d m with   
0.4

 4 E c I g h m 
h = column height between centerlines of beams
hm = height of infill panel
Ec = expected modulus of elasticity of frame material
Em = expected modulus of elasticity of infill material = 750f ‘m
Ig = moment inertia of column
dm = diagonal length of infill panel
t = thickness of infill panel

b. Shear strength based on sliding shear failure

The maximum shear strength based on the Mohr-Coulomb criteria:

f  0    n (3)
where,
0  cohesivecapacityof themortarbeds
  sliding frictioncoefficient along thebed joint
n  verticalcompression stress in theinfill walls

Maximum horizontal shear force is as follows:

Vf   0 t l m   N (4)
where,
t = infill wall thickness
lm = length of infill panel
N = vertical load in infill walls.
In this study, N is estimated directly as a summation of applied external vertical load on the panel and
the vertical component of the diagonal compression force RC. The external vertical load is zero for the
infill walls of the building, and only the vertical component of the strut compression force is
considered. Therefore, maximum shear force can be calculated as:

R c cos   0 t l m   R c sin  or: (5)

0 t l m
Vf 
1   tan (6)
hm
use  0  0.04f ' m ;   0.654  0.000515 f cm and tan 
lm

c. Maximum shear strength

According to ACI 530-88, the maximum shear strength of confined masonry walls is

Vmax / t l m  8.3 kg / cm 2 (7)

The shear strength Vm used in the analysis is the minimum value from a, b and c above.

3.4 Maximum displacement and initial stiffness

The displacement at maximum load can be estimated by Eqn. 8 (Madan et all. 1997):

U m   'm d m (8)

where, ε`m is masonry compression strain at the maximum compression stress, ε`m = 0.0018. .

The initial stiffness K0 can be estimated by Eqn. 9 (Madan et al. 1997):

K 0   Vm / U m  (9)
Madan et. Al. propose  = 2.0; however this value does not agree with experimental results and need
to be calibrated.

The lateral yielding force and Vy and yielding displacement Uy can be computed by considering
Figure 8 as follows:

Vm  K 0 U m Vy
Vy  and U y  (10)
1  K0

The value α = 0.2 is proposed by Madan et. al. However this value does not always agree with
experimental results and need to be calibrated.

The Vp and Up should be determined considering that the line connecting the peak of the envelope and
the point (Vp, Up) pass through the 80% post peak point. The drift at 80% post peak is estimated at
1%. Assuming Vp = 0.3Vm lead to Up = 3.5(0.01hm-Um) (Mostafaei and Kabeyasawa 2004).
4 Analysis Model
To generate the analysis model, RC elements are modeled as nonlinear beam-column as presented in
Section 2.1. Masonry panels are modeled as strut element as presented in Section 2.2. Both infill
panel with and without opening are modeled by a strut element. The analysis model is shown in
Figure 9.

Figure 9 Analysis model of the wall panel

5 Analysis Method
The analysis is performed nonlinear pushover method. As in the cyclic load experimentation, the
loads are applied at the top side corners of the wall model.

The pushover analyses are selected due to its ability to compare the performance of the walls from the
experiments with the ones from numerical models. Moreover, the pushover analyses are able to
produce important parameters such as maximum capacity and maximum displacement. Given that the
cyclic load experiment is a quasi-static test with a low rate velocity, the pushover analyses will
produce results with satisfied accuracy. The load-displacement curves from pushover analyses are
evaluated and compared with the envelopes of the hysteretic curve recorded during cyclic load
experiment.

6 Analysis Results and Comparison with Experimental Results

Strength parameters for wall panel as proposed by Madan et. al (1997) gives analysis results
that are very close to experimental results; however stiffness parameters need to be adjusted.
After calibration with experimental results the initial stiffness parameter () and the post yield
stiffness parameter () that give satisfactory analysis results are presented in Table 2.
Table 2 Stiffness parameters of the analysis model.

