Professional Documents
Culture Documents
EDITED BY:
Paola Mayorca, Sathiparan Navaratnaraj and Kimiro Meguro
July 2006
Preface
Masonry is the main currently used building material even though reinforced concrete and
steel have become more popular especially in industrialized countries. Thirty per cent of the
worlds population, or nearly 1,500,000,000 human beings, live in a home in unbaked earth.
Roughly 50% of the population of developing countries, the majority of rural population, and
at least 20% of urban and suburban populations live in earthen homes.
Unreinforced masonry structures are the most vulnerable type of buildings and have
caused more than 75% of the casualties during earthquakes in the past century. It is therefore
clear the urgency of retrofitting this existing housing stock to prevent casualties in future
seismic events. In order to effectively promote retrofitting, it is important to consider the
availability, acceptability and applicability of any proposed retrofitting scheme. In this
context, our group is proposing the use of polypropylene band (PP-band) meshes, which are
commonly used for packing, to retrofit unreinforced masonry houses. These bands are
inexpensive, strong, durable, light, and easy to handle.
The present report summarizes the developments of the methodology to retrofit masonry
houses by PP-band meshes as of October 2005. This is the outcome of a 3-year program
sponsored by the Ministry of Education, Culture, Sports, Science, and Technology of Japan
(MEXT). Although this research is still on-going, its basic framework has been already laid.
Some of the results presented in this report are still been investigated and therefore might be
updated in the future.
We sincerely hope that the technology presented in this report is widely used contributing
to effectively reduce the number of casualties in future earthquakes.
The editors
Table of contents
Preface ........................................................................................................................................ i
Table of contents .......................................................................................................................ii
List of figures ........................................................................................................................... vi
List of tables ............................................................................................................................. xi
List of photos..........................................................................................................................xiii
1. INTRODUCTION
1.1 Masonry........................................................................................................1-1
1.1.1 Masonry types .................................................................................1-2
1.2 Masonry seismic vulnerability .....................................................................1-4
1.3 Masonry failure modes during earthquakes .................................................1-7
1.4 Statement of the problem .............................................................................1-8
1.5 Objectives of the study .................................................................................1-9
1.6 Report outline .............................................................................................1-10
1.7 References ..................................................................................................1-11
ii
2.6.2 Walls..............................................................................................2-13
2.7 Summary ....................................................................................................2-15
2.8 References ..................................................................................................2-15
4. EXPERIMENTAL PROGRAM
4.1 Introduction ..................................................................................................4-1
4.2 Material Testing ...........................................................................................4-2
4.2.1 Brick ................................................................................................4-2
4.2.2 Mortar..............................................................................................4-6
4.2.3 Masonry...........................................................................................4-8
4.3 Shear wall tests...........................................................................................4-12
4.3.1 Specimen characteristics and construction process .......................4-12
4.3.2 Test setup.......................................................................................4-13
4.3.3 Testing program ............................................................................4-16
4.3.4 Test results.....................................................................................4-17
4.4 Summary ....................................................................................................4-21
4.5 References ..................................................................................................4-22
iii
5. NUMERICAL MODELING OF RETROFITTED MASONRY
5.1 Introduction ..................................................................................................5-1
5.2 Truss element ...............................................................................................5-1
5.3 Mesh reinforcement......................................................................................5-2
5.4 Model verification ........................................................................................5-3
5.4.1 Unreinforced masonry.....................................................................5-3
5.4.2 Reinforced masonry ........................................................................5-5
5.5 Parametric study ...........................................................................................5-7
5.5.1 Test setup effect ..............................................................................5-7
5.5.2 Unreinforced walls ..........................................................................5-8
5.5.3 Reinforced walls..............................................................................5-9
5.6 Summary ....................................................................................................5-11
5.7 References ..................................................................................................5-12
iv
7.2.2 Material and construction................................................................7-2
7.2.3 Instrumentation................................................................................7-2
7.2.4 Testing Program ..............................................................................7-3
7.3 Damage propagation and failure mechanism ...............................................7-4
7.4 Damage Level...............................................................................................7-5
7.5 Summary ......................................................................................................7-7
7.6 References ....................................................................................................7-7
v
List of figures
Figure 1-1 Distribution of earth masonry use and seismic hazard in the world [3] ..........1-12
Figure 1-2 Unbaked earth construction techniques [2] .....................................................1-13
Figure 1-3 Seismicity of the World [8] .............................................................................1-14
Figure 1-4 Breakdown of fatalities attributed to earthquake cause [4] .............................1-14
Figure 1-5 Relationship of the number of fatalities to the number of buildings
damaged in earthquakes [4].............................................................................1-15
Figure 1-6 Vulnerability functions for Masonry Type B Load bearing unit
block masonry according to the MSK Intensity Scale Definition
[4] ............................................................................................1-15
Figure 1-7 Research program task flow ............................................................................1-16
Figure 2-1 Parameter for the formulation of the Applied Element Method .........................2-17
Figure 2-2 Volume represented by a pair of normal and shear springs................................2-17
Figure 2-3 Brick-mortar spring representation.....................................................................2-18
Figure 2-4 Failure surfaces for the mortar-brick spring .......................................................2-18
Figure 2-5 Parameters used for the formulation of the constitutive law ..............................2-18
Figure 2-6 Tension cut-off yield value .................................................................................2-19
Figure 2-7 Coulomb friction yield value ..............................................................................2-19
Figure 2-8 Compression cap yield value ..............................................................................2-19
Figure 2-9 Comparison of tension model with experimental [8] .........................................2-20
Figure 2-10 Evaluation of standard deviation influence (GfI=0.012 Nmm/mm2) ............2-20
Figure 2-11 Evaluation of fracture energy influence (=0.3) ..............................................2-20
Figure 2-12 Comparison of shear model with experimental results [9] ...............................2-21
Figure 2-13 Evaluation of standard deviation influence.......................................................2-21
Figure 2-14 Evaluation of fracture energy influence (c=1Mpa, tan=0.73, =0.3).............2-21
Figure 2-15 Masonry shear wall considered for the analysis ...............................................2-22
Figure 2-16 Experimental cracking patterns [11].................................................................2-23
Figure 2-17 Structure discretization .....................................................................................2-23
Figure 2-18 Element size effect............................................................................................2-24
Figure 2-19 Number of springs effect ..................................................................................2-25
Figure 2-20 Force-displacement curves ...............................................................................2-26
Figure 2-21 Stress distribution Vertical normal stress (Case 1 in MPa) ...........................2-28
Figure 2-22 Stress distribution Shear stress (Case 1 in MPa) ...........................................2-30
Figure 2-23 Deformed shape (Case 1) (Scaling factor = 20)................................................2-31
Figure 2-24 Stress distribution Vertical normal stress (Case 2 in MPa) ...........................2-33
Figure 2-25 Stress distribution Shear stress (Case 2 in MPa) ...........................................2-35
vi
Figure 2-26 Deformed shape (Case 2) (Scaling factor = 20)................................................2-36
Figure 3-1 Force-deflection relations for unreinforced and retrofitted wall [3]................3-12
Figure 3-2 Load-deflection curves of coated and uncoated specimens [4] .......................3-12
Figure 3-3 Force-deflection relations for coated and uncoated specimens [5] .................3-12
Figure 3-4 Envelope curves of control and jacketed walls [6]..........................................3-13
Figure 3-5 Load-deformation curve at the middle of masonry panel surface [8] ...........3-13
Figure 3-6 Moment deflection curves [10] .......................................................................3-14
Figure 3-7 In-plane load versus shear strain [10]..............................................................3-14
Figure 3-8 Overall hysteretic response of the stabilized piers [11]...................................3-14
Figure 3-9 Hysteretic behavior of unreinforced and reinforced wall [12] ........................3-15
Figure 3-10 Typical example of application of proposed retrofitting .................................3-16
Figure 3-11 Force versus strain relation of 15.5mm PP-band.............................................3-17
Figure 4-1 Brick cores drilled from one brick unit (all dimensions in mm) .....................4-23
Figure 4-2 Axial stress versus axial and circumferential strain for brick cores ................4-23
Figure 4-3 Axial stress versus volumetric strain for brick cores.......................................4-23
Figure 4-4 Poisson ratio variation with axial stress for brick cores ..................................4-24
Figure 4-5 Young modulus variation with axial stress for brick cores .............................4-24
Figure 4-6 Compression loading direction........................................................................4-24
Figure 4-7 Mortar sand grain size distribution..................................................................4-25
Figure 4-8 Axial stress versus axial and circumferential strain for mortar cylinders .......4-25
Figure 4-9 Axial stress versus volumetric strain for mortar cylinders ..............................4-25
Figure 4-10 Poison ratio variation with axial stress for mortar cylinders ...........................4-26
Figure 4-11 Young modulus variation with axial stress for mortar cylinders.....................4-26
Figure 4-12 Compression test specimen nominal dimensions (all dimensions in mm) ..... 4-26
Figure 4-13 Compression stress versus deformation for masonry prisms (Set A)..............4-27
Figure 4-14 Compression stress versus Young modulus for masonry prisms (Set A)........4-27
Figure 4-15 Compression stress versus deformation for masonry prisms (Set B) ..............4-27
Figure 4-16 Compression stress versus Young modulus for masonry prisms (Set B) ........4-28
Figure 4-17 Shear test setup (all dimensions in mm)..........................................................4-28
Figure 4-18 Wall specimen (all dimensions in mm) ...........................................................4-29
Figure 4-19 PP-band mesh layout (all dimensions in mm) .................................................4-30
Figure 4-20 Wall testing setup (all dimensions in mm) ......................................................4-31
Figure 4-21 Full bridge used to measure the load in the steel rods.....................................4-32
Figure 4-22 Wall instrumentation .......................................................................................4-33
Figure 4-23 Sign convention (Positive direction is shown) ................................................4-34
Figure 4-24 Hole distribution Uniform (all dimensions in mm) ......................................4-34
Figure 4-25 Hole distribution Diagonal ...........................................................................4-35
Figure 4-26 Force deformation curves (VL=9kN)..............................................................4-35
vii
Figure 4-27 Force deformation curves (VL=30kN)............................................................4-36
Figure 4-28 Wall bottom uplift ...........................................................................................4-36
Figure 4-29 Wall shear deformation (VL=9kN) .................................................................4-37
Figure 4-30 Wall shear deformation (VL=30kN) ...............................................................4-37
Figure 4-31 Wall out-of-plane deformation (Bare wall VL=9kN)...................................4-38
Figure 4-32 Wall out-of-plane deformation (Bare wall w/mortar VL=9kN) ...................4-38
Figure 4-33 Wall behavior for different pre-compression loads (Bare wall w/mortar) ......4-39
Figure 4-34 Wall behavior for different pre-compression loads (Reinforced wall)............4-39
Figure 4-35 Force deformation relation detail (VL=9kN) ..................................................4-40
Figure 4-36 Normalized force-deformation relation (VL=9kN).........................................4-40
Figure 4-37 Normalized force-deformation relation (VL=30kN).......................................4-41
Figure 4-38 Horizontal force evolution (VL=9kN).............................................................4-41
Figure 4-39 Horizontal force evolution (VL=30kN)...........................................................4-42
viii
Figure 5-26 Force deformation curve for reinforced wall (Strong wall, pitch=45mm) ......5-36
Figure 5-27 Force deformation curve for reinforced wall (Strong wall, pitch=90mm) ......5-36
Figure 5-28 Force deformation curve for reinforced wall (Weak wall, pitch=45mm) .......5-36
Figure 5-29 Force deformation curve for reinforced wall (Weak wall, pitch=90mm) .......5-37
Figure 5-30 Deformed shapes of PP-band mesh reinforced walls (Strong wall) ................5-38
Figure 5-31 Deformed shapes of PP-band mesh reinforced walls (Weak wall) .................5-39
Figure 5-32 Comparison of vertical normal stress distribution Deformation =
5mm (in MPa) ............................................................................................5-40
Figure 5-33 Comparison of vertical normal stress distribution Deformation =
10mm (in MPa) ............................................................................................5-41
Figure 5-34 Comparison of shear stress distribution Deformation = 5mm (in MPa) ......5-42
Figure 5-35 Comparison of shear stress distribution Deformation = 10mm (in MPa) ....5-43
Figure 5-36 Residual to peak strength ratio variation for different conditions
of PP-band mesh reinforcement ......................................................................5-44
ix
Figure 6-18 Comparison between burned and unburned masonry wallettes
retrofitted by just overlapping of PP-band meshes..........................................6-28
Figure 6-19 Out-of-plane test setups...................................................................................6-29
Figure 6-20 Crack pattern types in out-of-plane test...........................................................6-29
Figure 6-21 Load versus deformation relation for burned brick out-of-plane
test ............................................................................................6-30
Figure 6-22 Load versus deformation relation for unburned brick out-of-plane
test ............................................................................................6-31
Figure 6-23 Out-of-plane load comparison between burned and unburned
brick specimen retrofitted by PP-band meshes................................................6-32
x
List of tables
Table 2-1 Material properties used in the analysis of the masonry walls.............................2-13
xi
Table 6-3 Mortar mix proportions in weight ..........................................................................6-3
Table 6-4 Result of the masonry direct shear tests .................................................................6-3
Table 6-5 Result of the masonry bond tests ...........................................................................6-4
Table 6-6 Results of the masonry compression tests (Burned brick) .....................................6-4
Table 6-7 Results of the masonry compression tests (Unburned brick) .................................6-4
Table 6-8 Mortar mix proportions in weight ..........................................................................6-5
Table 6-9 Number of specimens for diagonal tension test .....................................................6-5
Table 6-10 Displacement transducer specifications ...............................................................6-6
Table 6-11 Initial strength for burned brick specimens (Diagonal tension tests)...................6-7
Table 6-12 Initial strength for unburned brick specimens (Diagonal tension tests)...............6-8
Table 6-13 Number of specimen for out-of-plane test .........................................................6-10
Table 6-14 Mortar mix proportions in weight ......................................................................6-10
Table 6-15 Burned brick wallettes initial out-of plane strength ...........................................6-12
Table 6-16 Unburned brick wallettes initial out-of plane strength.......................................6-13
xii
List of photos
Photo 1-1 Traditional rural house in Kutch region of India (bhonga) [5] ............................1-17
Photo 1-2 Traditional bahareque house in El Salvador [6] ...............................................1-17
Photo 1-3 Rubble stone / fieldstone structure (by K. Meguro).............................................1-18
Photo 1-4 Adobe type of structure (by K. Meguro) .............................................................1-18
Photo 1-5 Simple stone structure [7] ....................................................................................1-19
Photo 1-.6 Massive stone structure (by P. Mayorca)............................................................1-19
Photo 1-7 Unreinforced brick structure (by P. Mayorca) .....................................................1-19
Photo 1-8 Unreinforced brick structure with RC roof (by K. Meguro)................................1-20
Photo 1-9 Confined masonry structure (by P. Mayorca)......................................................1-20
Photo 1-10 Completely collapsed masonry structure [9] .....................................................1-20
Photo 1-11 Out-plane failure of unreinforced masonry walls ..............................................1-21
Photo 1-12 Failure at the wall connection ............................................................................1-22
Photo 1-13 In-plane cracking of unreinforced masonry walls..............................................1-23
Photo 1-14 Cracking due to stress concentrations around openings ....................................1-24
xiii
Photo 4-7 Brick suction test ............................................................................................4-45
Photo 4-8 Mortar compression test setup .............................................................................4-45
Photo 4-9 Mortar specimen after compression test ..............................................................4-45
Photo 4-10 Mortar splitting tensile strength test setup .........................................................4-46
Photo 4-11 Mortar specimens after splitting tension test .....................................................4-46
Photo 4-12 Masonry compression test setup ........................................................................4-46
Photo 4-13 Typical failure in masonry compression test .....................................................4-47
Photo 4-14 Specimen failed in compression (Set A)............................................................4-47
Photo 4-15 Specimen failed in compression (Set B) ............................................................4-48
Photo 4-16 Bond test specimens...........................................................................................4-48
Photo 4-17 Bond test setup ............................................................................................4-49
Photo 4-18 Specimens after bond test ..................................................................................4-49
Photo 4-19 Masonry shear test setup....................................................................................4-49
Photo 4-20 Typical failure in masonry shear test .................................................................4-50
Photo 4-21 Specimens after masonry shear test ...................................................................4-50
Photo 4-22 Bottom connection detail ...................................................................................4-50
Photo 4-23 Template used for wall construction..................................................................4-51
Photo 4-24 Stopper detail ............................................................................................4-51
Photo 4-25 Detail of the loading surface treatment ..............................................................4-51
Photo 4-26 Test setup (Front view) .....................................................................................4-52
Photo 4-27 Test setup (Front view) .....................................................................................4-52
Photo 4-28 I-beam detail ............................................................................................4-53
Photo 4-29 Typical crack pattern of unreinforced wall (Bare wall) .....................................4-53
Photo 4-30 Typical crack pattern of reinforced wall (Reinforced wall w/diagonal holes)...4-54
Photo 4-31 Atypical crack pattern observed in Case No.4...................................................4-54
Photo 4-32 Detail of bottom connection in Case No.4.........................................................4-55
Photo 4-33 Crack pattern of bare wall w/mortar ..................................................................4-55
Photo 4-34 Failed connector detail.......................................................................................4-56
Photo 4-35 Loose connector detail .......................................................................................4-56
xiv
Photo 6-10 Failure patterns of brick masonry wallettes without and with retrofitting.........6-37
Photo 6-11 Epoxy failure at top edge ...................................................................................6-38
Photo 6-12 Brick crushing failure ........................................................................................6-38
Photo 6-13 PP-band meshes slip at wallette edges...............................................................6-39
Photo 6-14 Retrofitted masonry wallette for out-of-plane test.............................................6-39
Photo 6-15 Retrofitted masonry wallette for out-of-plane test.............................................6-40
Photo 6-16 Out-of-plane test setup.......................................................................................6-40
Photo 6-17 Failure patterns of brick masonry wallettes without and with retrofitting.........6-41
Photo 6-18 Specimen B-RE-2 failure...................................................................................6-41
xv
INTRODUCTION
1. Introduction
1.1 Masonry
Masonry is a composite structural material, which consists of units placed on top of each
other, laid dried or integrated with mortar. Stones, clay bricks, concrete blocks, among others
constitute the units while cement, lime or mud mortar is used to join them. Masonry is
anisotropic and heterogeneous. It is generally strong in compression, which is controlled by
the unit strength, and weak in tension, which is associated to the adherence between unit and
mortar.
Masonry and timber are the oldest building materials that are still used nowadays. In early
stages of masonry development, stones were used as units and no binding material was
employed. As tools became available and skills developed, stone units were shaped in regular
forms. The first bricks were made of mud or clay and dried by the sun. Later, it was found
that burning the bricks increased its strength and durability. Therefore, this practice became
common provided that there was combustible available.
Masonry was used as early as 9000-8000 BC near Lake Hullen, Israel [1]. The Walls of
Jericho, the Egyptian Pyramids at Giza, and the Sumerian Zigurats are masonry structures that
have endured until today. Both Greeks and Romans did also use masonry as a construction
material. Some examples of this are the Lion Gate at Mycenae, The Parthenon at Athens, The
Colosseum, and the Segovias aqueduct. Not only in Europe and Middle East was masonry
used but also in America and the Far East. The Inca citadel of Macchu Picchu, the Azteca
pyramids of Teotihuacan, the Indian Taj Mahal and the China Great Wall are just a few
examples of the widespread masonry use.
Masonry is still a main building material even though the use of other materials such as
concrete and steel has increased especially in industrialized countries. A large share of the
current world population lives in masonry houses. Thirty per cent of the worlds population,
or nearly 1,500,000,000 human beings, live in a home in unbaked earth. Roughly 50% of the
population of developing countries, the majority of rural population, and at least 20% of urban
and suburban populations live in earth homes [2]. Figure 1-1 shows the distribution of the
unbaked earth masonry use worldwide.
Masonry has a great variability around the world. The differences between the masonry
constituents (mainly the units), the construction techniques most of them traditional know-
1-1
INTRODUCTION
how, the structural configurations, etc. make masonry an extremely variable structural
material. For instance, two examples are presented. Photo 1-1 shows a Bhonga, commonly
used in Kutch region in India [5]. This structure consists of a single cylindrically shaped room
and a conical roof supported by the cylindrical walls. Many old Bhongas (constructed over
40-50 years) consist of adobe block walls with mud or lime mortar whereas the walls of
recently constructed Bhongas consists of cut stone or clay bricks in mud or lime mortar. This
type of house is quite durable and appropriate for prevalent desert conditions. On the other
hand Photo 1-2 shows a bahareque house [6], which is commonly used in Latin America
although its name varies from country to country. The foundation consists of either stones or
bricks and its main function is to transfer the loads to the ground and separate the walls from
the ground humidity. The main structure consists of wooden studs (bamboo is also used) and
cane spreaders attached with nails, wires, or vegetal fibers. The truss is filled with mud
composed of a mix of sand, clay and vegetal fibers. Although these two types of constructions
are quite different, both belong to the category of masonry structures.
3. Simple stone
Simple stone construction differs from fieldstone construction in that the building stones
have undergone some dressing prior to use. These hewn stones are arranged in the
construction of the building according to some techniques to improve the strength of the
structure (Photo 1-5).
1-2
INTRODUCTION
4. Massive stone
Buildings with very large stones are usually restricted to monumental constructions,
castles, large civic buildings, etc. These buildings usually possess great strength (Photo
1-6).
1-3
INTRODUCTION
Figure 1-3 shows the Seismicity of the World in the last century [8]. Three particularly active
zones can be observed: 1) the middle of the Pacific and Atlantic oceans, which are associated
with the mid-ocean ridges; 2) the western side of the Pacific Ocean from Alaska to New
Zealand, passing through the Aleutian Islands, Japan, Philippines and the islands of south-east
Asia and South Pacific; 3) the active mountain building zones at continental margins at the
western side of the Americas and at the trans-Asiatic zone running through the Himalayas, the
Caucasus Mountains to the Mediterranean and the Alps.
A comparison of the maps shown in Figure 1-1 and Figure 1-3 which show the distribution
of masonry use and the seismic activity in the world, respectively, leads to the conclusion that
there are wide regions in the world which are subjected to earthquakes and in which masonry
is one of the main construction materials. Therefore, it would be desired that masonry
buildings are earthquake resistant. Past and recent experiences have shown that this is not the
case.
About 75% of fatalities attributed to earthquakes are caused by the collapse of buildings.
Figure 1-4 shows the breakdown of earthquake fatalities by cause for each half of the past
century. It is clear that the greatest proportion of victims die due to the collapse of masonry
buildings. This percentage reduced from the 1st to the 2nd half of the century mainly because
of the increasing use of other building materials, such as reinforced concrete. Generally,
masonry structures fail even at low intensities of ground shaking and will collapse very
rapidly at high intensities.
Figure 1-5 shows the relationship between the number of people killed and the number of
buildings heavily damaged during 157 studied earthquakes. A building is considered heavily
damaged if its damage level, in the EMS scale (former MKS scale) is over D3. Table 1-1
shows the definition of damage levels for masonry houses according to this scale.
For each presented case, the predominant building type is shown. It can be observed that
the events with most of the building damage and hence casualties occurred in regions where
1-4
INTRODUCTION
masonry was the main construction material. It is also clear that the most destructive and
killer earthquakes occurred at locations where masonry structures were predominant.
Table 1-2 shows the differentiation of buildings into vulnerability classes as proposed in
[7]. In this table, vulnerability class A corresponds to the most vulnerable structure whereas
class F corresponds to the least vulnerable. It is clear that, compared with concrete and steel
constructions, masonry structures, especially unreinforced masonry, are extremely vulnerable
during earthquakes and have little chance of surviving a strong shake.
Simple stone O
Massive stone O
Unreinforced, with manufactured stone units O
Unreinforced, with RC floors O
Reinforced or confined O
Frame without earthquake-resistant design (ERD) O
REINFORCED
CONCRETE
1-5
INTRODUCTION
VIII, almost all masonry structures suffered some type of damage and at least 10% of the
constructions were completely collapsed.
1-6
INTRODUCTION
1-7
INTRODUCTION
1-8
INTRODUCTION
house performances during earthquakes in the INTERNET can have a large impact on the
population. Another far reaching approach is the in-situ demonstrations by which the general
public can see with the bare eye the different performance of non-retrofitted and retrofitted
structures under dynamic loads.
The technical and educational approaches are necessary for a successful improvement of
the existing housing stock quality. Any plan aimed at reducing the number of casualties due to
the collapse of masonry structures should take this into consideration.
1-9
INTRODUCTION
6. To numerically model the retrofitted masonry structure and capture its main features.
7. To use the numerical tool to identify the key issues for the optimization of the
proposed retrofitting technique
1-10
INTRODUCTION
Chapter 6 presents the second stage experimental program designed to further assess the
applicability of the proposed technique by discussing several relevant parameters such as
mesh orientation, installation procedure, etc. as well as the retrofitting effect for masonry of
different qualities. The program consisted of three parts: primary testing of PP-bands and
masonry, diagonal shear wallettes testing and out-of-plane wallette testing.
Chapter 7 and 8 are focused on the third stage experimental program which consisted of
small scale shaking table tests. In this program, four scale models were used to investigate
the seismic behavior of non-retrofitted and retrofitted models. The test results showed the
excellent performance of the PP-band retrofitted specimens.
Chapter 9 summarizes the findings of the first eight chapters, presents the final conclusions
of the present report, and proposes future research directions.
1.7 References
[1] Loureno, P.B., Computational strategies for masonry structures, Ph.D. Dissertation,
Delft University of Technology, Delft, The Netherlands, 1996.
[2] Houben, H. and Guillaud, H., Earth Construction A comprehensive guide, ITDG
Publishing, London, 1994.
[3] www.terracruda.com as retrieved on August 26th, 2003.
[4] Coburn, A. and Spence R., Earthquake Protection, John Wiley & Sons, West Sussex,
England, 1992.
[5] www.world-housing.net as retrieved on April 30th, 2003.
[6] Mayorca, P., Buildings and dwellings, in Konagai, K. (ed.) The January 13, 2001 Off
the Coast of El Salvador Earthquake, Japan Society of Civil Engineers, 2001.
[7] European Macroseismic Scale 1998 (EMS-98), Grnthal, G. (ed.), European
Seismological Commision, 1998.
[8] http://www.quakes.bgs.ac.uk/hazard/WSD_world_seismicity_map_medium.htm as
retrieved on May 31st, 2003.
[9] Meguro, K., Uehan, F., and Ramancharla, P. K., Damage to Masonry Structures, in
Sato, T. (ed.) A comprehensive survey of the 26 January 2001 earthquake (Mw7.7) in
the state of Gujarat, India, 2001.
[10] Meguro, K., Ohi, K., Mayorca, P., and Guzmn, R., Damage to buildings and dwellings,
in Konagai, K. (ed.) June 23, 2001 Atico Earthquake, Japan Society of Civil Engineers,
2001.
[11] Dolce, M., Masi, A., and Goretti, A., Damage to buildings due to 1997 Umbria-March
earthquake, in Bernardini, A. (ed.) Seismic Damage to Masonry Buildings, Balkema,
Rotterdam, 1999
1-11
INTRODUCTION
Figure 1-1 Distribution of earth masonry use and seismic hazard in the world [3]
1-12
INTRODUCTION
1-13
INTRODUCTION
1-14
INTRODUCTION
Figure 1-5 Relationship of the number of fatalities to the number of buildings damaged in
earthquakes [4]
Figure 1-6 Vulnerability functions for Masonry Type B Load bearing unit block masonry
according to the MSK Intensity Scale Definition [4]
1-15
INTRODUCTION
1-16
INTRODUCTION
Photo 1-1 Traditional rural house in Kutch region of India (bhonga) [5]
1-17
INTRODUCTION
1-18
INTRODUCTION
1-19
INTRODUCTION
1-20
INTRODUCTION
1-21
INTRODUCTION
1-22
INTRODUCTION
1-23
INTRODUCTION
1-24
NUMERICAL MODELLING OF MASONRY STRUCTURES
2. Numerical modeling of
masonry structures
2.1 Introduction
Compared to other construction materials, masonry has a great variability around the world.
Not only due to the different characteristics of its components, brick and mortar, but also due
to its different construction practices. The great variability of the material together with the
limited economic resources make it difficult to carry out large experimental studies for all
types of existing masonry.
In this context, there is a need for structural analysis techniques that can assess the
vulnerability of masonry structures as well as to design appropriate countermeasures. It has
being pointed out that the research in this field is still limited and most of the analyses have
focused on the pre-peak regime [1]. Therefore, a numerical tool by which masonry can be
analyzed in the pre- and post-peak regime given a limited number of parameters obtained
from simple experiments is needed.
In this chapter, a review of the existing approaches for masonry modeling is presented.
Next, the numerical method employed as well as the material constitutive law are introduced.
Finally, the numerical model is validated comparing its results with experimental data.
2-1
NUMERICAL MODELLING OF MASONRY STRUCTURES
cracking of the material ([4], [5], [6]). Using macro modeling, it is possible to analyze large
structures. However, discontinuities or details cannot be considered. In this study, simplified
micro modeling is adopted because it allows a degree of detail sufficient to discuss masonry
vulnerability and to design countermeasures.
where: A=[Ax, Ay]T and B=[Bx, By]T. Considering the theory of small deformations, i.e.
sin() and cos()1, Eq. 2-1 can be reduced to:
u n u 4 u 6 B y u1 + u 3 A y
= Eq. 2-2
u s u 5 + u 6 B x u 2 u 3 Ax
The potential strain energy of the ith spring pair in the local coordinate system is given by:
2-2
NUMERICAL MODELLING OF MASONRY STRUCTURES
T
1 u Kn 0 u n
= n Eq. 2-3
2 us 0 K s u s
where Kn and Ks are the normal and shear spring stiffness, respectively. The contribution of
the spring pair to the structure stiffness is calculated using the Castigliano Theorem.
2
K ij = Eq. 2-4
u i u j
The ith spring pair contribution to the stiffness matrix is given by:
Kn 0 K n Ay Kn 0 Kn By
0 Ks K s Ax 0 Ks K s Bx
K n A y K s Ax K n Ay2 + K s Ax2 K n Ay K s Ax K n A y B y K s Ax B x
Ki '= Eq. 2-5
Kn 0 K n Ay Kn 0 Kn By
0 Ks K s Ax 0 Ks K s Bx
K n B y K s Bx K n A y B y K s Ax B x K n By K s Bx K n B y + K s Bx
2 2
Finally, the stiffness matrix is rotated to obtain the stiffness matrix in the global coordinate
system, Ki=RKiRT:
K K 12
K i = 11 Eq. 2-6
K 21 K 22
where, R is the rotation matrix, K21=K12 and is the angle between the local and the global
coordinate systems. The complete structure stiffness matrix is assembled by adding up the
contributions of all the springs.
The remaining issue is how to evaluate the stiffness of a pair of springs. For this purpose,
it is assumed that each spring is representing a portion of the media (Figure 2-2) (specifically
its mechanical characteristics). In this fashion, the spring properties are formulated as:
2-3
NUMERICAL MODELLING OF MASONRY STRUCTURES
E d t G d t
Kn = and Ks = Eq. 2-10
a a
where t is the element thickness, E, Young modulus, G, Shear modulus, and the other
variables are described in Figure 2-2.
The equation of motion is solved incrementally:
where u&& and u& are the incremental accelerations and velocities, M is the mass matrix, C
is the damping matrix, and F is the incremental force. Because mass and damping are
assumed concentrated at the element centroids, both M and C are diagonal. To this end, a
mass proportional damping is considered.
1 a th th
= + Eq. 2-12
Kneq Eb t d E m t d
1 a th th
= + Eq. 2-13
Ks eq Gb t d G m t d
where Eb and Gb the Youngs and shear modulus of brick and similarly Em and Gm for the
mortar. Other variables are defined in Figure 2-3.
A major development of the Applied Element Method was the introduction of the
plasticity theory for the handling of the nonlinear material behavior. Both associative and
non-associative plastic potentials were included as needed for masonry modeling.
2-4
NUMERICAL MODELLING OF MASONRY STRUCTURES
material exhibits plastic behavior, f i ( , ) = 0 for at least one of the yield functions.
= e + p Eq. 2-14
= D e Eq. 2-15
g
p = Eq. 2-16
if only one yield function is active. In the special case of an associative flow rule, f and g are
identical. Both the flow rule and the yield function are functions of the stress and the
hardening parameter . In order to get simple numerical algorithms, both variables are usually
separated. The scalar hardening parameter is related to the plastic strain rate by:
= ( )
p
T
p Eq. 2-17
The plastic multiplier and the yield function must fulfill the Kuhn-Tucker conditions
during the whole loading process:
i 0, f i 0, i f i = 0 Eq. 2-18
In the case that there is more than one active yield function, the plastic strain increment in
the corner can be evaluated with (Koiter, 1953):
g 1 g
p = 1 + 2 2 Eq. 2-19
In Eq. 2-19, it was assumed that only two yield surfaces define the corner. This, however,
does not lack generality. The yield surfaces could eventually be coupled by introducing
composite hardening parameters,
1c = 1c ( 1 , 2 ) 2c = 2c ( 1 , 2 ) Eq. 2-20
2-5
NUMERICAL MODELLING OF MASONRY STRUCTURES
where the subscript c stands for composite and 1= 1(1p) and 2= 2(2p) are
calculated according to Eq. 2-17.
e = p = D ( p ) Eq. 2-21
At the next stage n+1, the problem is strain driven, i.e. is known, and the updating of
is straightforward. The remaining problem is the updating of p and . In the frame of an
implicit Euler backward algorithm, this problem is transformed into a constrained
optimization problem governed by the Kuhn-Tucker conditions given in Eq. 2-18. For single
surface plasticity, this leads to:
D 1 ( n +1 trial
n +1 ) + n +1 = 0
p
n +1 = n +1 ( n +1 , n +1 )
p
Eq. 2-22
f n +1 ( n +1 , n +1 ) = 0
n +1 = n + D n +1
where trial . Given that np+1 can be expressed as a function of n+1 according
to Eq. 2-16, the final problem is reduced to determining n+1, n+1, and n+1.
If the plastic potential has separate variables, it is possible to obtain the updated stress as a
function of the updated plastic multiplier
n +1 = n +1 ( n +1 ) Eq. 2-23
Combining Eq. 2-16 and Eq. 2-22.b, the hardening parameter rate can be expressed as:
n +1 = n +1 ( n +1 , n +1 ) Eq. 2-24
Inserting Eq. 2-23 and Eq. 2-24 in Eq. 2-22.c it is possible to obtain the yield function in
terms of the plastic multiplier rate, i.e. f n +1 ( n +1 ) = 0 .
In the case of multisurface plasticity, the Euler backward algorithm results in:
D 1 ( n +1 trial
n +1 ) + 1, n +1 + 2 , n +1 = 0
p p
1, n +1 = 1, n +1 ( n +1 , 1, n +1 , 2, n +1 )
c c p p
2, n +1 = 2, n +1 ( n +1 , 1, n +1 , 2, n +1 )
c c p p
Eq. 2-25
f 1, n +1 ( n +1 , 1c, n +1 ) = 0
f 2, n +1 ( n +1 , 2c, n +1 ) = 0
in which:
2-6
NUMERICAL MODELLING OF MASONRY STRUCTURES
g 1 g 2
1p, n +1 = 1, n +1 2p, n +1 = 2, n +1 Eq. 2-26
Similarly to Eq. 2-23 and Eq. 2-24, the following can be obtained:
And considering the same formula manipulation, the active yield functions can be expressed
as functions of the plastic multiplier rates, i.e.
f 1, n +1 ( 1, n +1 , 2, n +1 ) = 0
Eq. 2-30
f 2, n +1 ( 1, n +1 , 2, n +1 ) = 0
For the yield functions and hardening/softening laws proposed in the present study, the close
form solution of the plastic multiplier rates could be obtained.
2-7
NUMERICAL MODELLING OF MASONRY STRUCTURES
f 1 (, 1 ) = 1 Eq. 2-31
f 2G If
f t 1 tI 1 1
Gf ft
1 =
Eq. 2-32
2G If
0 1 >
ft
Figure 2-6 depicts the tension cut-off yield value. An associated flow rule is considered, i.e.
g1(,1)=f1(,1). The solution of the Kuhn-Tucker conditions leads to the following
expressions for the plastic multiplier rate:
2 f t G If f t 2 1 2G If trial 2G If
1
f t 2 2G If K n ft
1 = Eq. 2-33
trial 2G If
1 >
Kn ft
c 2G IIf
c1 II 2 2
Gf c
2 =
Eq. 2-35
2G IIf
0 2 >
c
g2 = c Eq. 2-36
Although it has being observed that the mortar interfaces exhibit dilatancy when sheared, this
effect will not be taken into account in the present model. The expressions for the plastic
multiplier rate are:
2-8
NUMERICAL MODELLING OF MASONRY STRUCTURES
II trial
trial
2cG II
c 2
2 + 2 tan G II
trial
2G
trial
f f f
2G IIf
2
c 2 2G IIf K s c
2 = Eq. 2-37
trial trial 2G IIf
2 >
K s trial c
f 3 (, 3 ) = m 3 ( 3 ) Eq. 2-38
m ( A( 3 c )2 + f m ) 3 r
3 ( 3 ) = Eq. 2-39
fr 3 >r
fi fm
3 c
c2
A= Eq. 2-40
f fm
r 3 >c
( r c )2
32 + C1 3 + C 2 = 0 Eq. 2-41
Ks Kn
C1 = + + 2( 3 c ) Eq. 2-42
mA A
trial
trial m trial mf r
trial Eq. 2-44
3 =
Ks + m Kn2
2-9
NUMERICAL MODELLING OF MASONRY STRUCTURES
G IIf f t
2 = 1 Eq. 2-45
G If c
whereas if f2 is active:
G If c
1 = 2 Eq. 2-46
G IIf f t
In case that both yield surfaces are active, the corresponding hardening parameter rates
would be:
2 2
G If c G IIf f
1 = 12 + II 2 2 = I t 1 + 22 Eq. 2-47
G ft G c
f f
The situation between the linearized compression cap and the shear yielding surface is
different because they represent two phenomena that are not related. The compression cap
stands for the interaction between mortar and brick, which results in the tensile splitting
failure of the latter caused by the difference in the poisons modulus of the two materials. On
the other hand, the shear yielding surface is representing the debonding between mortar and
brick. For this reason, no coupling between yielding surfaces f2 and f3 is considered.
The manipulation of Eq. 2-25 for the cases in which f1/f2 or f2/f3 yielding surfaces are
active leads to a quadratic equation on one of the relevant plastic multiplier rates. Since it is
always possible to express one of the plastic multiplier rates as a function of the other, the
system of equations can be solved without the need of any numerical method.
2-10
NUMERICAL MODELLING OF MASONRY STRUCTURES
fb f
+ t =1 Eq. 2-48
f 'b f 't
where fb and ft are the principal compression and tensile stresses, respectively, and fb and ft
are the uniaxial compression and tensile strengths, respectively.
2-11
NUMERICAL MODELLING OF MASONRY STRUCTURES
becomes wider, the rate at which springs enter the softening regime becomes smaller. The
strongest springs, which emulate the portions of the brick-mortar interface that have full
bonding, will continue loading while the weakest ones will be unloading. Because the fracture
energy is considered constant for all the springs, the softening rates will be different. The
combination of both effects results in the overall smooth softening.
In order to observe the effect of the fracture energy in the model, three cases with 0.009,
0.012, and 0.015Nmm/mm2 were studied. The standard deviation was kept constant and
equal to 0.3. The results are shown in Figure 2-11. It can be observed that the curvature is not
affected much. The first portion of the softening curve has different slopes because it is
directly related with the fracture energy value. However, the latter portion, which is related to
the scatter of the spring properties, is basically the same.
By adjusting the material properties, tensile strength and fracture energy, and the strength
standard deviation, the modeling of the element test in tension was performed and agreed well
with the experimental observation (Figure 2-9).
The second set of experiments consisted of two bricks connected with mortar and
subjected to direct shear under three different normal compression loads. The experimental
setup was carefully designed to subject the specimen to a pure shear state. The mortar
interface cohesion and friction coefficient were 1Mpa and 0.73, respectively. Figure 2-12
shows the envelopes of the experimental results. The first portion of the softening curve is
related to the decohesion of the mortar interface whereas the final branch is related to the
residual shear strength due to the friction between interfaces.
In this group of analysis, the previously used element and spring discretization was kept.
Again, it was assumed that the mortar cohesion was not constant but varied according to a
lognormal distribution. The load was applied under deformation control.