Initial Stiffness Post Yield Stiffness

Model Opening Parameter Parameter
 
Madan This Madan This
et. al. Study et. al. Study
Model 1 No 2 6 0.2 0.05
Model 2 Yes 2 4 0.2 0.2
After calibration of the initial stiffness parameter () and the post yield stiffness parameter (a)
the comparison of experimental and analysis results are presented in Figure 10 and Figure 11
for Model 1 and Model 2, respectively.

60

50

40
Load (kN)

30
Experiment
20 Analysis

10

0
-25 0 25 50 75 100 125
Displacement (mm)

Figure 10 Pushover curve of Model 1

70
60
50
Load (kN)

40
30
Experiment
20
Analysis
10
0
-25 0 25 50 75 100 125
-10
-20
Displacement (mm)

Figure 11 Pushover curve of Model 2

The experimental results show that Window opening when protected by timber frame does not reduce
the strength of the wall panel significantly; however the stiffness properties are significantly altered.
The initial stiffness parameter is reduced by a factor of 1.5.

The analysis results show that the wall with opening can also be modeled by a simple strut. The
modification is required for stiffness parameters.
7 Concluding Remarks
The numerical analysis for confined masonry structure is conducted to calibrate structural parameters
of the analysis model. The structural model consists of reinforced concrete element and masonry
panel. Nonlinear beam-column element is used to model the RC members and strut element is used to
model the masonry wall. Nonlinear pushover analyses are performed as appropriate for the models.
The numerical models were developed using material properties obtained from the material testing.

Strength parameters for wall panel as proposed by Madan et. al (1997) gives analysis results that are
very close to experimental results; however stiffness parameters need to be adjusted. After calibration
with experimental results the initial stiffness parameter () and the post yield stiffness parameter ()
that give satisfactory analysis results are shown in the following table

Initial Stiffness Post Yield Stiffness

Model Opening Parameter Parameter
 
Madan This Madan This
et. al. Study et. al. Study
Model 1 No 2 6 0.2 0.05
Model 2 Yes 2 4 0.2 0.2

Window opening when protected by timber frame does not reduce the strength of the wall panel
significantly; however the stiffness properties are significantly altered. The initial stiffness parameter
is reduced by a factor of 1.5.

The analysis results show that the wall with opening can also be modeled by a simple strut. The
modification is required for stiffness parameters.

ACKNOWLEDGEMENT
Parts of this research are supported by Institut Teknologi Bandung (ITB) Program Riset dan Inovasi
KK (RIK). The supports are gratefully acknowledged.

REFERENCES
[1] European Committee of Standardization (CEN). (1996), Design of Masonry Structures. Part 1-1: General
Rules for Buildings-Reinforced and unreinforced Masonry. ENV 1996 1-1 Eurocode 6, UK
[2] FEMA 306 (1998), Evaluation of Earthquake Damage Concrete and Masonry Wall Buildings,
Basic Procedures Manual, Federal Emergency Management Agency, Washington, D.C.
[3] Mostafaei, H. and Kabeyasawa, T. (2004). Effect of Infill Masonry Walls on the Seismic
Response of Reinforced Concrete Buildings, Bull. Earthquake Research Institute, The
University of Tokyo, 79 ,133-156.
[4] Tomaževič, M. (1999). Earthquake-Resistant Design of Masonry Buildings, Imperial Collage
Press.
[5] Madan, A., Reinhorn, A.M., Mandar, J.B., Valles, R.E. (1997). Modeling of Masonry Infill
Panels for Structural Analysis, Journal of Structural Engineering ASCE. 123:10, 1295-1302.
[6] Muhammad, Dwipayana Ali (2011), Kajian Eksperimental Kinerja Dinding Bata Terkekang
Portal Beton Bertulang, Master Thesis, Institut Teknologi Bandung.
[7] Suarjana, M., Kusumastuti, D., Pribadi, K. S. and Rildova (2012) An Experimental Study on
the Effect of Opening on Confined Masonry Wall under Cyclic Lateral Loading, Proceeding
15th World Conference on Earthquake Engineering (15WCEE), Lisboa, Portugal,