The effect of the standard deviation was studied first. Figure 2-13 shows the results of
three cases. As in the tension test, the post-peak curve becomes smoother as the standard
deviation increases. The peak strength slightly increases as well. The fracture energy
influence was also observed as shown in Figure 2-14. In this case again, the post-peak curves
present two portions, one related to the fracture energy itself and the other related to the
standard deviation. The former varies according to the energy variation whereas the latter is
almost constant as expected due to the constant standard deviation.
The test results presented in [9] were modeled by adjusting the material properties, i.e.
cohesion, friction, and fracture energy, and the cohesion standard deviation. Because the three
reported tests were carried out on similar specimens, it was assumed that the set of properties
was the same for the three cases and only the pre-compression load varied. The obtained
results were in good agreement with the experimental data as shown in Figure 2-12.
2-12
NUMERICAL MODELLING OF MASONRY STRUCTURES
2.6.2 Walls
In order to test the ability of the model to simulate the behavior of walls, the tests carried out
within the scope of the CUR project [11] were used. This testing program was extensively
instrumented and the uniformity of the material was given especial attention. Solid walls as
shown in Figure 2-15 were tested in shear. The specimen dimensions were
9901140100mm3. They were conformed by 18 layers of 4.5 bricks each. The brick
dimensions were 21052100mm3 and the mortar thickness was 10mm. The top and bottom
layers were embedded in steel beams, which were used to fix the wall to the floor and to
apply the loads on top.
At first, the wall was subjected to a vertical preload, p, while the upper horizontal beam
was kept horizontal. After application of the vertical load, the upper beam was fixed and the
racking test started. In order to keep the horizontality of the upper beam, the vertical load was
increased and so was the wall shear strength due to friction. This loading and boundary
conditions are hardly observed in the reality. However, for the purpose of verifying the
numerical technique, this dataset was chosen.
Two cases from the experimental program were simulated. In one case, the vertical pre-
compression was 30 kN while in the other it was 121 kN. The summary of the material
properties used for the simulation is shown in Table 2-1. The brick modulus of elasticity was
obtained from samples of each test while the mortar modulus of elasticity was fixed to fit the
initial stiffness of the observed force-displacement curve.
Table 2-1 Material properties used in the analysis of the masonry walls
P Eb Em ft c GfI GfII
Case tan
(kN) (kN/mm2) (kN/mm2) (N/mm2) (N/mm2) (Nmm/mm2)
1 30 170.0 4.0 0.25 0.35 0.018 0.125 0.75
2 121 170.0 10.3 0.16 0.22 0.018 0.050 0.75
The cracking patterns observed in the experiments are shown in Figure 2-16. In the testing
program, two walls under a pre-compression load of 30kN were tested.
In order to study the effect of the element size on the simulation results, two discretization
meshes were considered as shown in Figure 2-17. Furthermore, the analysis was performed
twice on the first discretization considering 10 and 5 springs per element side to observe the
effect of the number of springs.
At first, the mesh discretization effect was investigated. Figure 2-18 shows the results of
the analysis of Case 1 and 2 for two configurations. The analysis results did not change
considerably.
As observed in Figure 2-18, the analysis results are not affected by the element
discretization. However, the computational time is directed influenced. More elements imply
2-13
NUMERICAL MODELLING OF MASONRY STRUCTURES
more degrees of freedom, which increases the size of the stiffness matrix as well as its
bandwidth.
Figure 2-19 shows the results of the analysis for Cases 1 and 2 considering different
number of springs per element side. It can be observed that the number of springs does not
largely affect the results.
Although the number of degrees of freedom does not increase when the number of springs
per element side increases, the number of times that the elastoplasticity equations are solved
increases. This, however, is not as computationally burdensome as the increase of the stiffness
matrix size.
As a result of the previous discussion, it was decided to perform the analyses considering
(72 elements / brick) and 5 springs / element side. Figure 2-20 shows the force-deformation
curves obtained numerically and experimentally for Cases 1 and 2. Because the structure is
not perfectly symmetric, Case 1 was analyzed twice varying the direction of the horizontal
load (Case 1a). It can be observed that there was almost no difference in the structural
response although the diagonal cracking position slightly shifted.
In general, the simulation agrees very well with the first portion of the experimental curve.
However, the agreement decreases in the later portion. The reason for this is that at the current
stage, the linearized compression cap has not being included in the numerical model yet.
Because of the boundary conditions in the experiments in question, as the horizontal
deformation increases, a diagonal compression strut develops along the loaded diagonal.
Without the cap, the compression stresses in this region are not limited. As a consequence, the
shear strength from the friction mechanism is also unlimited.
Three mechanisms govern the masonry shear strength depending of the magnitude of the
shear displacements. At relatively low deformations, the shear friction mechanism is the
predominant. As the displacements increase, the diagonal cracking of the units controls the
behavior. At the last stage of large deformations, the last mechanism of masonry crushing in
compression is the critical. The two first stages have being well captured by the model in its
present situation.
Figure 2-21 and Figure 2-22 show the vertical normal and shear stress distribution at four
stages of the wall loading. At 1mm deformation, only cracking at the first and upper most
mortar layer has occurred and therefore stresses have being released in this zone. However,
the rest of the wall behaves fairly continuous. The compressed diagonal is clearly observed.
As the deformation increases, the diagonal tension crack occurs. As a result, the compression
strut divides in two with the diagonal crack in between them. Stresses along the crack are
released. The stresses in the two diagonal struts increase along with the horizontal
displacement. The shear strength, which at this stage relied only on the friction between
interfaces, increases accordingly.
2-14
NUMERICAL MODELLING OF MASONRY STRUCTURES
Figure 2-23 shows the structure deformed shapes at two loading stages. At 2mm
deformation, mainly two cracks at the bottom and upper most mortar layers are observed. As
the deformation increases, the diagonal crack appears. This sequence agrees well with the
observed in the experiments.
Figure 2-24 to Figure 2-26 show the stress distribution and deformed shapes for Case 2.
The mechanism in this case is similar to the previous one, although the stresses are higher in
this case. Because of this, the lack of the compression cap model is more evident. Regarding
the crack pattern, it is in well agreement with the experimental observation.
2.7 Summary
The present chapter presented a brief review of the existing methods for modeling masonry
structures. Micro modeling, simplified micro modeling and macro modeling combined with
continuous and discrete numerical methods are used to simulate masonry behavior.
Considering the scope of the present study, the simplified micro modeling and a discrete
analysis approach was selected.
The selected numerical method was the Applied Element Method (AEM), which was
adapted to consider the particular features of masonry structures. In the framework of
elastoplasticity, a simplified constitutive model for masonry was proposed and implemented
in the AEM.
In order to validate the numerical model, the simulations of element tests and shear wall
tests were carried out. The results were compared with the experimentally obtained data. In
general, a good agreement was found. The effect of the mesh refinement and the number of
springs per element side was also investigated.
Although in the model in its present state does not include the compression cap, it suffices
for the objectives of the present study.
2.8 References
[1] Loureno P. B., Computational strategies for masonry structures, Ph.D. Dissertation,
Delft University of Technology, Delft, The Netherlands, 1996.
[2] Gambarotta, L. and Lagomarsino, S. Damage models for the seismic response of brick
masonry shear walls. Part I: The mortar joint model and its applications, Earthquake
Engineering and Structural Dynamics, 26, 1997, p.423-439.
[3] Guinea, G.V., Hussein, G., Elices, M., and Planas, J., Micromechanical modeling of
brick-masonry fracture, Cement and Concrete Research, 30, 2000, p.731-737.
[4] Lee J. S., Pande G. N., Middleton, J., and Kralj, B. Numerical modeling of brick
masonry subject to lateral loadings, Computers and Structures, 61-4, 1996, p.735-745.
2-15
NUMERICAL MODELLING OF MASONRY STRUCTURES
[5] Anthoine, A., Derivation of the in-plane elastic characteristics of masonry throught
homogenization theory, Int. J. Solids Structures, 32-2, 1995, p.137-163.
[6] Lofti, H. R. and Shing, P. B., An appraisal of smeared creack models for masonry shear
wall analysis, Computers & Structures, 41-3, 1991, p.413-425.
[7] Tagel-Din H. and Meguro K.: Applied Element Method for simulation of nonlinear
materials: theory and application for RC structures, Struct. Engrg./Earthquake Engrg.,
17(2), 137s-148s, 2000.
[8] Pluijim, R. Van Der, Material properties of masonry and its components under tension
and shear, Proc. 6th Canadian Masonry Symposium, Canada, p.675-686, 1992.
[9] Pluijim, R. Van Der, Shear behavior of bed joints, Proc. 6th North American Masonry
Conf., Philadelphia, p.125-136, 1993.
[10] Ang, A. H.-S. and Tang, W. H., Probability concepts in engineering planning and
design, Vol. 1, John Wiley & Sons. 1975.
[11] Rajimakers, T. M. J. and Vermeltfoort, A. T., Deformation controlled meso shear tests
on masonry piers, Rep. B-92-1156, TNO BOUW/TU Eindhoven, Build. and Constr.
Res., Eindhoven, The Netherlands, 1992 (in Dutch).
2-16
NUMERICAL MODELLING OF MASONRY STRUCTURES
u5
Fsi,us
Kn B u4
Ks u6
Fni,un Ks
u2 A Kn
u3 u1
Figure 2-1 Parameter for the formulation of the Applied Element Method
a Ks Kn
Kn Ks
a
Volume represented by a spring pair
Y
2-17
NUMERICAL MODELLING OF MASONRY STRUCTURES
a
th
Mortar
Kb Km Kb
Brick Brick
CL Kneq CL
Coulomb friction, f 2
Cap model, f 3
Intermediate Y.F.
Initial Y.F. Tension, f 1
Residual Y.F.
Initial Y.F.
c
m
1
fi ft
Figure 2-5 Parameters used for the formulation of the constitutive law
2-18
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-19
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-20
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-21
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-22
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-23
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-24
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-25
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-26
NUMERICAL MODELLING OF MASONRY STRUCTURES
Figure 2-21 Stress distribution Vertical normal stress (Case 1 in MPa) (continues)
2-27
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-28
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-29
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-30
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-31
NUMERICAL MODELLING OF MASONRY STRUCTURES
(a) Deformation = 1 mm
(b) Deformation = 2 mm
Figure 2-24 Stress distribution Vertical normal stress (Case 2 in MPa) (continues)
2-32
NUMERICAL MODELLING OF MASONRY STRUCTURES
(c) Deformation = 3 mm
(d) Deformation = 4 mm
2-33
NUMERICAL MODELLING OF MASONRY STRUCTURES
(a) Deformation = 1 mm
(b) Deformation = 2 mm
2-34
NUMERICAL MODELLING OF MASONRY STRUCTURES
(c) Deformation = 3 mm
(d) Deformation = 4 mm
2-35
NUMERICAL MODELLING OF MASONRY STRUCTURES
2-36
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3.1 Introduction
As presented in the previous chapters, unreinforced masonry buildings are extremely
vulnerable during earthquakes. For this reason, several retrofitting methods have being
proposed to improve its strength, ductility, and energy dissipation capability. Some studies
have mainly focused on the restoration of monuments or historical buildings while others
have targeted masonry panel infills. These applications and the strengthening of masonry
houses in developing countries are different in the sense that in the latter case, the main issue
is the economy and availability of retrofitting materials.
In this chapter, a review of the existing methods for masonry strengthening is presented as
well as a proposal of a new efficient technique for masonry retrofitting in remote areas of
developing countries.
3-1
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
The third strategy consists of drilling vertical and/or horizontal holes into an unreinforced
wall and inserting either conventional reinforcing bars or pre-stressing tendons. Flexural and
shear strength as well as ductility can be improved through this methodology.
Finally, the last category includes the addition of steel braces or frames around the
unreinforced masonry piers in order to share the lateral forces and increase the lateral strength
and deformation capacity of the structural system.
In the following sections, a brief introduction of some of these strengthening techniques is
presented.
3-2
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
The test showed that both strength and stiffness of the damaged walls could be restored
with grout injections (Figure 3-1). Furthermore, the strength and ductility of the test walls
could be enhanced with the introduction of steel reinforcement. Grout injection proved to be a
reliable means for bonding the new reinforcement to the masonry.
3-3
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
wire) was placed and covered with 25-mm thick cement mortar. The meshes were anchored to
the wall by 50-mm long nails for wood driven by hand next to the wire intersection. Metal
bottle caps were left between the wall surface and the mesh to ease the placement of mortar
behind the mesh and to improve the mortar-masonry bond. Prior to placement of mortar, wall
surfaces were saturated. The mortar was placed manually using masonry trowels.
The test showed a more uniform inclined crack pattern and a remarkably higher strength in
all specimens rehabilitated by jacketing as compared to the control masonry specimens
(Figure 3-4.) The energy dissipated by the jacketed specimens was also higher. The
contribution of steel welded wire meshes to strength depended on the amount of horizontal
reinforcement, deformation applied, type of anchor and mortar quality.
Zegarra et al (1997) investigated the reinforcement of adobe houses with galvanized
welded wire steel meshes and cement mortar overlay. Two house modules with and without
strengthening were tested on a shaking table. The wire diameter was 1mm and the mesh pitch,
20mm. The mesh was attached to the house walls with 64mm long nails placed at 250mm
pitch. Additionally, holes were drilled through the walls at 500mm pitch and the meshes on
both sides of the wall were connected through wires. The wall holes were later filled with
cement mortar. In this method, the steel mesh is not placed on all the house wall surfaces but
only at the intersections between walls and at the walls with long unsupported length. After
the mesh is set, a 20mm thick cement mortar is laid over the walls.
The results of the shaking table tests showed that the reinforced house performed well and
could withstand the imposed excitation whereas the unreinforced specimen failed. This
technique was applied in some houses in Peru. During the last 2001 Atico Earthquake, one of
the houses that belonged to this program was located in the region subjected to strong shaking.
The house showed very little damage, which was within the limits of repairing. Most
important of all, it allowed its users to safely evacuate their homes during the event.
3-4
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
composite fibers (carbon) in the form of woven fabrics with a polymer-modified cement
matrix to form structural overlays for structural components. The surfaces of the relevant
structural members are cleaned in order to remove old paint and weak or weathered surfaces.
The first layer of polymer-modified cement the matrix is applied to the surface of the
member. Subsequently, a textile fabric of carbon,- the reinforcement- is pressed into the fresh
cement. If necessary the last two steps are repeated until the required stiffness and strength of
the overlay is reached. Finally, a covering sheet of polymer-modified cement is applied. In
this way a laminated composite is produced on the surface of the structural member.
In this experimental program 330.24m3 and 220.24m3 walls with a reinforcing
overlay of three layers of unidirectional carbon fabric and a polymer-modified mortar were
tested out-of-plane. Load perpendicular to the wall plane was applied with a pressure bag
while no vertical load other than the self-weight acted. The walls were loaded and unloaded
quasi-statically in increasing load steps. The wall was able to sustain a horizontal load of
120kN for a wall mass of 3,900kg. This capacity is much larger than the required by any
seismic code.
Another study on the use of FRP laminates for masonry strengthening was that of Ehsani
et al (1999). In this case, vertical glass FRP strips were installed on the wall surface. First the
wall was cleaned with a steel brush and dust and any loose particles were removed with high
air pressure. The surface of the wall where the fabrics were to be attached was coated with a
thin layer of a two component water-based primer. Second, composite fabric strips were cut
and laid on a plastic sheet, and the mixed epoxy was poured on the fabric and spread over the
whole fabric using a trowel, ensuring that the fabric was saturated with epoxy. Next, the
saturated composite strips were bonded to the wall surface by hand pressure and pressed with
a roller. Finally, a small layer of epoxy was put on the fabric for protection and
instrumentation purposes.
Three half-scale unreinforced masonry walls, 0.711.22m2, were constructed of solid clay
brick. The specimens were subjected to a prescribed out-of-plane load and displacement
history. The load was applied by means of an airbag system. The specimens were subjected to
approximately 20 cycles and the wall was subjected to two cycles for each maximum
displacement prescribed.
The tests showed that the ultimate flexural strength of the walls significantly increased; the
applied pressure varied from 10 to 32 times the unit weight of the wall per surface area.
Deflections as much as 2.5% of the wall height were observed for walls with unidirectional
fabric. Inelastic behavior was observed as a result of brickwork softening and delamination of
the GFRP strips.
FRP rods have also been successfully utilized for masonry strengthening. Tumialan et al
(2000) reported the use of near surface mounted FRP rods. The technique is similar to the one
3-5
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
proposed by Binda el al (1999) with the difference that instead of steel bars, FRP rods are
used. In this case, however, out-of-plane and in-plane strengths are enhanced. In the first case,
the FRP rods are placed vertically in slots grooved on the masonry wall. In the second one,
the rods are placed horizontally. Part of the mortar layer is cut out using a grinder.
To investigate the out-of-plane behavior, four masonry specimens were constructed with
different amounts of reinforcement. The wallettes (610122095mm3) were tested under
simply support conditions. The experimental results showed that the strength increased 4, 10,
and 14 times with respect to the original masonry capacity depending on the amount of
reinforcement (Figure 3-6).
The in-plane behavior was evaluated by testing 16251625150mm3 specimens in
diagonal compression. The strengthened walls showed a larger deformation capacity than the
control wall, which had a brittle failure. The shear capacity also increased in a range of 30 to
80% (Figure 3-7).
3-6
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
and =15.9mm A325 structural steel bolts were used to fasten the vertical and diagonal steel
strips to the walls. The steel strips were welded together at the center of the wall as well as to
300mm long 15015016mm steel angles anchored into the concrete footing and top beam
using 400mm long high-strength anchor bolts.
Six large-scale walls (18001800200mm3) were constructed and tested. Two vertical
servo-controlled actuators applied axial compression to the specimens and a third one was
positioned horizontally and supported by a frame to apply horizontal deformation reversals.
An identical axial load of 100kN was applied to all specimens. The experimental results
showed an increase in in-plane strength, ductility, and energy dissipation capacity (
Figure 3-9). The lateral load resistance was increased more than three times with respect to
the non-strengthened control wall.
3-7
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-8
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
Figure 3-10 shows an example of application of the proposed retrofitting procedure. The
figure shows a typical unreinforced masonry house. In the plan, the arrangement of PP-band
meshes as well as the distribution of connectors is presented. As mentioned in the previous
paragraphs, these are placed close to the intersection of two walls. In this case, the first
connector line is located at 75mm from the face of the contiguous wall. Similarly, at the wall
edges, connectors are placed at 150mm from the wall borders. Section A-A shows the detail
of mesh folding around the window opening and the anchoring of the mesh to the foundation.
To determine the modulus of elasticity and ultimate strain, 5 bands were tested in tension
as shown in Photo 3-14 and Error! Reference source not found.. The band was fixed on one
end and pulled from the other. A load cell was used to determine the applied load. Before
starting the test, two lines perpendicular to the band axis were drawn. The initial distance
between these lines was 200mm. The increment of this distance was measured with a ruler at
regular load increments. The results of the test are shown in Figure 3-11.
Three out of five bands failed before reaching 1.2KN, which is below the value given in
the catalogue. All of them, however, exhibited a large deformation capability, more than 10%
strain in all of the cases. The force-deformation curve is fairly bilinear with an initial and
residual stiffness of 16.5 and 8.9 KNm/m, respectively.
Given the relatively low stiffness of the PP-band, it is not expected that it will contribute to
increase the masonry wall strength. However, given its large deformation capacity, it might
improve the structure ductility. Therefore, it was decided to proceed with further testing to
verify the suitability of the procedure.
3-9
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3.5 Summary
In this chapter, a review of existing retrofitting methods for masonry structures was presented.
These were mainly categorized as: grout and epoxy injections, surface coatings, reinforced or
post-tensioned cores, and addition of structural elements. Most of the presented methods
focused on the restoration of monumental buildings or reinforcement of masonry infill panels.
For strengthening masonry structures in earthquake prone areas in developing countries
attention should be focused on providing methods that are economic, simple to execute, and
which make use of locally available materials. In this context, a novel retrofitting technique is
proposed based on the use of polypropylene bands, which are commonly used for packing.
Preliminary testing is performed to initially assess the suitability of the materials to be used.
Based on this preliminary evaluation, further experimental programs were designed and are
reported in subsequent chapters.
3.6 References
[1] Lizundia, B.; Holmes, W. T.; Longstreth, M.; Kren, A.; Abrams, D. P., Development of
Procedures to Enhance the Performance of Rehabilitated URM Buildings, NIST GCR
97-724-1, 1997, pp. 437.
[2] Binda, L., Modena C., Valluzi M.R., and Zago R., Mechanical Effects of Bed Joint
Steel Reinforcement in Historic Brick Masonry Structures, Structural Faults and
Repairs 99, 8th International Conference, London, UK, 1999.
[3] Manzouri, T., Shing, P.B., Schuller, M.P., and Atkinson R.H., Repair of Unreinforced
Masonry Structures with Grout Injection Techniques, Proceedings of the Seventh North
American Conference, University of Notre Dame, June 1996, pp. 472-483.
[4] Reinhorn, A.M. and Prawel S.P., Ferrocement for Seismic Retrofit of Structures,
Proceedings of the 2nd International Symposium on Ferrocement, Bangkok, Tailand,
January 1985, pp. 157-172.
[5] Prawel S.P., Reinhorn, A.M. and Qazi S.A., Upgrading the Seismic Resistance of
Unreinforced Brick Masonry Using Ferrocement Coatings, Proceedings of the 8th
International Brick/Block Masonry Conference, Dublin, Ireland, September 1988, pp.
785-791.
[6] Alcocer, S.M., Ruiz, J., Pineda, J.A., and Zepeda J.A., Retrofitting of Confined
Masonry Walls with Welded Wire Mesh, Proceedings of the 11th World Conference on
Earthquake Engineering, 1996, Paper No. 1471.
[7] Zegarra, L. San Bartolome, A., Quiun, D., and Villa Garcia, G., Reinforcement of
Existing Adobe Houses, AridLands Newsletter, No. 47, May 2000.
3-10
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
[8] Kolsch, H., Carbon Fiber Cement Matrix (CFCM) Overlay System for Masonry
Strengthening, Journal of Composites for Construction, Vol. 2, No. 2, May 1998.
[9] Ehsani, M.R. and Saadatmanesh H., Repair and Strengthening of Earthquake-Damaged
Concrete and Masonry Walls with Composite Fabrics, Proceedings of the 1st
International Conference on Composites in Infrastructure ICI96, Tucson, Arizona,
January 1996.
[10] Tumialan, J.G., Galati, N., Namboorimadathil, S.M., and Nanni, A., Strengthening of
Masonry with FRP Bars, ICCI 2002, San Francisco, CA, June 10-12
[11] Rai, D.C., Hysteretic Behavior of Unreinforced Masonry Piers Strengthened with Steel
Elements, 11th World Conference on Earthquake Engineering, 1996, Paper No. 501.
[12] Taghdi, M., Bruneau, M., and Saatcioglu, M., Seismic Retrofitting of Low-Rise
Masonry and Concrete Walls using Steel Strips, Journal of Structural Engineering, Vol.
126, No. 9, September 2000.
[13] http://sipan.inictel.gob.pe/ceresis/proyect/limitaciones/limitaciones.htm as retrieved on
June 5, 2003
[14] Industrial Materials Product Guide, Sekisui Jushi Corporation, 2002
(in Japanese).
3-11
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
Figure 3-3 Force-deflection relations for coated and uncoated specimens [5]
3-12
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
Figure 3-5 Load-deformation curve for middle of masonry panel - at surface [8]
3-13
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-14
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-15
Connector distance (250350mm)
A A Wooden lintel
Door
Top/Bottom fixing
PPband mesh
Detail in Photo 33
Connector
Footing
Overlap (300mm) 150mm 75mm
PLAN
3-16
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-17
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-18
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-19
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
Photo 3-7 Detail of the wall with PP-band on one side only
3-20
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-21
STRENGTHENING OF MASONRY STRUCTURES IN SEISMIC PRONE REGIONS
3-22
EXPERIMENTAL PROGRAM
4. Experimental program
4.1 Introduction
In the previous chapters the development of a numerical technique for the analysis of
unreinforced masonry and the proposal of an economic method for masonry retrofitting was
presented. In order to verify the suitability of the proposed retrofitting technique as well as to
create an additional dataset for the verification of the numerical tool, an experimental program
was designed and executed.
The experimental program had two parts: material testing and shear wall testing. The first
concentrated in the evaluation of the masonry parameters needed to feed the numerical model,
i.e. bond, cohesion, compression strengths. As discussed previously, the variability of
masonry properties is very high. Even if the same brick type and mortar mix are used,
differences in the construction procedure can lead to masonry assemblies with considerably
different properties. Because of this, the material testing program was also oriented to identify
the ranges in which the masonry properties varied for the same constituent materials.
Although a great care was put in keeping identical conditions for brick and mortar during the
construction, variability in the final product, i.e. masonry was observed.
The second part of the experimental program consisted of wall testing. The objective of
this group of tests was to assess the retrofitting efficiency. A complete evaluation would
include the examination of the retrofitting effect on each of the different failure modes
observed in masonry structures, such as in plane, out-of-plane, wall connection, etc. However,
at this stage, only the mesh effect on the in-plane behavior was examined. Because this was
the first time that this type of strengthening was used, the preparation of the specimens also
served as an opportunity to develop and optimize the mesh preparation and installation
process.
4-1
EXPERIMENTAL PROGRAM
4.2.1 Brick
Clay burnt wire-cut bricks were used for the present experimental program. The nominal unit
dimension was 21010060 mm3.
Compression strength
For brick compression tests, five cores (=50mm, h=100mm) were prepared. These cores
were drilled from brick units as shown in Figure 4-1. Among the five cores, three of them
were instrumented with strain gages in order to obtain the poisons ratio as well as the
modulus of elasticity. The loading plates acted directly on the sample. Table 4-2 shows a
summary of the test results. The modulus of elasticity was defined as the secant stiffness at
one third on the maximum observed stress.
4-2
EXPERIMENTAL PROGRAM
The results of the compression tests on SPC-1, 2, and 3 are shown in Figure 4-2 to Figure
4-5. The brick behavior is basically linear until failure, which is explosive. The modulus of
elasticity is almost constant through the whole loading history whereas the poisons ratio
increases from 0.12 to 0.23, especially in the last half of the loading history.
Another set of compression tests was carried out with half brick specimens as
recommended by the ASTM C-67 and JIS A 1108. Five samples were tested and the
compressive strength was measured. Thin wooden sheets were placed between the loading
plate and the specimen to avoid any concentration of stresses (Photo 4-3). Test results are
summarized in Table 4-3.
The compression strength obtained with the second set of specimens was approximately
70% of that obtained with the core specimens. This might be due to: a) specimen shape and
size and/or test setup; b) brick anisotropy. In general, core specimens with aspect ratio h/d=2
give smaller strengths than cubic specimens mainly because of the different stresses imposed
by the boundary conditions, i.e. the confining effect of the loading plates reduces as the aspect
ratio of the sample increases. However, in this case, core samples had larger strength than
parallelepiped samples suggesting that this might not be the cause of the discrepancy. In case
of the core samples, the loading surface was more uniform than in case of the half brick
4-3
EXPERIMENTAL PROGRAM
samples. This may have cause concentration of stresses and/or flexural compression stress
states in the latter that may have led to premature failure of the sample.
Other possible reason for the strength discrepancies is the anisotropy of the brick itself.
The loading axes of the core and half brick samples are not coincident as shown in Figure 4-6.
The brick used in the current experimental program was fabricated by extrusion along the axis
parallel to the brick smaller dimension. This may result in different properties along different
axes.
Although further testing is needed to clarify the above-described situation, this is beyond
the scope of the present study. The main purpose of the material testing is to obtain
parameters for the modeling. In this context, the mechanical properties in compression of both
mortar and brick separately are not as important as the behavior of the combination of both,
i.e the masonry.
The same type of cores used for the evaluation of the compression strength was used to
determine the brick splitting tensile strength. The test setup is shown in Photo 4-5. In this case,
only the peak load was recorded. The splitting tensile strength was calculated as:
2 Pmax
fbt = Eq. 4-1
Dt
where: Pmax, maximum registered load, D, core diameter, and t, core height. The test results
are shown in Table 4-4
Photo 4-2 shows the cores after the test. The failure plane is very well defined.
4-4
EXPERIMENTAL PROGRAM
The brick capacity to absorb water largely affects the masonry strength. If the brick absorbs
too much water from the mortar mix, no water will be left for the cement hydration. On the
other hand, the mechanism of bond between mortar and brick heavily relies on the brick
capacity to absorb some mortar water, which carries cementitious materials dissolved in it.
Therefore, a balance should be attained.
Brick suction is a property that measures how much water a brick can absorb through its
laying surface within a certain time. To measure this property, a brick surface is left in contact
with water for one minute while keeping the water surface constant and 2 or 3 mm above the
brick bottom face (Photo 4-7). The difference in the brick weight before and after the test is
called suction or initial rate of absorption (IRA). The standard exposed area is 194cm2 but
when testing bricks of different dimensions, suction is calculated proportionally to the area
relation. Several studies have tried to correlate suction and masonry bond strength. Although
there is still controversy, construction standards recommend IRAs between 10 and 20g. If
bricks have tendency to absorb more water, it is recommended to wet them before the laying.
Brick suction was measured for the bricks using in the present program. The results are
shown in Table 4-5.
Because of the high IRA found, it was decided to soak the bricks before construction and
let them dry for two hours before the laying. The IRA obtained after this procedure was
within the recommended ranges.
4-5
EXPERIMENTAL PROGRAM
4.2.2 Mortar
For the present testing program cement mortar was used. It was decided to use a cement to
sand proportion of 1:4.5 in volume. In order to keep the mortar mix uniform, the volume
proportion was converted to weight proportion as shown in Table 4-6.
Three 500g sand samples were tested to investigate the aggregate grading. The obtained grain
size distribution is shown in Figure 4-7 together with the limits recommended by the ASTM
C-144 for mortar sands. The average fineness modulus was 1.74.
As observed in Figure 4-7, the sand used to prepare the specimens fell outside the limits
recommended by the ASTM for certain grain sizes. However, according to other standards [1]
the sand could be used for this purpose. Because the objective of the test program was to
evaluate the effectiveness of the masonry retrofitting technique as well as to provide data for
the verification of the numerical simulation, it was not considered necessary to adjust the
grain distribution curve because all the specimens and test pieces for material testing were
built with the same material.
Compression strength
For evaluating the mortar compression strength, five cylinders (=50mm, h=100mm) were
cast and tested at 28 days. Cylinder and masonry wall curing conditions were similar. The
mortar average unit weight was 1.9tn/m3. The typical failure mode is observed in Photo 4-9.
The results of the tests are shown in Table 4-7 and Figure 4-8 to Figure 4-11.
4-6
EXPERIMENTAL PROGRAM
It can be observed that while the specimen stress is below 70% of the maximum stress, the
Poissons ratio remains almost constant. However, as the stress increases, the Poissons ratio
increases even beyond the theoretical value of 0.5. This fact has being reported by other
researchers ([4], [5]) and might be due to the presence of cracks in the test piece, which affect
the strain gage measurements. Another reason could be the effect of the glue used to paste the
strain gages to the mortar cylinders, which becomes more apparent for low strength materials
as is the case of masonry mortar.
The compression strength of the mortar is low when compared to the values found in the
literature for similar cement/sand proportions. However, several points should be considered
when assessing this test results. First, the mortar tested was the one used for building the
masonry walls. As a rule, this mortar has larger flow than the mortar used to evaluate the
standard compression strength according to the ASTM C-109 [1]. Because the w/c is larger,
the strength reduces. Second, the shape of the test pieces is cylindrical whereas the strengths
reported in the literature are evaluated using the ASTM standard, which recommends cubic
specimens. It is usual that the cylindrical specimen strength is approximately 80% of the
cubic specimen strength. Finally, the large fine content of the sand implied a larger amount of
sand for the same volume. The quantity of water needed to achieve certain mortar workability
is directly related to the amount of sand. More sand requires more water, thus implying a
reduced w/c ratio.
The tensile splitting test results are shown in Table 4-8. The strength was calculated using Eq.
4-1.
The splitting tensile strength was approximately 9% of the compression strength. Photo
4-10 and Photo 4-11 show the test setup and the specimens after the test.
4-7
EXPERIMENTAL PROGRAM
4.2.3 Masonry
All samples were cured under the same conditions in which the walls were cured, i.e. water
spray for 14 days after construction. Because the walls were originally scheduled to be tested
at the age of 42 days, the samples for evaluation of material properties were tested at the same
age. Due to logistical limitations, this does not stand for the second set of compression test
samples, which was test at 56 days.
Compression Strength
Five test pieces per set were prepared for evaluating the compression strength of the masonry
units. The specimen consisted of five brick units and four mortar joints as shown in Figure
4-12. Although the joints were intended to be 10mm thick, there was a large variation in the
joint thickness, which resulted in large variation of the specimen heights as shown in Table
4-9. In the table djmax and djmin stand for the thicker and thinner joint thickness, respectively.
The specimen upper and lower faces were leveled with cement paste. The test was carried
out under load control conditions and the deformation between the upper and lower most
brick centers was measured as shown in Photo 4.12.
As expected, the specimens failed at the bricks due to the splitting tension stresses induced
by the mortar. However, a remarkable difference was observed. The specimens of Set A
showed a poor bonding between mortar layers and clay bricks, which resulted in the
separation of both elements as observed in Photo 4.14. On the other hand, an intimate bonding
between brick and mortar was observed in the specimens belonging to Set B (Photo 4.15).
The stress-deformation and stress-modulus of elasticity relationships are shown in Figure
4-13 to Figure 4-16. The masonry compression strength as well as the modulus of elasticity
are summarized in Table 4-10 and Table 4-11. The modulus of elasticity was defined as the
secant stiffness at one third of the compression strength.
4-8
EXPERIMENTAL PROGRAM
It was observed that both the average strength and modulus of elasticity of Set A are
slightly less than 70% of the corresponding properties for Set B. This difference can be
attributed to several reasons: 1) mortar thickness variation; 2) brick conditions during the
laying; 3) workmanship. The first effect is observed in specimen MC-3a in which the average
joint thickness was 15.5mm. As a result, the modulus of elasticity was 63% of the average in
the same set. Although the compressive strength also diminished, this was not so dramatic.
The brick conditions and workmanship are not expected to largely change within each set.
Therefore, the difference between MC-3a and the rest of its group is purely the effect of the
joint thickness variation.
Although the brick preparation was kept as uniform as possible, i.e. constant soaking and
drying periods, it was observed that the brick surface drying was strongly affected by the air
temperature, humidity and brick exposed surface. Particularly, the surface of the bricks
belonging to Set A while preparing the test pieces was wetter than that of the bricks in Set B.
This may have undermined the bond between brick and mortar. Indeed, Photo 4.14 shows that
the poor bonding between these two elements led to the breaking of only one brick and its
separation from the rest of the specimen. As a consequence, there was no transfer of stresses
to the remaining bricks, which were almost undamaged. On the other hand, the specimen
4-9
EXPERIMENTAL PROGRAM
shown in Photo 4.15 (Set B) developed a good bonding between brick and mortar and
therefore three out of five bricks were broken as a result of the compression stresses imposed.
It is obvious that the energy required to break the later piece is greater than the required to
break the former.
The issues discussed in the two previous paragraphs are directly related to the human
factor or workmanship. An experienced mason will be able to keep a uniform mortar joint
thickness as well as to judge whether a brick is ready for laying or not. It has already been
recognized that the pressure applied to the brick while positioning it over the fresh mortar, the
time lapse between mortar spreading and brick placing, the realignment or tapping of the
brick once it has being placed over the fresh mortar, among other factors affect the strength of
the resulting masonry. Although only one person was in charge of the construction of the
masonry walls and the procedure was kept as uniform as possible, it is likely that the human
factor affected the product result due to the limited experience of the technician in charge.
Bond strength
The American standard, ASTM C-952, recommends two ways to evaluate the mortar-brick
interface bond strength. Neither of them was used for the present testing program. Instead,
five 404060mm3 brick pieces were placed over a brick overlaid with a 10mm thick mortar
layer as shown in Photo 4.16. Before performing the test, the mortar was cut along the borders
of the small brick piece to force the transferring of the pull out force only through the bond
between mortar and brick. The test was carried out under load control conditions and the
maximum strengths observed are summarized in Table 4-12. Two test pieces of Set A were
lost in the process of preparation. One piece broke during handling at early stage of curing
and the other while cutting the mortar around the small brick piece.
The results show a large scatter, which is characteristic of bond tests. Although it was
expected that the failure would occur on the upper mortar-brick interface, it was observed that
in some cases the situation was reversed. The bond strength corresponding to Set B was more
4-10
EXPERIMENTAL PROGRAM
than twice the strength of Set A. The reasons for this have being discussed in the previous
section.
Because the test setup has not being used for masonry before, there is little discussion on
the influence of the setup on the test results. Because the actual tested area is small (4040
mm3), the mortar could be easily spread in its entire surface and the actual contact area was
almost 100%. This is not true at the wall mortar-brick interfaces where in most cases, not
100% of the brick surface is in contact with mortar. Therefore, a penalty factor to consider
this effect may be introduced. Additionally, in the specimen used for the test the amount of
water available for cement hydration is different than the one available at the walls. This is
because: 1) a large area of mortar is exposed to the air and evaporation might be larger; 2) the
amount of water available in the smaller brick piece is smaller and can be easily lost because
the dimensions of the piece are small. On the other hand, during the wall construction, the
bricks are lightly hit with either the shovel or a rubber hammer and are subjected to the
weight of the structure over it. These two effects might increase the bond strength inside the
wall mortar-brick interfaces.
Shear strength
Five specimens were prepared to evaluate the shear strength of the masonry units used in the
present testing program. The test setup is shown in Photo 4.19 and Figure 4-17. In order to
obtain a force-deformation relationship for masonry in shear, force and deformations were
measured at a frequency of 100Hz. Unfortunately, this sampling frequency was not large
enough and only the peak load could be registered. These results are shown in Table 4-13.
In this group of material tests, a large variation of the cohesion between Set A and B was
observed as well as a high coefficient of variation especially for the second group.
Particularly, the strength of MS-1b was relatively high. The average and coefficient of
variation of the second group was recalculated without considering MS-1b. These results are
shown in parenthesis in Table 4-13.
4-11
EXPERIMENTAL PROGRAM
The reasons for the discrepancies between both sets are basically the same as the ones
discussed before. However, it should be added that in the particular case of Set A, one of the
two mortar joints was 15mm thick instead of 10mm. This was due to a construction
imperfection. Because the failure plane most of the times lies on the interface between brick
and mortar, this is not expected to largely influence the obtained results. However, if the crack
crosses the mortar joint to shift from one mortar brick interface to the other, as shown in
Photo 4.20, the mortar thickness may have some influence. Nevertheless, none of the samples
exhibit this pattern.
Each wall was constructed in one day. Attention was given to keep constant the
construction conditions, especially the mortar mix proportions and the brick suction. Mortar
was prepared with a mixer, in small quantities in order to limit the retempering. Before
4-12
EXPERIMENTAL PROGRAM
starting the wall construction, the brick IRA was determined to check that it was within the
recommended ranges.
In order to ensure the geometrical uniformity of the wall, a wooden template as shown in
Photo 4.23 was used. Because the time lapsed between the mortar placing and the brick
positioning has a strong influence in the masonry properties, this parameter was controlled
during the construction.
The walls were cured with water spray for 14 days in which they were covered with plastic
sheets. At the end of the curing process, the upper channel was installed.
In order to prevent the wall sliding, two stoppers at the wall toes were provided as shown
in Photo 4.24. Grouted was poured at the location were the actuators would apply the load to
the wall in order to smooth the surface and prevent stress concentration (Photo 4.25).
For the retrofitted walls, two meshes as shown in Figure 4-19 were prepared per wall. The
mesh pitch was 45mm and the inclination angles, 50. The mesh pitch was chosen so that each
brick would be crossed by at least three bands and as a result would not fall even in case it
broke. The mesh overlapping length was 250mm. The construction procedure was as
described in the previous chapter.
Because the bricks were very strong, the connectors were placed at the mortar interface. A
total of 27 connectors were used. A cement mortar mix (cement:sand::1:3) was used for the
protection overlay of 8mm thickness.
4-13
EXPERIMENTAL PROGRAM
The specimen lower plate and the loading frame were not perfectly even. Because masonry
is an extremely brittle material, trying to bolt both elements without any preparation would
cause the cracking of the lower most mortar joint. In order to prevent this, a thin layer of
gypsum was placed between these surfaces while setting the specimen for testing. After the
gypsum hardened, the specimen was bolted to the frame.
The three I-beams, which serve as guides for the vertical rods, were welded to the top
channel once the wall was set in the frame. The channel served not only as a means to install
the I-beams but also to distribute the reaction loads from the rods. These rods, in addition to
be used for applying the vertical pre-compression, gave stability to the system and prevented
the wall overturning.
In order to prevent the out-of-plane wall movement, a guide was provided on the wall side
opposite to the actuator.
Instrumentation
Loads were measured at the actuator and the steel rods. In the first case, the actuator built-in
load cell was used. For the steel rods, a full bridge, as shown in Figure 4-21 was installed at
every rod. Before the testing program started, the accuracy of the bridge measurements was
verified by loading the bars and comparing the bridge measured load to the one measured
with a load cell. In all the cases the agreement was satisfactory.
The displacements were measured with linear potentiometers (LVDT) and laser
displacement meters at the locations shown in Figure 4-22. The instrument specifications are
given in Table 4-15 and Table 4-16. LVDT D1 was type CDP-50 while the remaining were
type CPD-25.
4-14
EXPERIMENTAL PROGRAM
LVDT D1 and laser L4 aimed at obtaining the wall displacement at the actuator level.
LVDT D2 and D9 measured the relative horizontal displacement between the frame and the
specimen bottom. The first measured the 1st layer brick directly whereas the second measured
the stopper movement. LVDT D3 and D4 recorded the bottom corner uplift while D5 and D6
were intended at measuring the wall shear deformation. LVDT D7 and D8 registered the
relative vertical displacement between the frame and the channel-plate lower fitting. Lasers
L1 to L3 and L5 to L7 recorded the wall deformation profile. Figure 4-23 shows the
measurements sign convention.
In order to register temperature variations, two temperature gauges were installed at two
vertical steel rods, one on each side of the wall. During the present testing program, the
temperature variations were within the range of auto-compensation of the measuring
instruments.
Loading procedure
The first stage was the application of the vertical pre-compression load. For this, the bolts at
the bottom end of the vertical rods were slowly closed while monitoring the force increment.
During this process, attention was put in order not to generate any premature local stresses in
the specimen.
The second stage was the fixing of the horizontal steel bars together with the positioning
of the actuator. The horizontal bars were slightly tensed to guaranty the contact between
loading plates and the specimen. After fixing the actuator and related fittings, the forces on
the vertical rods were unbalanced and therefore, they were readjusted.
The final stage was the horizontal loading. Although, it would have being desirable to
investigate the cyclic behavior of the specimens, due to time limitations, only one loading
cycle was applied to all the walls. This stage consisted in 5 steps. In the first step, the wall
was loaded until shear cracking was observed. The second step consisted on pushing the wall
4-15
EXPERIMENTAL PROGRAM
10mm more in the same direction. In the third step, the actuator displacement direction was
reversed and the specimen was loaded until the diagonal shear crack in the opposite direction
appeared. In the fourth step, the wall was loaded 10mm more in the same direction. Finally,
the wall was unloaded.
4-16
EXPERIMENTAL PROGRAM
4-17
EXPERIMENTAL PROGRAM
Table 4-18 Peak load and corresponding deformations observed in the shear walls
Case No. Case name Peak load (kN) Deformation (mm)
1 Bare wall 170.0 5.69
2 Bare wall w/ holes 170.0 5.00
3 Bare wall w/mortar (VL=3kN) 309.3 9.69
4 Bare wall w/mortar (VL=90kN) 320.5 8.91
5 Reinforced wall (VL=3kN) 270.0 7.92
6 Reinforced wall w/ holes (VL=3kN) 193.7 5.17
7 Reinforced wall (VL=90kN) 290.0 7.85
8 Reinforced wall w/holes (VL=90kN) 280.0 9.12
Figure 4-28 shows the pre-peak displacement at the wall bottom for Cases No. 1 and 3. It
can be observed that the uplift is rather large. This implies that the wall rotation constitutes an
important portion of the deformation observed at the actuator level. The deformation jump
observed at a horizontal load of 30kN corresponds to the appearing of the bottom crack. If the
force deformation relations for these two cases are considered, it is clear that Case No. 1 has
an apparently lower stiffness. This is mainly due to the larger rotation, deduced from the
larger bottom uplift that the wall in Case No. 1 experienced.
In order to appreciate the real shear deformation that the walls experienced, diagonal
compression and expansion should be examined. Figure 4-29 shows the diagonal
deformations for Cases No. 1 to 3. This graph clearly shows that the modulus of rigidity is
fairly constant even for Case No. 1.
The maximum shear strain experienced by the walls was calculated using Eq. 4-2.
T + C D6 D5
= = Eq. 4-2
L 1260
4-18
EXPERIMENTAL PROGRAM
Although a relatively constant modulus of rigidity was observed in Group A, this was not
found in the Group. Figure 4-30 shows the diagonal deformations of this group of walls. It
can be observed that the reinforced wall with diagonal holes had a lower modulus of rigidity
than the two other cases. Thus, the larger deformation observed in this wall was more related
to its lower stiffness than to the wall rotation.
If for a moment, Case No.1 is set aside, the relation between the peak strengths of the
remaining walls in Group A can be reasonably explained. Case No.2 exhibits the lowest
strength because it did not have mortar overlay and in addition, the wall had uniformly drilled
holes. Case No.6, which had a similar distribution of holes as Case No.2, presented a larger
strength that could be attributed to the mortar overlay. Due to the low modulus of elasticity of
the PP-band mesh, it can be assumed that the mesh did not contribute to the peak strength.
Case No.5 was obviously stronger than Case No.6 because the wall did not have any holes.
This is also the case when Cases No.3 and 2 are compared. The former, in addition of not
having any holes, had the cement mortar overlay. This contributed to increase the wall
strength.
Cases No.5 and 3 corresponded to walls without holes and mortar overlay. However, there
is a remarkable difference in strength. This can be attributed to the better bonding between the
mortar overlay in the case of the unreinforced wall. During the construction process, because
of the small pitch of the PP-band mesh, setting the mortar overlay was difficult and air
pockets were observed. This poor bonding might have undermined the contribution of the
mortar to the wall strength.
In order to quantify the mortar contribution to the wall strength, the peak strengths of
Cases No.2, 3, 5 and 6 were compared. The difference between Case No.2 and 6 can be
attributed to the mortar overlay (th=8mm 2layers), with poor bonding, and is equal to
23.7kN. If this contribution is subtracted from Case No.5, 270-23.7=246.3kN, the resulting is
the strength of the wall without any mortar. The relation between the strength of Case No.3
and the previous result would give the effective contribution of the mortar assuming a perfect
bonding to the wall. In this case, the contribution would be 309/246.3=1.25, or a 25% strength
increase.
The previously obtained value can be roughly calculated if it is considered that the mortar
overlay contribution is a certain percentage of the mortar within the masonry wall. For
instance, the wall mortar thickness was 100mm whereas the mortar overlay thickness, in Case
No.3, was 10mm (th=5mm 2layers). This would imply an increment of 10%. If in addition
to this, the higher compression strength of the mortar overlay (fc=12.3MPa) is taken into
account, the increment would be 12.3 / 8.2 0.1 = 0.122 or 12.2% (Notice that the masonry
strength compression strength is 8.2). In this it is assumed that the mortar tensile strength is
4-19
EXPERIMENTAL PROGRAM
proportional to the square root of the compression strength. This preliminary result and the
value obtained in the experiments are in the same order. The mortar contribution to the
structure strength is lost after the first diagonal cracking appears.
The relatively high strength of the wall in Case No.3 can also be explained if it is
considered that in this case the mortar was particularly strong. If this would be the case, the
diagonal cracks, which mainly run along the mortar interfaces (Photo 4.29), would cross more
bricks on its way. This was actually the case as observed in Photo 4.33.
The peak strengths observed in the walls in Group B can be explained in a similar fashion.
The bare wall with mortar exhibit a slightly higher strength due to the better bonding of the
mortar overlay. The walls with uniform and diagonal holes exhibit basically the same strength.
This suggests that the diagonal holes had almost no effect on the wall peak strength.
Coming back to Case No.1, the observed peak load does not fit within the scheme
described in the previous paragraphs. This might be due to a premature failure of the wall.
This hypothesis is supported by the fact that this was the first tested wall and it the test setup
was adjusted. The main change was in the guide used to prevent the wall from having out-of-
plane displacements. In the first experiment, this guide was rather flexible and indeed some
deviation of the wall from the actuator axis was observed.
Figure 4-31 and Figure 4-32 show the vertical rod force evolution for Cases No.1 and 3. In
the graphs, the forces at the rods next to the actuator are depicted. In Case No.3, the forces on
both sides of the wall are almost identical throughout the loading process suggesting a rather
good balance. On the other hand, in Case No.1, a discrepancy between the forces on both rods
can be identified at an early loading stage. This fact would suggest that the wall in Case No.1
experienced out-of-plane deformations that led to its premature failing.
A comparison between the deformation curves of the unreinforced and reinforced walls
show the effect of the mesh reinforcement. The mesh presence contributed, at first, to smooth
the bottom flexural crack propagation. This can be observed in Figure 4-35 in which the
deformation jump observed when the flexural bottom crack appeared is more marked in the
case of unreinforced walls. In Case No.2, the jump is not so clear because the wall had a crack
at the bottom as a result of the specimen handling.
The effect described in the previous paragraph is not so evident in Group B because the
flexural crack propagation was smoother due to the larger pre-compression load. This is
further appreciated in Figure 4-33 and Figure 4-34 where two sets of unreinforced and
reinforced walls with different pre-compression loads are displayed. These two graphs also
show the relative small difference in peak strengths due to the initial pre-compression load.
Another effect of the mesh presence was the residual strength after the appearance of the
first diagonal crack. In the case of the reinforced walls, the residual strength was over 50% of
the peak strength except for Case No.7. This can be observed in the normalized graphs shown
4-20
EXPERIMENTAL PROGRAM
in Figure 4-36 and Figure 4-37. Furthermore, even at the largest deformation during the first
half of the loading cycle, the reinforced wall strength never fell below 45% of the peak
strength.
Case No.7 deserves special attention because effect described in the previous paragraph
was not observed. The specimen evaluation after the experiment showed that the mesh
connectors did not perform well. In the other reinforced walls, no connector broke and the
meshes were tightly attached to the masonry wall. This was not observed in Case No.7 in
which loose or broken connections affected the mesh performance (Photo 4.34, Photo 4.35).
In spite of this, this failed retrofitted wall gave the opportunity to observe the importance of
the wall connectors in the performance of the strengthened wall.
Another effect of the reinforcement was that the strengthened walls exhibit a smoother
behavior after the diagonal cracking whereas the unreinforced walls were unstable and
characterized by additional cracking and consequent sharp strength reduction. These features
can be observed in Figure 4-38 and Figure 4-39 in which the horizontal force evolution is
observed as a function of the loading step.
During the experiments, it was observed that the unreinforced walls failed in a brittle
manner whereas the reinforced walls, especially those of Group B, failed in a ductile way. It
was also observed that the small fragments of failed brick and mortar, which under seismic
out-of-plane inertial forces may fall and affect the people nearby the structure, were kept
within the PP-band mesh.
4.4 Summary
The present chapter introduced the experimental program, which was designed and executed
for two main purposes: 1) to verify the suitability of the retrofitting technique with PP-band
meshes, and 2) to create an additional dataset for the verification of the developed numerical
tool. The experimental program also served to investigate the variability of masonry
properties cause mainly by the construction process as well to develop and optimize the
retrofitting construction process.
The experimental program consisted of two parts: material testing and shear wall testing.
In the first, brick, mortar and masonry properties were determined, especially those that were
considered key input parameters for the numerical modeling. Two sets of masonry specimens
were prepared, one before and other after the shear wall construction. The material properties
were remarkable different. The groups constructed later exhibited almost twice the strength
observed in the group constructed earlier. This fact points out the importance of putting
attention not only to the constituent materials but also to the workmanship especially in the
case of masonry structures.
4-21
EXPERIMENTAL PROGRAM
The shear wall testing allowed the analysis and evaluation of the performance of the PP-
band mesh retrofit. Although it would have being desirable to subject the specimens to a
cyclic horizontal force, due to time limitations this was not possible. Only one loading cycle
was investigated. In spite of this, the experimental observations showed the goodness of the
propose retrofitting. The masonry residual strength was higher in the case of the reinforced
walls. Moreover, the failure mode was ductile compared to the brittle failure observed in the
unreinforced walls. In general, the behavior of the walls was more stable and larger
deformations could be sustained.
In spite of the high quality of the masonry used in the present testing program, the
beneficial effects of the PP-band mesh could be observed. Therefore, it is expected that when
this retrofitting procedure is applied to materials with properties closer to the ones observed in
developing countries, the goodness of the proposed procedure would be further highlighted.
The present shear wall testing program mainly aimed at investigating the effect of the PP-
band mesh on the in-plane behavior of masonry walls. Further evaluation of the technique for
the out-of-plane behavior and the enhancement of the adjacent wall connections are, thus,
necessary.
4.5 References
[1] Specifications for Masonry Structures, ACI 530.1-92 / ASCE 6-92. With ASTM
References, July, 1992.
[2] Cmara Peruana de la Construccin (CAPECO). 1997. Norma E.070 Albailera.
Lima, Per (in Spanish)
[3] Concrete compressive strength testing methods, JIS A1108
(in Japanese)
[4] McNary, W.S. and Abrams, D.P., Mechanics of Masonry in Compression,
Proceedings of the American Society of Civil Engineers, Journal of Structural
Engineering, Vol. 111, No. 4, 1985, pp. 857-870.
[5] Park. T. and Priestly, M.J.N., Seismic Design of Reinforced Concrete and Masonry
Buildings, John Wiley & Sons Inc., 1992, 744p.
4-22
EXPERIMENTAL PROGRAM
Figure 4-2 Axial and circumferential strain versus vertical stress for brick cores
Figure 4-3 Volumetric strain versus axial stress for brick cores
4-23
EXPERIMENTAL PROGRAM
4-24
EXPERIMENTAL PROGRAM
Figure 4-8 Axial and circumferential strain versus vertical stress for mortar cylinders
Figure 4-9 Volumetric strain versus axial stress for mortar cylinders
4-25
EXPERIMENTAL PROGRAM
4-26
EXPERIMENTAL PROGRAM
Figure 4-13 Compression stress versus deformation for masonry prisms (Set A)
Figure 4-14 Young modulus versus compression stress for masonry prisms (Set A)
Figure 4-15 Compression stress versus deformation for masonry prisms (Set B)
4-27
EXPERIMENTAL PROGRAM
Figure 4-16 Young modulus versus compression stress for masonry prisms (Set B)
4-28
A
985 See Top Connection detail
Grout Hole to pour grout Grout capping
4.5@210 + 4@10
150
EXPERIMENTAL PROGRAM
Channel (h=150)
Top connection
1072
1040
15@60 + 14@10
4-29
EXPERIMENTAL PROGRAM
985
38.5 16 @ 56.75=908 38.5
65
100
13 @ 70 = 910
1240
1040
100
65
4@56.75=227 100 492.5 492.5 100 4@56.75=227
1639
Connector location
Figure 4-19 PP-band mesh layout
4-30
EXPERIMENTAL PROGRAM
4-31
EXPERIMENTAL PROGRAM
A2 D2
e
D1 A1
D1
A1 D2
A2
Figure 4-21 Full bridge used to measure the load in the steel rods
4-32
EXPERIMENTAL PROGRAM
D1
L4
L5 L1
L6 L2
D3 D4
L7 L3
D7
D9 D2
D8
L4
L1
L5
D5 D6
L2 L6
D7 D8
L3 L7
D2 D9
4-33
EXPERIMENTAL PROGRAM
D1
HL
D6 D5
D9 D2
D3 D4
D8 D7
Figure 4-23 Sign convention (Positive direction is shown)
35 35
6@70 6@70 13@70
4-34
EXPERIMENTAL PROGRAM
4-35
EXPERIMENTAL PROGRAM
4-36
EXPERIMENTAL PROGRAM
4-37
EXPERIMENTAL PROGRAM
Figure 4-31 Evaluation of the wall out-of-plane deformation (Bare wall VL=9kN)
Figure 4-32 Evaluation of the wall out-of-plane deformation (Bare wall w/mortar
VL=9kN)
4-38
EXPERIMENTAL PROGRAM
Figure 4-33 Wall behavior for different pre-compression loads (Bare wall w/mortar)
Figure 4-34 Wall behavior for different pre-compression loads (Reinforced wall)
4-39
EXPERIMENTAL PROGRAM
4-40
EXPERIMENTAL PROGRAM
4-41
EXPERIMENTAL PROGRAM
4-42
EXPERIMENTAL PROGRAM
4-43
EXPERIMENTAL PROGRAM
4-44
EXPERIMENTAL PROGRAM
4-45
EXPERIMENTAL PROGRAM
4-46
EXPERIMENTAL PROGRAM
4-47
EXPERIMENTAL PROGRAM
Photo 4.16 Masonry bond test specimens Masonry bond test setup
4-48
EXPERIMENTAL PROGRAM
4-49
EXPERIMENTAL PROGRAM
4-50
EXPERIMENTAL PROGRAM
4-51
EXPERIMENTAL PROGRAM
4-52
EXPERIMENTAL PROGRAM
4-53
EXPERIMENTAL PROGRAM
Photo 4.30 Typical crack pattern of reinforced wall (Reinforced wall w/diagonal holes)
4-54
EXPERIMENTAL PROGRAM
4-55
EXPERIMENTAL PROGRAM
4-56
NUMERICAL MODELING OF RETROFITTED MASONRY
5. Numerical modeling of
retrofitted masonry
5.1 Introduction
The experimental results presented in the previous chapter showed the effect of reinforcing
masonry walls with PP-band meshes. The discussion presented there, however, is limited to
the conditions observed during the testing program, i.e. material characteristics and
experimental setup. Masonry characteristics widely vary around the world and the boundary
conditions observed during the tests do not reflect the real structure boundary conditions. In
this sense, the numerical approach is useful to discuss a wider variety of materials and also to
verify the influence of the test setup on the observed behavior.
This chapter presents the extension of the numerical model introduced in Chapter 2 for the
analysis of masonry structures reinforced with meshes. A new element, i.e. mesh, was
introduced in the formulation of the AEM. In order to validate the model, the dataset obtained
during the experimental program presented in Chapter 4 was used. Because the experiment
setup included steel bars to stabilize the system, truss elements were also added to the AEM.
This was verified independently from the mesh model, using the unreinforced masonry test
results.
After both mesh and truss models were validated, the effect of the test setup on the
observed results was investigated. Parametric studies of both unreinforced and reinforced
walls were also carried out to identify the parameters that influence most the structural
behavior.
5-1
NUMERICAL MODELING OF RETROFITTED MASONRY
the concrete spring closest and parallel to it. If the rebar orientation does not coincide with
that of any of the concrete springs, a new inclined spring element is created [2].
The steel bars used in the experiments behave differently. In this case, the masonry wall
and steel bars deform independently except for the points where they are bolted (Figure 5-1).
In order to represent this condition in the AEM, a truss element was introduced.
The truss element is included straight forwardly by adding the truss stiffness to the degrees
of freedom of the elements where the bar connections are located. Although the steel bars
have shear stiffness, this is small compared to their axial stiffness. Thus, only the latter was
considered in the formulation as shown below:
sin 2 sin 2
cos 0 cos 2
2
0
2 2
sin 2 sin 2
sin 2 0 sin 2 0
2 2
0
EA 0 0 0 0 0
K bar =
L cos 2
sin 2 sin 2 Eq. 5-1
0 cos 2 0
2 2
sin 2 sin 2 0
sin 2
sin 2 0
2 2
0 0 0 0 0 0
where E, is the bar Youngs modulus, A, the bar area, L, the bar length, and is the bar
inclination angle with respect to the global axes.
Given the conditions of the experimental setup introduced in Chapter 4, it is clear that the
vertical steel rods could only develop tension stresses and that almost no bending moment
could develop at the bar connections. Furthermore, the stresses in the bar did not exceed the
elastic range. Therefore, the steel material model used was the one shown in Figure 5-3 and
both bar ends were considered hinged.
5-2
NUMERICAL MODELING OF RETROFITTED MASONRY
spring as well. However, because these springs connect elements that were previously
unrelated, the stiffness matrix skyline needs to be modified accordingly.
The bands, when separated, have no shear stiffness. However, in the present application
they are combined to form a mesh. To take into consideration this effect, a shear stiffness
equivalent to 5% of the normal stiffness was considered for each band.
The material model used was elastoplastic bilinear as shown in Figure 5-6. It is needless to
say that no compression developed on the bands and thus this fact was considered in the
model. This material representation has the advantage that it can be directly adapted for other
materials such as steel.
5-3
NUMERICAL MODELING OF RETROFITTED MASONRY
The mortar modulus of elasticity was adjusted to fit the initial stiffness of the force
deformation curve. This is acceptable if it is considered that the mortar properties obtained
from the cylinders and those observed in the mortar joint vary due to differences in curing
conditions, brick confinement effect, etc.[3]
Because the friction between mortar and brick was not measured, a value was picked from
the literature [4]. Similarly, the fracture energies were calculated considering the experimental
data presented in [5] and [6]. These test results showed that the bond strength reduced to 90%
of the original value at a normal deformation of 0.15mm. In the case of the cohesion, a similar
decay was observed for a shear deformation of 0.20mm. The fracture energies were then
calculated considering similar strength decays for similar deformations.
Figure 5-8 shows the force deformation curves corresponding to two experiments and one
simulation. The model captures the main structural behavior features: bottom flexural
cracking, shear diagonal crack, and subsequent strength drop. After the appearance of the
flexural crack a drop in the simulated force deformation curve is observed. This effect was
also experimentally observed.
In the experiments, the diagonal shear occurred very fast and was accompanied by the
sliding of the top portion of the broken wall. A drastic reduction of strength was also observed.
In the static simulation the sudden crack appearance was captured by successive iterations
that followed the breaking onset. When the wall reached its peak strength, at first some
springs entered the plastic range and consequently residuals were needed to back-map the
spring stresses to the failure surface. The redistribution of these residuals caused other springs
to enter the plastic regime. This chain reaction generated the strength drop observed in Figure
5-8.
The static model, however, could not replicate the dynamic sliding of the broken wall.
Thus, there is a difference in the displacements of the experimental and simulation curves
immediately after the strength drop. It could be argued that during the experiments only the
data just before and after the shear crack was registered and therefore, the behavior between
this two points is unknown. Therefore, the simulated and observed curves should only be
compared at these two particular positions. Even in such a case, there is still some
discrepancy in the values of the residual strength.
Figure 5-9 shows the cracking sequence obtained in the simulation. At 5mm, only the
flexural bottom crack is observed and the rest of the wall rotates as a rigid body. Shortly after
5mm, the shear diagonal crack occurs and the top portion of the wall slides as observed in the
figures corresponding to a deformation of 10 and 15mm. After the shear crack occurs, the
bottom flexural crack closes as the top and bottom portions of the wall separate. All this
sequence is consistent with the behavior observed experimentally.
5-4
NUMERICAL MODELING OF RETROFITTED MASONRY
Although the stresses in the wall were not measured during the experiments, the stress
distribution and evolution obtained from the simulation is presented. At 5mm, when the wall
basically behaved as a rigid body, the distribution of normal and shear stresses is uniform.
After the shear crack occurrence the stresses are released and the load finds another load
transfer path. In the simulation, two paths are identified. The first, which carries most of the
load, goes from the actuator, through the brick just beneath it down to the base. The second
goes from the actuator through the upper wall portion down to the stopper on the wall bottom
right corner. The stresses on both paths slightly increase with the deformation.
Although the simulation results do not perfectly match the experimental quantitative
observations, qualitatively, the model can capture the main features of the observed behavior.
Regarding the mesh band parameters, the values obtained from the experiment were
considered as shown in Table 5-3.
5-5
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-13 shows the comparison between the experimental results and the simulation.
The model captured the main features of the behavior, i.e flexural cracking, shear cracking
and strength drop. In the experiments, a strength drop accompanied the flexural crack
appearance. This was smaller in the case of the reinforced walls than in the unreinforced cases.
Comparing the simulation results for both, unreinforced and reinforced walls, a similar trend
was observed. However, in both cases the drop was larger than the observed during the
experiments.
The observed and simulated stiffness also show some discrepancies, which are directly
related to the bottom crack propagation and opening. In the experiments, the mortar crack at
this location had a width of approximately 1 to 2 mm. If the mortar overlay effectively
connected the PP-band mesh and the wall, the free band length is the same as the mortar crack
width, i.e. 1 to 2 mm. The connectors considered in the model do not reflect this situation. In
the model, the free band length is approximately 80mm, resulting in a more flexible band. The
underestimation of the band stiffness at the location of the bottom crack may be the reason
why the PP-band effect on the bottom crack propagation and wall rotation was smaller than
the experimental one.
As mentioned in the previous section, there is no experimental data registering the shear
crack occurrence process. However, if the measured points just before and after the crack
appearance are compared with the numerical simulation curve, a good agreement is observed.
During the experiments, the wall strength slightly increased after the strength drop
associated with the shear crack. In the simulation this effect was observed. However, the
strength increase rate was larger in the later case.
Figure 5-14 shows the deformed shapes of the reinforced walls at 5, 10 and 15mm
deformation. The main difference between the unreinforced and reinforced wall deformed
shapes is that in the former case, the upper portion of the wall freely slid after the diagonal
crack. On the other hand, in the latter case, the bricks on the bottom corner of the upper wall
portion stuck to the base and rotated. This increased the friction resistance and the strength
increment rate observed after the peak. The deformed shapes also show that the shear crack,
which was concentrated in the case of the unreinforced wall, was better distributed in the case
of the reinforced wall.
Figure 5-15 and Figure 5-16 show the distribution of vertical normal and shear stresses in
the wall. As in the case of the unreinforced wall, the diagonal cracking causes stress release
along the crack and the redefinition of the load transfer paths. In the present case, however,
the stresses are equally distributed between the two observed paths due to the presence of the
PP-band mesh. Although the overall strength of the wall is not improved by the reinforcement,
the stress state inside the wall is considerably affected causing a more uniform stress
redistribution.
5-6
NUMERICAL MODELING OF RETROFITTED MASONRY
5-7
NUMERICAL MODELING OF RETROFITTED MASONRY
A total of 27 cases were considered. In order to discuss the obtained results, the force
deformation curve was simplified as shown in Figure 5-18. The discussion mainly focuses on
the peak strength, Ppeak, residual to peak strength ratio, Ppeak/Pres, and the residual stiffness,
Kres=(Pfinal Pres)/(Dfinal Dcr).
Figure 5-19 shows three cases in which the only varying parameter is the bond strength. It
is clear that this has almost no effect on the peak strength and its effect is minimum in the
residual strength. The bond mainly influences the onset of the bottom flexural crack
appearance and the subsequent strength drop.
Figure 5-20 shows three cases in which the cohesion strength varies. As expected, the peak
strength is higher for the larger cohesion case. The residual strength is also larger. However,
because the strength drop is larger in the case of the strongest wall, the Ppeak/Pres ratio is not
affected considerable. The residual stiffness is also very similar for the three presented cases.
Figure 5-21 shows the effect of the friction variation on the observed behavior. It is clear
that the peak strength increases as the friction increases. The strength drop that follows the
shear crack occurrence is almost constant and therefore a larger Ppeak/Pres ratio for larger
friction coefficients is observed. The residual stiffness is also affected. A larger friction
coefficient results in a larger residual stiffness.
The results of the parametric study are summarized in Figure 5-22 to Figure 5-24. As
mentioned earlier, both cohesion and friction affect the wall peak strength. However, only the
latter improves the residual to peak strength ratio and the residual stiffness. This suggests that
in order to effectively improve the masonry wall behavior in both the pre- and post-peak
regimes, the interface friction in the interface is a key parameter to be paid attention.
Friction is a physical phenomena rather than chemical as is the case of cohesion. In order
to increase the interface friction, the sand used for the mortar is most important. Sands with a
large fineness modulus and rough shape enhance the mortar interface friction. Furthermore,
5-8
NUMERICAL MODELING OF RETROFITTED MASONRY
strong sands, which can resist larger compression stresses without crushing, do not loose its
frictional properties even under large compression stresses.
The mesh pitch is a parameter that affects the behavior of the reinforced walls. A larger
reinforcing material ratio may lead to a better strengthening performance. This is in direct
relation to the connector amount and location because they ensure that the reinforcement
material is fully used. More connectors reduce the band free lengths, increasing the
reinforcement stiffness. Figure 5-25 shows the combinations of mesh pitches and connector
locations considered in the parametric study.
The relation between the wall and reinforcement strengths and stiffness may also affect the
reinforced wall behavior. Considering that the structures that are the target of the present
study are several times weaker than the walls used in the experimental program, it is
reasonable to investigate the variation of the failure mechanism, and the structure stiffness
and strength when the reinforcement is applied to such weak materials.
The properties of weak masonry, which are available in the literature [7], indicate that their
strength can be roughly considered equal to 10% of the strength of the walls used in the
experiments. For the present parametric study, the weak wall properties were defined as to
10% of the corresponding values obtained in the material tests presented in Chapter 4. It
would be unrealistic to reduce the friction value by a factor of 10. Therefore, this parameter
was reduced by only 25%. Totally, eight cases were considered with the conditions presented
in Table 5-7.
5-9
NUMERICAL MODELING OF RETROFITTED MASONRY
Table 5-7 Conditions that were considered in the parametric study of reinforced walls
Connection Mesh pitch Wall type
Case
Full Partial 45 90 Strong Weak
1
2
3
4
5
6
7
8
Figure 5-26 and Figure 5-27 show the force deformation relations for the strong wall cases.
It can be seen that the strength drop after initial cracking is larger in case of partial mesh
connection and that this effect is more dramatic for the mesh with 90mm-pitch.
The mesh pitch and connection condition also influences the residual strength. Larger
pitches and fewer connectors reduced the reinforcement beneficial effects. A large pitch mesh
adequately connected can perform better than a small pitch mesh with fewer connectors.
Figure 5-28 and Figure 5-29 show the force deformation curves for the reinforced weak
walls. The structure failure mechanism was not changed but the residual strength after the
shear cracking was larger than in the unreinforced case. Because of the low stiffness of the
PP-band mesh, it has little effect before the wall cracks.
Although a full connection of the mesh improves its performance, this effect is smaller in
the case of the weak walls than in the strong walls. The connection effect eventually
disappears as the mesh pitch increases.
Figure 5-30 shows the deformed shape for the reinforced walls (strong walls). In the case
of the walls reinforced with PP-bands, the main effect observed is the better distribution of the
wall cracks and in the case of the fully connected mesh at 45mm pitch, a certain restrain to the
upper portion of the wall sliding.
Figure 5-31 show the deformed shapes for the weak walls reinforced with PP-band. It is
again observed that the fully connected meshes prevent the sliding of the upper wall portion.
This effect is clear for the 45mm pitch mesh and to some extend for the 90mm pitch mesh.
Figure 5-32 to Figure 5-35 show the stress distribution of two weak walls, one
unreinforced and the other reinforced with PP-band mesh (Case 5 in Table 5-7.) As
mentioned in the discussion on Section 5.4.2, before the shear crack occurrence the stress
distribution is quite uniform. After the cracking, however, the stresses are released along the
crack and transferred through alternative loading paths. The stresses are better distributed in
the reinforced wall than in the unreinforced one.
5-10
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-36 shows the peak to residual strength ratio for the reinforced walls. The case
numbering is consistent with the one shown in Table 5-7, i.e. the four cases to the left
correspond to strong walls whereas the remaining four correspond to weak walls. The
Pres/Ppeak for unreinforced walls is 0.20 and 0.25 for strong and weak walls, respectively. In all
the cases, this ratio is increased by the presence of PP-bands, especially in the case of the
weak walls.
As discussed earlier, the connection condition affects the wall residual strength. In general,
a fully connected mesh shows higher Pres/Ppeak than a partially connected one for the same
mesh pitch. This difference reduces as the pitch increases. The connection effect is larger for
the strong walls. This causes that a fully restrained mesh with 90mm pitch exhibits larger
residual strength than a partially restrained mesh with half the pitch.
The fact that the Pres/Ppeak ratio is comparable for both strong and weak walls suggests that
the PP-band contribution is related to the strength of the structure that it reinforces. In
absolute terms the strong wall post-peak strength is increased more than the weak wall
strength. The reinforcement effect can be separated in two components: the reinforcement
strength itself and its effect on the masonry behavior. The former depends on the
reinforcement only whereas the latter mainly depends on the masonry properties. In the case
of the PP-band mesh the first component is relatively small (even when a weak wall is
considered) and most of the effect corresponds to the second component. Thus, when a wall is
reinforced with a PP-band mesh, the improvement it gets is directly related to the wall
properties.
5.6 Summary
The present chapter presents the extension of the numerical modeling introduced in Chapter 2
for the analysis of retrofitted masonry structures. In order to verify the model accuracy, the
experimental dataset presented in Chapter 4 was utilized. Due to the experiment setup
characteristics, a truss element was introduced in the Applied Element Method to model the
external steel bars that were used. The mesh model was also introduced to account for the
retrofitting
After presenting the model formulation the unreinforced and reinforced masonry wall
experiments were simulated. A good agreement was found between the simulation results and
the experimental observations. Force displacement curves and crack pattern evolution were
compared. Some discrepancies were observed and their reason explained.
The effect of the experimental setup on the observed behavior was investigated. It was
found that the external steel bar diameter had a direct influence in the observed stiffness and
5-11
NUMERICAL MODELING OF RETROFITTED MASONRY
strength. However, after the shear crack occurs, this effect gradually reduces as the
deformation increases.
A parametric study of the unreinforced walls was carried out. It was found that the
parameter that influenced the most the pre- and post-peak behavior is the friction coefficient.
A larger value of this parameter results in a larger peak strength, residual to peak strength
ratio, and residual stiffness.
A parametric study of the reinforced walls was also carried out. Conditions such as the
connection patterns, mesh pitch and masonry strength were varied. The analysis showed that
meshes with larger pitches but good connection with the masonry wall can perform as well as
meshes with smaller pitches but poor connections. The mesh reinforcement effect was
roughly separated in two components: the first directly related to the reinforcement strength
itself whereas the second, to the masonry properties. It was found that for PP-band meshes the
later is larger and therefore the mesh contribution is directly related to the strength of the
masonry wall that is strengthened.
5.7 References
[1] Tagel-Din H. and Meguro K.: Applied Element Method for simulation of nonlinear
materials: theory and application for RC structures, Struct. Engrg./Earthquake Engrg.,
17(2), 137s-148s, 2000.
[2] Nagashima, H.: RC
, (), 2003.
[3] Amrhein, J.E., Reinforced masonry engineering handbook 5th Ed., Masonry Institute of
America, 1998.
[4] Atkinson, R.,H., Noland, J.L. and Abrams, D.P., A deformation failure theory for stack-
bond brick masonry prisms in compression, Melbourne, Australia, Vol. 1, pp.577-592,
1985.
[5] Pluijim, R. Van Der, Material properties of masonry and its components under tension
and shear, Proc. 6th Canadian Masonry Symposium, Canada, p.675-686, 1992.
[6] Pluijim, R. Van Der, Shear behavior of bed joints, Proc. 6th North American Masonry
Conf., Philadelphia, p.125-136, 1993.
[7] Moreira Peate, R. A. and Roales Inestroza, J. A., Diseo, construccin y control de
calidad de estructuras de adobe para vivienda rural, Thesis submitted to the
Centroamerican University Jos Simen Caas, in Spanish, 1998.
5-12
NUMERICAL MODELING OF RETROFITTED MASONRY
u5 Element j
u4
u6
Bar axis
u2
u1
u3
Element i
5-13
NUMERICAL MODELING OF RETROFITTED MASONRY
Panel boundary
Reinforcement mesh Band 1
Band 2 Mesh spring
for band 2
l2
Connector
l3
l1
Element boundaries
Connector
5-14
NUMERICAL MODELING OF RETROFITTED MASONRY
Ks
Kn u5
B
u4
u6
u2
u3
Fni,un u1
Fsi,us Kn
A
Y Ks
X
5-15
NUMERICAL MODELING OF RETROFITTED MASONRY
5-16
NUMERICAL MODELING OF RETROFITTED MASONRY
5-17
NUMERICAL MODELING OF RETROFITTED MASONRY
5-18
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-10 Stress distribution - Vertical normal stresses (in MPa) (continues)
5-19
NUMERICAL MODELING OF RETROFITTED MASONRY
5-20
NUMERICAL MODELING OF RETROFITTED MASONRY
5-21
NUMERICAL MODELING OF RETROFITTED MASONRY
5-22
NUMERICAL MODELING OF RETROFITTED MASONRY
5-23
NUMERICAL MODELING OF RETROFITTED MASONRY
5-24
NUMERICAL MODELING OF RETROFITTED MASONRY
(c) Deformation = 15 mm
Figure 5-14 Reinforced wall deformed shape (Scale factor = 5)
5-25
NUMERICAL MODELING OF RETROFITTED MASONRY
5-26
NUMERICAL MODELING OF RETROFITTED MASONRY
5-27
NUMERICAL MODELING OF RETROFITTED MASONRY
5-28
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-17 Effect of the steel bar diameter on the structure strength and stiffness
5-29
NUMERICAL MODELING OF RETROFITTED MASONRY
5-30
NUMERICAL MODELING OF RETROFITTED MASONRY
5-31
NUMERICAL MODELING OF RETROFITTED MASONRY
(a) ft=0.5MPa
(b) ft=1.00MPa
(c) ft=1.50MPa
Figure 5-22 Variation of Ppeak with friction and cohesion
5-32
NUMERICAL MODELING OF RETROFITTED MASONRY
(a) ft=0.5MPa
(b) ft=1.00MPa
(c) ft=1.50MPa
Figure 5-23 Variation of Ppeak / Pres with friction and cohesion
5-33
NUMERICAL MODELING OF RETROFITTED MASONRY
(a) ft=0.5MPa
(b) ft=1.00MPa
(c) ft=1.50MPa
Figure 5-24 Variation of Kres with friction and cohesion
5-34
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-25 Configurations considered for the parametric study of reinforced walls
5-35
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-26 Force deformation curve for reinforced wall (Strong wall, pitch=45mm)
Figure 5-27 Force deformation curve for reinforced wall (Strong wall, pitch=90mm)
Figure 5-28 Force deformation curve for reinforced wall (Weak wall, pitch=45mm)
5-36
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-29 Force deformation curve for reinforced wall (Weak wall, pitch=90mm)
5-37
NUMERICAL MODELING OF RETROFITTED MASONRY
5-38
NUMERICAL MODELING OF RETROFITTED MASONRY
5-39
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-32 Distribution of vertical normal stresses Deformation = 5mm (in MPa)
5-40
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-33 Distribution of vertical normal stresses Deformation = 10mm (in MPa)
5-41
NUMERICAL MODELING OF RETROFITTED MASONRY
5-42
NUMERICAL MODELING OF RETROFITTED MASONRY
5-43
NUMERICAL MODELING OF RETROFITTED MASONRY
Figure 5-36 Residual strength to peak strength ratio variation for different conditions
of PP-band mesh reinforcement
5-44
MATERIAL TESTING PROGRAM
6.1 Introduction
In the previous chapters the proposal of an economic method for masonry retrofitting was
presented. In order to verify the suitability of the proposed retrofitting technique, an
experimental program was designed and executed. Recently, scaled model specimens for
structural tests tend to become more and more popular. A real scale model test makes it
possible to obtain data similar to real structures. However, it requires large size testing
facilities and large amount research funds, so it is difficult to perform parametric tests. In this
experimental program scale models were used to investigate the statics and seismic
behavior of masonry walls. This experimental program consisted of two parts: material testing
and shaking table test.
The material testing program consisted of three sub parts: primary testing on PP-bands and
masonry, diagonal tension testing and out-of-plane testing. The first one concentrated on the
evaluation of the masonry parameters, i.e compression, tensile and bond strengths. The
variability of masonry properties is very high. Even if the same brick type and mortar mix are
used, differences in the construction procedure can lead to masonry assemblies with
considerably different properties. Because of this, the material testing program was also
oriented to identify the ranges in which the masonry properties varied for the same constituent
materials. Although a great care was put in keeping identical conditions for brick and mortar
during the construction, variability in the final product, i.e. masonry was observed.
The second part of the material testing consisted of diagonal tension tests for both non-
retrofitted and retrofitted masonry wallettes, to evaluate the beneficial effects of proposed PP-
band mesh reinforcement method. In addition to these, efficiency of different meshes
orientations was also examined. The third part of material test consisted of out-of-plane test to
investigate the PP-band mesh effectiveness in walls exhibiting out of plane action.
6-1
MATERIAL TESTING PROGRAM
bands were fixed on one end and axial tensile force was applied on other end. The initial
distance between two ends was 150 mm. The test was carried out under displacement control.
The results are shown in Figure 6-1. To calculate the stress in the band, its nominal cross
section 15.50.6mm2 was used. As the matter of fact, the band thickness is not uniform due to
the surface corrugation that it presents. The tensile strengths as well as the modulus of
elasticity are summarized in Table 6-1.
All the bands exhibited a large deformation capacity, more than 13% strain. The stress-
strain curve is fairly bilinear with an initial and residual modulus of elasticity of 3.19GPa and
1.06GPa respectively. Given its large deformation capacity, it is expected that it will
contribute to improve the structural ductility. Step-by-step failure was observed after PP-band
reached its peak strength. This behavior was observed due to individual failure of fibers in
PP-band (Photo 6-2) and was influenced by the deformation control loading.
In the primary material testing program, four different types of mortar mix were used. In
case of burned brick specimens, Cement/Water ratio of 0.14 and 0.25 were used. In case of
unburned, it was 0.33 and 0.45. Weight proportion of the mortar mixes is shown in Table 6-3.
6-2
MATERIAL TESTING PROGRAM
In the following sections, a brief discussion of some of these test results is presented.
Even a higher Cement/Water ratio was used for unburned brick specimens, it was observed
that they had lower shear strength than burned brick specimens. This behavior was observed
due to different failure mechanism. In unburned brick specimen case, the poor bonding
between mortar and brick led to separation along the interface. But in burned brick case,
failure occurred along the mortar. This difference is shown in Photo 6-4.
6-3
MATERIAL TESTING PROGRAM
6-4
MATERIAL TESTING PROGRAM
mortar joint thickness was 5mm for both cases. Cement/Water ratio 0.14 and 0.33 were used
for burned and unburned brick, respectively. For the present testing program, it was decided
to use a cement, lime, sand proportion of 1:7.9:20 in weight for burned units and in case of
unburned units it was 1:2.8:8.5 in weight. The mortar mix weight proportion is shown in
Table 6-8.
In this study totally 18 masonry wallettes were constructed. Nine test pieces per each
burned and unburned brick were prepared. For each case, three specimens were tested without
retrofitting. Next three specimens were retrofitted with PP-band meshes whose borders were
connected with epoxy and wire connectors. Remaining three specimens were retrofitted by
just overlapping of PP-band meshes and wire connectors. A total of 4 wire connectors were
used to attach the mesh with the masonry wallettes. The number of specimen types tested in
this experiment program is summarized in Table 6-9.
In addition to these, one more specimen was tested, to observe the efficiency of different
mesh orientations. Two types of PP-band mesh arrangement shown in Photo 6-7 were used:
Type-1: PP-band mesh oriented parallel to the masonry joints.
Type-2: PP-band mesh oriented 45 from the masonry joints.
Both had mesh pitch equal to 40mm. A total of 4 wire connectors were used to attach the
mesh with the masonry. In the retrofitted case epoxy was used for connecting PP-meshes
from both sides. In this way, any performance differences can be attributed only to the mesh
orientation.
The specimens were named according to the following convention: M-S-T-O-N in which:
M: B: Burned
U: Unburned
S: P: With external paste
X: Without external paste
T: NR: Non-retrofitted
RE: Retrofitted by PP-band meshes whose borders were connected with epoxy and
wire connectors
6-5
MATERIAL TESTING PROGRAM
The loading rate was 0.3mm/min and 2mm/min for the non-retrofitted and retrofitted cases
respectively. The retrofitted wallettes were applied up to 50mm vertical displacement.
6-6
MATERIAL TESTING PROGRAM
regained due to the PP-band mesh effect. Although at the end of the test almost all the mortar
joints were cracked, the retrofitted wallettes did not lose stability.
Figure 6-9 shows the shear compression strength variation with vertical deformation for
the non-retrofitted and retrofitted burned brick specimens. Although the specimen preparation
was kept as uniform as possible, some variation in test results was observed even before the
first crack. This difference can be attributed to several reasons: 1) mortar thickness variation,
2) Brick condition during laying, and 3) workmanship. After the first crack, it was observed
different behavior in diagonal tension strength in the retrofitted case. The main reasons for
that are the non uniformity of PP-band mesh pitch and workmanship in attaching mesh with
masonry wallettes. Considering average behavior of the specimens, results are discussed
below.
In the non-retrofitted case, the average initial strength was 1.64kN and there was no
residual strength after the first crack. In the retrofitted case, although the initial cracking was
followed by a sharp drop, at least 50 % of the peak strength remained. Subsequent drops were
associated with new cracks like the one observed at the deformation of 5mm in specimen B-
X-RE-H-1. After this, the strength was regained by readjusting and packing by PP-band mesh.
When the strength exceeded 3.0kN individual PP-bands started to fail. However, this did not
reduce considerably strength of the specimen, because stresses redistributed to other PP-bands.
The specimen quickly recovered its strength. The final strength of the specimen was equal to
3.0kN relatively higher than initial strength of 1.5kN. Initial compression strength of non-
retrofitted and retrofitted masonry wallettes are summarized in Table 6-11.
Table 6-11 Initial strength for burned brick specimens (Diagonal tension tests)
Diagonal tension strength (kN)
Specimen
Non-retrofitted Retrofitted
B-X-NR-0-1 1.85
B-X-NR-0-2 1.62
B-X-NR-0-3 1.47
B-X-RE-0-1 1.51
B-X-RE-0-2 2.03
B-X-RE-0-3 2.56
Average 1.64 2.03
For these cases, additionally to the vertical deformation, horizontal deformation was also
measured using LVDT to calculate the shear strain. Shear stress (SS) and shear strain () were
calculated with the following equation:
0.707 P
Ss =
An Eq. 6-1
where P is the applied load, and
6-7
MATERIAL TESTING PROGRAM
w+h
An = tn
2 Eq. 6-2
w is the specimen width, h the specimen height, t specimen total thickness, and n the
percentage of the gross area of the unit that is solid.
H + V
=
Lg Eq. 6-3
is the shear strain, V the vertical shortening, H the horizontal extension, and Lg the
vertical gauge length.
Figure 6-11 shows the shear stress variation with shear strain for the non-retrofitted and
retrofitted specimens.
Table 6-12 Initial strength for unburned brick specimens (Diagonal tension tests)
Specimen Diagonal tension strength (kN)
Non-retrofitted Retrofitted
U-X-NR-0-1 0.59
U-X-NR-0-2 0.89
U-X-NR-0-3 0.95
U-X-RE-0-1 0.92
U-X-RE-0-2 1.04
U-X-RE-0-3 0.96
Average 0.81 0.97
6-8
MATERIAL TESTING PROGRAM
6-9
MATERIAL TESTING PROGRAM
Cement/Water ratios equal to 0.25 and 0.45 were used for burned and unburned brick
respectively considering stability of the specimens during handling. For the present testing
program, it was decided to use a cement, lime, sand proportion of 1:4:11.2 in weight for
burned and in case of unburned it was 1:1.8:6.2 in weight as shown in Table 6-14.
6-10
MATERIAL TESTING PROGRAM
Tests were performed to characterize the engineering properties of the material used in
investigation. The average tensile strength of burned brick and unburned brick masonry
obtained from bond test were 0.1616MPa and 0.0056MPa respectively.
The specimens were named according to the following convention: M-T-O-N in which:
M: B: Burned
U: Unburned
T: NR: Non-retrofitted
RE: Retrofitted by overlapping of PP-band meshes and wire connectors
N: Specimen number
6-11
MATERIAL TESTING PROGRAM
case. The main reason for that is non uniformity of PP-band mesh pitch and workmanship in
attaching mesh with masonry wallettes. Results are discussed below, considering average
behavior of the specimens.
In the non-retrofitted case, the average initial strength was 0.67kN and there was some
residual strength remaining for further small amount of deformation after the first crack. This
behavior was observed due interlocking between bricks and also the application of load under
displacement control method. In the retrofitted case, although the initial cracking was
followed by a sharp drop, at least 30% of the peak strength remained. Initial out-of-plane load
strength of non-retrofitted and retrofitted masonry wallettes are summarized in Table 6-15.
6-12
MATERIAL TESTING PROGRAM
In retrofitted cases, except for U-RE-3, a sharp strength drop after initial crack was not
observed. Also behavior of U-RE-3 varied from other two specimens, because of different
failure type. In specimen U-RE-3, cracks occurred at one side of the support and propagated
to mid span as shown Figure 6-22. In other two specimens, crack initiated at mid-span as
same as non-retrofitted specimens.
Figure 6-23 shows the comparison between burned and unburned brick wallettes out-
of-plane behavior. As expected, initial strength of the burned brick specimen was relatively
higher than that of unburned brick one. Even higher cement/water ratio used for unburned
brick, the poor bonding between mortar and unburned brick led to bricks separation from the
mortar. But in burned brick case, failure occurred along the mortar. This behavior highly
influenced the initial strength of the specimens.
After initial strength drop, the mesh presence positively influenced the wallette
behavior. Both types of brick wallettes show the same behavior in strength up to vertical
deformation equal to 7mm. but that point, in unburned brick case, individual crushing was
observed. Due to that overall strength of the unburned brick wallettes considerably reduced
compared with burned brick wallettes. There after if two type of bricks are compared, almost
40% different in strength was observed.
6.6 Summary
The present chapter introduced the experimental program, which was executed for two main
purposes: 1) to verify the suitability of the retrofitting technique with PP-band meshes, and 2)
to create a dataset for the verification of future numerical modeling.
The testing program consisted of three parts: primary testing of PP-bands and masonry,
diagonal tension testing and out-of-plane testing. Diagonal tension tests and out-of-plane tests
allowed the analysis and evaluation of the performance of the PP-band mesh retrofit in both
in-plane and out-of-plane behavior. In the retrofitted case, larger residual strength after the
formation of the first diagonal cracks was observed in diagonal tension tests. Furthermore, as
6-13
MATERIAL TESTING PROGRAM
deformation increased, the wallette achieved strength higher than the initial cracking strength.
Moreover, the failure was ductile compared to the brittle failure observed in the unreinforced
walls. This behavior was also observed in out-of-plane tests after crack at mid-span.
The present testing program was limited mainly to a PP-band mesh pitch equal to 40mm.
Further testing with different PP-band mesh pitch, is necessary to evaluate this effect.
6.7 Reference
[6-1] P.Mayorca, Strengthening of Unreinforced Masonry Structures in Earthquake Prone
Regions, Ph.D. Dissertation, University of Tokyo, Japan, 2003.
[6-2] N.Galati, J.G.Tumialan, A.Nanni, A.La Tegola, Influence of Arching Mechanism in
Masonry Walls Strengthening with FRP Laminates, University of Missouri, Rolla, Italy.
[6-3] T.Okada, F.Kumazawa, S.Horiuchi, M.Yamamoto, A.Fujioka, K.Shinozaki, Y.Nakano,
Shaking Table Tests of Reinforced Concrete Small Scaled Model Structures, Bull. ERS,
No.22 (1989).
[6-4] ASTM, E72 Standard Test Methods of Conducting Strength Tests of Panels for Building
Construction, American Society for Testing and Materials, Philadelphia, Pa, (1998).
6-14
MATERIAL TESTING PROGRAM
100
50
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
strain
121 50
75
6-15
MATERIAL TESTING PROGRAM
0.06
0.04
0.02
0.00
0.00 0.05 0.10 0.15
Deformation (mm)
0.20
Shear stress (MPa)
0.15
0.10
0.05 B-SP4
B-SP5
B-SP6
0.00
0.00 0.05 0.10 0.15
Deformation (mm)
6-16
MATERIAL TESTING PROGRAM
0.008
Shear stress (MPa)
0.006
0.004
0.002 U-SP1
U-SP2
U-SP3
0.000
0.00 0.05 0.10 0.15
Deformation (mm)
0.006
Shear stress (MPa)
0.004
0.002
U-SP1
U-SP2
U-SP3
0.000
0.00 0.05 0.10 0.15
Deformation (mm)
6-17
MATERIAL TESTING PROGRAM
20
Compression stress (MPa)
15
10
5
B-SPC 1
B-SPC 2
B-SPC 3
0
0 0.5 1 1.5 2 2.5
Deformation (mm)
5
Compression stress (MPa)
1 U-SPC 1
U-SPC 2
U-SPC 3
0
0 0.3 0.6 0.9 1.2 1.5
Deformation (mm)
6-18
Diagonal compression test - Burned brick
5
4
MATERIAL TESTING PROGRAM
Figure 6-9 Force versus vertical deformation (Diagonal tension test Burned brick)
Vertical deformation (mm)
6-19
Diagonal compression test - Burned brick
3
2.5
2
MATERIAL TESTING PROGRAM
1.5
Figure 6-10 Force versus vertical deformation (Diagonal tension test Burned brick)
6-20
Diagonal compression test - Burned brick
0.25
0.20
MATERIAL TESTING PROGRAM
0.15
Stress (MPa)
0.10
0.05
B-X-NR-0-1
B-X-RE-H-1
B-X-RE-H-2
B-X-RE-H-3
0.00
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12
Shear strain
Figure 6-11 Shear stress versus shear strain (Diagonal tension test Burned brick)
6-21
Diagonal comprssion test - Unburned brick
3.5
U-X-NR-H-1
U-X-NR-H-2
3.0 U-X-NR-H-3
U-X-RE-H-1
U-X-RE-H-2
2.5 U-X-RE-H-3
MATERIAL TESTING PROGRAM
2.0
1.5
0.5
0.0
0 5 10 15 20 25 30 35 40
Figure 6-12 Force versus vertical deformation (Diagonal tension test Unburned brick)
6-22
Diagonal compression test - Unburned brick
0.1
0.08
MATERIAL TESTING PROGRAM
0.06
0.04
Figure 6-13 Shear stress versus shear strain (Diagonal tension test Unburned brick)
6-23
MATERIAL TESTING PROGRAM
40
35
30
Diagonal compression test comparison
25
Vertiacal Deformation (mm)
20
15
10
U-X-RE-H-1
U-X-RE-H-2
U-X-RE-H-3
B-X-RE-H-1
B-X-RE-H-2
B-X-RE-H-3
5
0
5
Figure 6-14 Force versus vertical deformation - comparison between burned and unburned
brick specimen retrofitted with PP-band meshes whose borders were connected with epoxy
6-24
Diagonal compression test - Burned brick
4.5
4.0
3.5
3.0
MATERIAL TESTING PROGRAM
2.5
2.0
band mesh.
Compression Force (kN)
1.5
1.0
B-X-RE-H-1
0.5 B-P-RE-H-1
B-P-RE-I-1
0.0
0 5 10 15 20 25 30 35 40
Vertical Deformation (mm)
Figure 6-15 Comparison between masonry wallettes retrofitted by Type 1 and Type 2 PP-
6-25
Diagonal compression test - Burned brick
5
4
MATERIAL TESTING PROGRAM
(Burned brick)
Compression Force (kN)
B-X-NR-0-1
B-X-RE-H-1
B-X-RE-H-2
1
B-X-RE-H-3
B-X-RO-H-1
B-X-RO-H-2
B-X-RO-H-3
0
0 5 10 15 20 25 30 35 40
Vertical deformation (mm)
Figure 6-16 Behavior of masonry wallettes with mesh edges fully or partially connected
6-26
Diagonal compression test - Unburned brick U-X-NR-H-1 U-X-RE-H-1
3.5
U-X-RE-H-2 U-X-RE-H-3
U-X-RO-H-1 U-X-RO-H-2
U-X-RO-H-3
3.0
2.5
MATERIAL TESTING PROGRAM
2.0
1.5
(Unburned brick)
Compression Force (kN)
1.0
0.5
0.0
0 5 10 15 20 25 30 35 40
Vertical Deformation (mm)
Figure 6-17 Behavior of masonry wallettes with mesh edges fully or partially connected
6-27
Diagonal compression test comparison
5.0
B-X-RO-H-1
B-X-RO-H-2
B-X-RO-H-3
4.0 U-X-RO-H-1
U-X-RO-H-2
U-X-RO-H-3
MATERIAL TESTING PROGRAM
3.0
2.0
0.0
0 5 10 15 20 25 30 35 40
Vertical Deformation (mm)
Figure 6-18 Comparison between burned and unburned masonry wallettes retrofitted by just
6-28
MATERIAL TESTING PROGRAM
Type 1 failure
2
1
Type 2 failure
6-29
Out-of-plane test - Burned brick
1.6
1.4
1.2
1.0
MATERIAL TESTING PROGRAM
0.8
0.6
Figure 6-21 Load versus deformation relation for burned brick out-of-plane test
6-30
Out-of-plane test - Unburned brick
1.0
0.8
MATERIAL TESTING PROGRAM
0.6
0.4
Figure 6-22 Load versus deformation relation for unburned brick out-of-plane test
6-31
Out-of-plane test - comparison
1.6
1.4
1.2
1.0
MATERIAL TESTING PROGRAM
0.8
0.6
Figure 6-23 Out-of-plane load comparison between burned and unburned brick specimen
6-32
MATERIAL TESTING PROGRAM
6-33
MATERIAL TESTING PROGRAM
Photo 6-4 Failure pattern of burned and unburned in direct shear test
6-34
MATERIAL TESTING PROGRAM
6-35
MATERIAL TESTING PROGRAM
6-36
MATERIAL TESTING PROGRAM
Photo 6-10 Failure patterns of brick masonry wallettes without and with retrofitting
6-37
MATERIAL TESTING PROGRAM
6-38
MATERIAL TESTING PROGRAM
6-39
MATERIAL TESTING PROGRAM
6-40
MATERIAL TESTING PROGRAM
Photo 6-17 Failure patterns of brick masonry wallettes without and with retrofitting
6-41
SHAKING TABLE TESTS BURNED BRICK MASONRY
7.1 Introduction
In the previous chapter the material testing program designed to verify the applicability of the
proposed retrofitting technique was presented. In order to understand the dynamic response of
masonry houses with and without PP-band mesh retrofitting, crack patterns, failure behavior, and
overall effectiveness of the retrofitting technique, shaking table tests were carried out. These tests
were also intended to collect a dataset that will be used in the future to calibrate the numerical
model.
A real scale model test makes possible to obtain data similar to real structures. However, it
requires large size testing facilities and large amount research funds, so it is difficult to execute
parametric tests by using full scaled models. Recently, structural tests of scaled models became
larger and larger as the overall behavior of the system can be understood from scaled model also.
In this experimental program scale models were used to investigate the seismic behavior of
masonry houses and effectiveness of PP-band retrofitting technique. This chapter describes the
response characteristics obtained by the shaking table tests carried out.
7-1
SHAKING TABLE TESTS BURNED BRICK MASONRY
order to keep the simplicity of the structure for future numerical modeling.
Both specimens consisted of 18 rows of 44 bricks in each layer except at the opening levels.
Construction process took place in two days, first 11 rows in first day and remaining rows in
following day. Specimen U-B-40 was not retrofitted and specimen R-B-40 was retrofitted by PP-
band meshes after construction.
7.2.3 Instrumentation
The shaking table system (Photo 7-3) available at Institute of Industrial Science, The University
of Tokyo is capable of controlling six degrees of freedom and operating in frequencies ranging
from 0.1 to 50 Hz. It has a maximum displacement of 100mm and maximum weight of the
specimen is 2 tons. In this experiment, one direction shaking was applied for simplicity of motion.
To assess the global and local behavior, specimens were instrumented to measure
accelerations and displacements. During the tests, twelve accelerometers, four with three-
dimensional measurement capacity and eight with one-dimensional measurement capacity were
installed at the locations shown in Figure 7-3. The number of accelerometers was 12, 4 and 4 in
7-2
SHAKING TABLE TESTS BURNED BRICK MASONRY
The measured data were recorded continuously throughout the tests. The sampling rate was
1/500 sec in the all runs.
7-3
SHAKING TABLE TESTS BURNED BRICK MASONRY
Initially, a sweep acceleration signal with amplitude 0.05g was applied to identify the
dynamic properties of the models. Both U-B-40 and R-B-40 were subjected to a sequence of
sinusoidal excitations as shown in Table 7-4 until the final damage state was attained. A total of
46 and 62 test runs were applied to U-B-40 and R-B-40, respectively.
The shaking characteristics are summarized in Figure 7-5 where the expected and excited
model base peak accelerations are presented. It was observed that the expected input acceleration
amplitudes slightly differ from measured acceleration of base of the model in the case of higher
amplitude input acceleration.
7-4
SHAKING TABLE TESTS BURNED BRICK MASONRY
was totally separated from the specimen. It was removed from specimen before next
test run proceed.
In case of R-B-40, many cracks were observed in all the walls. Particularly, in north
and south wall more cracks that at the same stage in U-B-40 were observed. Although
PP-band mesh kept the structure integral during the shaking, it allowed the sliding of
the bricks along these cracks to some extent. (Figure 7-11)
Run 46 In U-B-40, existing cracks widened and connection between adjacent walls weaken.
As a result each of the connecting walls became an independent structure, which is the
worst-case scenario, because each is supported only at the bottom. Therefore, the walls
subjected to out-of-plane load failed leading to the structure collapse. This was the last
run for U-B-40. In R-B-40, although extensive cracking was observed, as the PP-band
mesh kept the structure integral during the shaking, it prevent the structure from
collapse. (Figure 7-12)
Run 61 Although at the end of this run almost all the mortar joints were cracked, the specimen
did not lose stability. This input motion was 1.25 times the acceleration and 8 times the
displacement applied in Run 46. Another important point to mention is that the
retrofitted model sustained 15 more runs with higher input energy before this run.
This indicates that retrofitted house could withstand base displacements 8 times larger and
velocities 3 times higher than the non-retrofitted house. Photo 7-4 and Photo 7-5 show non-
retrofitted and retrofitted model at the final stage.
7-5
SHAKING TABLE TESTS BURNED BRICK MASONRY
Table 7-5 Performance of Non-retrofitted Building Model with Different JMA Intensities
Acceleration Frequency (Hz)
(g) 2 5 10 15 20 25 30 35
1.4
1.2
1.0
0.8 DG4 DG3 DG2 DG2 DG1 DG1
0.6 DG5 DG3 DG3 DG2 DG1 DG1 DG1
0.4 DG4 DG3 DG2 DG2 DG1 DG1 DG1
0.2 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.1 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.05 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
Table 7-6 below shows the performance of the retrofitted building model at different JMA
intensities.
Table 7-6 Performance of Retrofitted Building Model with Different JMA Intensities
Acceleration Frequency (Hz)
(g) 2 5 10 15 20 25 30 35
1.4 DG3 DG3 DG3
1.2 DG2 DG2 DG2 DG2
1.0 DG4 DG2 DG2 DG2 DG2
0.8 DG4 DG2 DG2 DG2 DG2 DG2 DG1 DG1
0.6 DG3 DG2 DG2 DG2 DG2 DG1 DG1 DG1
0.4 DG2 DG2 DG2 DG2 DG2 DG1 DG1 DG1
0.2 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.1 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.05 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
The retrofitted building performed at moderate damage level until Run 46 at which the non-
retrofitted building collapsed. Moreover, moderate damage level of performance was maintained
7-6
SHAKING TABLE TESTS BURNED BRICK MASONRY
for 10 more runs leading to intensity JMA 6-. Even in a very high level of shaking, Run 62 with
intensity JMA 6+, the building survived from collapse. Considering the pre-damage level of the
building caused by previous loadings, it can be interpreted that this retrofitting technique can
achieve reasonable safety even in the worst case scenario earthquakes of JMA 7. Further, this
technique may also be applicable to retrofit earthquake damaged buildings as it was effective
even after masonry had severe cracks. As the model was already considerably deformed beyond
the limit of measurement system the test was stopped after Run 62. It should be noted again that
this building model survived 15 more shakings in which many runs had intensities higher than
JMA 5-.
7.5 Summary
The present chapter introduced the shaking table testing program, which is executed for two main
purposes; 1) to understand the dynamic response of masonry houses with and without PP-band
mesh retrofitting, and 2) to create a dataset for the verification of future numerical modeling.
Two identical 1:4 scaled models were built using burnt bricks as masonry units and cement,
lime and sand (1:8:20) mixture as mortar. This mix was specially designed to obtain mechanical
properties similar to those found in masonry houses in developing countries even though the
construction materials used were those available in Japan.
From the tests it was observed that, partial collapse of the non-retrofitted house occurred in
Run 44 and the building totally collapsed in Run 46. Under the same shaking, the retrofitted
house stood as one piece. The retrofitted house was shaken until the Run 62. The structure
exhibited remarkable deformations because of the large number of failed mortar joints.
Nevertheless, the PP-band mesh provided confinement and held the disintegrated elements
together preventing the collapse. Considering overall performance of both specimens, PP-band
can effectively increase the seismic capacity of masonry houses and therefore reduce the number
of casualties in the coming earthquakes.
7.6 Reference
[1] T. Okada, F. Kumazawa, S. Horiuchi, M. Yamamoto, A. Fujioka, K. Shinozaki, Y. Nakano,
Shaking table tests of reinforced concrete small scaled model structures, Bull. ERS, No.22
(1989).
[2] Sergio M. Alcocer, Juan Guillermo Arias, Alejandro Vazquez, Response assessment of
Mexican confined masonry structures through shaking table tests, 13th World Conference
on Earthquake Engineering, Vancouver, B.C., Canada, August 1-6, 2004, Paper No. 2130.
[3] Roko Zarnic, Samo Gostic, Adam J. Crewe and Colin A.Taylorm, Shaking table tests of
1:4 reduced scale models of masonry infilled reinforced concrete frame buildings,
Earthquake Engineering and Structural Dynamics. 2001, Vol.30, 819-834.
7-7
SHAKING TABLE TESTS BURNED BRICK MASONRY
[4] Miha Tomazevic and Iztok Klemenc, Verification of seismic resistance of confined
masonry buildings, Earthquake Engineering and Structural Dynamics. 1997, Vol.26, 1073-
1088.
[5] P. Mayorca, Strengthening of unreinforced masonry structures in earthquake prone regions,
Ph.D. Dissertation, University of Tokyo, Japan, 2003.
[6] D. Benedetti, P. Carydis and M. P. Limongelli, Evaluation of the Response of Masonry
Buildings Based on energy functions, Earthquake Engineering and Structural Dynamics.
2001, Vol.30, 1061-1081.
[7] European Macroseismic Scale 1998 (EMS-98), Grnthal, G. (ed.), European Seismological
Commision, 1998.
7-8
SHAKING TABLE TESTS BURNED BRICK MASONRY
7-9
SHAKING TABLE TESTS BURNED BRICK MASONRY
7-10
SHAKING TABLE TESTS BURNED BRICK MASONRY
L4 L7
L2
L5 L1
L3
L6
7-11
1.8
Expected base acceleration
1.6 Retrofitted speci base acceleration
Non-retrofitted speci base acceleration
1.4
1.2
1.0
0.8
Acceleration (g)
SHAKING TABLE TESTS BURNED BRICK MASONRY
0.6
0.4
0.2
3
6
9
12
15
18
21
24
27
30
33
36
39
42
45
48
51
54
57
60
Loading Run
7-12
SHAKING TABLE TESTS BURNED BRICK MASONRY
INITIAL STAGE
U-B-40
R-B-40
7-13
SHAKING TABLE TESTS BURNED BRICK MASONRY
U-B-40
R-B-40
7-14
SHAKING TABLE TESTS BURNED BRICK MASONRY
U-B-40
R-B-40
7-15
SHAKING TABLE TESTS BURNED BRICK MASONRY
U-B-40
R-B-40
7-16
SHAKING TABLE TESTS BURNED BRICK MASONRY
U-B-40
R-B-40
7-17
SHAKING TABLE TESTS BURNED BRICK MASONRY
U-B-40
R-B-40
7-18
SHAKING TABLE TESTS BURNED BRICK MASONRY
U-B-40
R-B-40
7-19
SHAKING TABLE TESTS BURNED BRICK MASONRY
7-20
SHAKING TABLE TESTS BURNED BRICK MASONRY
7-21
SHAKING TABLE TESTS BURNED BRICK MASONRY
7-22
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
8-1
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
Table 8-2 Mechanical properties of masonry
U-A-40 R-A-40
Compressive strength (MPa) 4.40 4.26
Shear strength (MPa) 0.0064 0.0055
Bond strength (MPa) 0.0045 0.0050
Diagonal compression strength (MPa) 0.81 0.97
Both specimens consisted of 18 rows of 44 bricks in each layer except at the opening levels.
The construction process took place in two days, first 11 rows in the first day and remaining rows
construct in the following day. The specimen U-A-40 was not retrofitted, but specimen R-A-40
was retrofitted by PP-band meshes after construction.
8.1.3 Instrumentation
To assess the global and local behavior, specimens were instrumented to measure accelerations
and displacements. During the tests, ten accelerometers, four with three-dimensional
measurement capacity and six with one-dimensional measurement capacity were installed at the
locations shown in Figure 8-1. The number of accelerometers was 10, 4 and 4 in the exciting,
transverse and vertical direction respectively.
Five lasers, in N-S direction were used to measure displacements. The locations of laser
measuring instruments are shown in Figure 8-1. Lasers L1, L2, L3 aimed at obtaining the North
wall deformation at the top level in the motion direction. Laser L4 recorded the facade wall
deformation at centre. The instrument specifications are given Table 8-3.
The measured data were recorded continuously throughout the tests. The sampling rate was
1/500 sec in the all runs.
8-2
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
considered because of its adequacy for later use in the numerical modeling. The numbers given in
the Table 8-4 show the loading sequence followed for the two tests.
Initially, a sweep acceleration signal with amplitude 0.05g was applied to identify the
dynamic properties of the models. Both U-A-40 and R-A-40 were subjected to a sequence of
sinusoidal excitations as shown in Table 8-4, until the final damage state was attained. A total of
44 and 53 test runs were applied to U-A-40 and R-A-40, respectively.
The shaking characteristics of model excited on shaking table are summarized in Figure 8-2
where the expected and excited model base peak accelerations are presented. It was observed that
the expected input acceleration amplitudes slightly differed from measured acceleration of the
model base in the case of higher amplitude input acceleration.
8-3
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
propagated up to the wall top layer. In R-A-40 cracks propagated horizontally from top
and bottom corner of the openings. Also, major cracks were observed in walls
perpendicular to the shaking direction (Figure 8-5).
Run 37 In U-A-40,X type crack was observed in both walls, which are in direction of the
shaking. In R-A-40, X crack was observed only on top part of the walls with
openings (Figure 8-6).
Run 40 In U-A-40, a large crack was observed at the connection between adjacent walls. In
case of walls parallel to the shaking direction, the top part of the west wall (above the
door opening) was totally cracked. In case of R-A-40, many cracks were observed in
all the walls. Particularly, in north and south wall more cracks were observed than at
the same stage of U-A-40. Although PP-band mesh kept the structure integral during
the shaking, it allowed the sliding of the bricks along these cracks to some extent.
(Figure 8-7)
Run 43 In U-A-40 the top part of the wall (above the window opening) totally separated from
the specimen and fell. (Figure 8-8)
Run 44 In U-A-40, existing cracks widened and the connection between adjacent walls
weakened. As a result each of the connecting walls became independent structures,
which is the worst-case scenario, supported only at the bottom. The walls subjected to
out-of-plane loads worked as cantilevers and failed. This finally led to the structure
collapse. This was the last run for U-A-40. In R-A-40, although extensive cracking
was observed, as the PP-band mesh kept the structure integral during the shaking, it
prevent the structure from collapse. (Figure 8-9)
Run 53 Although at the end of this run almost all the mortar joints were cracked, the specimen
did not lose stability (Figure 8-10). This input motion had twice the acceleration and
12.5 times the displacement that the input motion corresponding to Run 44. Another
important point to mention is that the retrofitted model sustained 9 more runs than U-
A-40 with higher input energy before this run.
The retrofitted house could withstand base displacements 12.5 times larger and velocities 5
times higher than the non-retrofitted house. Photo 8-3 and Photo 8-4 show the non-retrofitted and
retrofitted model at the final stage.
8-4
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
different runs. Table 8-5 shows the non-retrofitted building model damage with different JMA
intensities.
Table 8-5 Performance of Non-retrofitted Building Model with Different JMA Intensities
Acceleration Frequency (Hz)
(g) 2 5 10 15 20 25 30 35
1.4
1.2
1.0
0.8 DG4 DG3 DG2 DG2 DG1 DG1
0.6 DG4 DG3 DG2 DG2 DG1 DG1
0.4 DG4 DG4 DG2 DG2 DG2 DG1 DG1
0.2 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.1 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.05 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
Table 8-6 below shows the performance of the retrofitted building model at different JMA
intensities.
Table 8-6 Performance of Retrofitted Building Model with Different JMA Intensities
Acceleration Frequency (Hz)
(g) 2 5 10 15 20 25 30 35
1.4 DG3
1.2 DG3
1.0 DG3
0.8 DG4 DG2 DG2 DG2 DG2 DG2 DG1 DG1
0.6 DG3 DG2 DG2 DG2 DG2 DG2 DG1 DG1
0.4 DG3 DG2 DG2 DG2 DG2 DG1 DG1 DG1
0.2 DG2 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.1 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
0.05 DG1 DG1 DG1 DG1 DG1 DG1 DG1 DG1
The retrofitted building exhibited moderate damage after Run 44, at which the non-retrofitted
building collapsed. Even at a very high level of shaking, Run 53 with intensity JMA 6+, the
building survived from collapse. Further, this technique may also be applicable to retrofit
8-5
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
earthquake damaged buildings as it was effective even after masonry had severe cracks. As the
model was already considerably deformed beyond the limit of measurement system the test was
stopped after Run 53. It should be noted again that this building model survived 9 more shakings
in which many runs had intensities higher than JMA 5-.
8.4 Summary
The present chapter introduced the shaking table testing program, which is executed for two main
purposes: 1) to understand the dynamic response of masonry houses with and without PP-band
mesh retrofitting, and 2) to create a dataset for the verification of future numerical modeling.
Two identical 1:4 scaled models were built using burnt bricks as masonry units and cement,
lime and sand (1:3:8) mixture as mortar. From the tests it was observed that, partial collapse of
the non-retrofitted house occurred in Run 43 and the building totally collapsed in Run 44. Under
the same shaking, the retrofitted house stood as one piece. The retrofitted house was shaken until
Run 54. The structure exhibited remarkable deformations because of the large number of failed
mortar joints. Nevertheless, the PP-band mesh provided confinement and held the disintegrated
elements together preventing the collapse. Considering overall performance of both specimens, it
is concluded that PP-band can effectively increase the seismic capacity of masonry houses and
therefore reduce the number of casualties in the coming earthquakes.
8.5 References
[1] T. Okada, F. Kumazawa, S. Horiuchi, M. Yamamoto, A. Fujioka, K. Shinozaki, Y. Nakano,
Shaking table tests of reinforced concrete small scaled model structures, Bull. ERS, No.22
(1989).
[2] Sergio M. Alcocer, Juan Guillermo Arias, Alejandro Vazquez, Response assessment of
Mexican confined masonry structures through shaking table tests, 13th World Conference
on Earthquake Engineering, Vancouver, B.C., Canada, August 1-6, 2004, Paper No. 2130.
[3] Roko Zarnic, Samo Gostic, Adam J. Crewe and Colin A.Taylorm, Shaking table tests of
1:4 reduced scale models of masonry infilled reinforced concrete frame buildings,
Earthquake Engineering and Structural Dynamics. 2001, Vol.30, 819-834.
[4] Miha Tomazevic and Iztok Klemenc, Verification of seismic resistance of confined
masonry buildings, Earthquake Engineering and Structural Dynamics. 1997, Vol.26, 1073-
1088.
[5] P.Mayorca, Strengthening of unreinforced masonry structures in earthquake prone regions,
Ph.D. Dissertation, University of Tokyo, Japan, 2003.
[6] D.Benedetti, P. Carydis and M.P.Limongelli, Evaluation of the Response of Masonry
Buildings Based on energy functions, Earthquake Engineering and Structural Dynamics.
2001, Vol.30, 1061-1081.
8-6
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
[7] European Macroseismic Scale 1998 (EMS-98), Grnthal, G. (ed.), European Seismological
Commision, 1998.
8-7
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
8-8
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
8-9
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
INITIAL STAGE
U-A-40
R-A-40
8-10
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
RUN 25 5Hz / 0.2g
U-A-40
R-A-40
8-11
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
RUN 34 25Hz / 0.8g
U-A-40
R-A-40
8-12
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
RUN 37 20Hz / 0.8g
U-A-40
R-A-40
8-13
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
RUN 40 15Hz / 0.8g
U-A-40
R-A-40
8-14
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
RUN 43 10Hz / 0.8g
U-A-40
R-A-40
8-15
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
U-A-40
R-A-40
8-16
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
R-A-40
8-17
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
8-18
SHAKING TABLE TESTS UNBURNED BRICK MASONRY
8-19
CONCLUSIONS AND RECOMMENDATIONS
9. Conclusions and
recommendations
9.1 Introduction
As mentioned at the beginning of the present report, unreinforced masonry is one of the most
used construction materials in the world. It is also, unfortunately, the most vulnerable during
earthquakes. More than 30% of the worlds population presently lives in earthen structures. It
is expected that unreinforced masonry dwellings will continue to shelter important portions of the
population of developing countries. Consequently, a large number of casualties due to masonry
collapse are expected in future earthquakes.
To tackle the previously stated problem, an economic and efficient retrofitting scheme for
unreinforced masonry is needed. This method should stress the use of locally available materials
and workmanship. Masonry is a material that has a great variation around the world. Although it
would be desirable to test any proposed retrofitting technique on every type of masonry, this
is practically impossible. This situation could be overcome by means of a powerful numerical
analysis tool.
In this background, the present research focuses not only on the proposal of a technique for
strengthening unreinforced masonry dwellings but also on the development of a numerical tool
for the analysis of non-retrofitted and retrofitted masonry structures. The findings of this on-
going work are discussed below. Further developments will be reported in the near future.
A new retrofitting technique for unreinforced masonry using polypropylene bands, commonly
used for packing, was proposed and implemented. These bands are inexpensive, strong, durable,
light, and worldwide available. A step-by-step construction procedure was developed. Issues
such as how to connect the bands, how to prepare the mesh and how to protect it were addressed.
Conclusions and findings are classified in two categories: relevant to the experimental program
and relevant to the numerical modeling.
9-1
CONCLUSIONS AND RECOMMENDATIONS
9-2
CONCLUSIONS AND RECOMMENDATIONS
9-3
CONCLUSIONS AND RECOMMENDATIONS
mid-span deformations equal to 7mm. After that point, brick crushing was observed in the
unburned unit case. Due to this, the overall strength of the unburned brick wallettes
considerably reduce compared with that of burned ones. There after if the two types of brick
specimens are compared, almost 40% difference in strength was observed.
9-4
CONCLUSIONS AND RECOMMENDATIONS
5. The parametric study of the reinforced walls showed that the mesh pitch and connection
conditions strongly influence the in-plane behavior. A large pitch mesh appropriately
connected can have similar effect as a smaller pitch mesh with fewer connectors. The effect of
properly installed cover mortar to improve the connection between mesh and wall was also
recognized through the numerical simulation.
6. The reinforcement effect could be separated in two components: the first is related to the
reinforcement strength itself whereas the second is related to the mesh effect on the masonry
behavior. The first component is relatively small especially in the small deformation range.
9-5
CONCLUSIONS AND RECOMMENDATIONS
9-6
CONCLUSIONS AND RECOMMENDATIONS
from very simple experiments. The material model used in the present study included easy to
obtain parameters such as bond, cohesion, compression strength, etc. However, it also included
fracture energies in tension and shear. During the experimental program, it was realized that
obtaining these parameters is not easy due to the brittleness of the phenomena. Very accurate
and high frequency sampling measuring devices are needed to estimate them.
The objective of developing a numerical tool is to analyze a wide variety of masonry
structures. If the material model input parameters are not available, the most probable situation
is that the missing parameters would be estimated. Unfortunately, there is little information
available in the literature regarding the ranges in which the fracture energies of masonry mortar
joint fluctuate. Thus, it would be desirable to increase this experimental dataset.
Another issue that needs to be addressed is the understanding of the mortar properties,
especially the Youngs modulus, in the masonry assembly. The properties obtained through
the testing of mortar cylinders are different from those of the mortar inside masonry because of
the different conditions during the construction, aspect ratio, and restraint provided by the
bricks. The experimental research in this direction is limited and needs to be expanded in order to
be able to obtain reliable parameters for the numerical modeling.
The extension of the Applied Element Method for the 3D analysis is also necessary. This
improvement will allow the discussion of masonry out-of-plane behavior and the interaction
between adjacent walls. As mentioned before, these phenomena are observed during earthquakes
and may result in the structure collapse. The retrofitting effect on these failure modes can be also
investigated by the 3D approach. This point is currently under development.
Finally, the verification of the model under dynamic loads is a necessary step to analyze the
shaking table experimental data.
9-7