Professional Documents
Culture Documents
by
A THESIS
CALGARY, ALBERTA
JANUARY, 2009
ISBN: 978-0-494-51131-2
Abstract
There is a tremendous need for affordable, safe housing in developing regions, yet with
the economic hardships of many of these nations, traditional materials such as concrete
and steel are unaffordable. One of the potential solutions to this problem is the use of
indigenous materials, which can greatly reduce the cost of housing projects.
An experimental testing programme was developed to find the suitability of using the
steel reinforcement in masonry shear walls. A total of seven masonry shear walls were
tested in-plane under cyclic loading conditions. One wall contained conventional steel,
while the others contained Tonkin cane reinforcement. In addition to these cyclic tests,
material testing was conducted on all components of the walls. A complete analysis of
the results was carried out, showing great potential for bamboo as a low cost housing
reinforcing material.
iii
Acknowledgements
I would like to extend my uttermost appreciation to all of the people that have helped me
with the completion of this thesis, especially to my academic supervisor, Dr. Shelley
Lissel, whose patience, guidance and expertise defined this project. I would also like to
thank my interim academic supervisor, Dr. Nigel Shrive, for his advice and constant
willingness to aid me. I would like to also express my gratitude to all of the professors
and administrative staff of the Civil Engineering Department, with all the help they have
provided along the way. The laboratory work that I undertook was paramount in this
research, and could not have been completed with out the technical staff at the laboratory,
as I would like to thank all of them, especially Dan Tilleman, Terry Quinn, Don Anson,
Mirsad Berbic and Don McCullough. I would also like to thank master mason George
Larocque for constructing all of my wall specimens, as well as Cliff Lomenda and H.O.
Block for the generous donation of block. The financial support of the University of
Calgary Civil Engineering Department, the Schulich Student Activities Fund (SSAF), and
the Energy and Environment Access Expansion Program are also gratefully
acknowledged. All of my family and friends and colleagues have given me so much
especially like to acknowledge Gerd Birkle, Mark Hagel, Iain Gidley, Jocelyn Dickie,
iv
Dedication
Words cannot express my gratitude for all that you have done for me.
v
Table of Contents
Approval Page................................................................................................................ ii
Abstract......................................................................................................................... iii
Acknowledgements ........................................................................................................iv
Dedication .......................................................................................................................v
Table of Contents ...........................................................................................................vi
List of Tables..................................................................................................................ix
List of Figures .................................................................................................................x
List of Symbols and Abbreviations ...............................................................................xiv
BIBLIOGRAPHY .......................................................................................................132
vii
APPENDIX A: ADDITIONAL PHOTOS..................................................................140
A.1 Initial Testing Frame.........................................................................................140
A.2 Final Testing Frame ..........................................................................................141
A.3 Wall QS.B.1......................................................................................................142
A.4 Wall QS.B.2......................................................................................................143
A.5 Wall QS.B.3......................................................................................................144
A.6 Wall QS.B.4......................................................................................................145
A.7 Wall QS.B.5......................................................................................................146
A.8 Wall QS.B.6.L ..................................................................................................147
A.9 Wall QS.S.1 ......................................................................................................148
viii
List of Tables
Table 2.1 - Bamboo vs Other Construction Material Properties (Abang & Aband
1983) .....................................................................................................................18
Table 4.1 - Calculation Results for Bilinear Idealization and Ductility .........................100
Table 4.5 - Shear Equation Comparison for Walls with Diagonal Tension Failure .......123
ix
List of Figures
Figure 1.2 - World Distributions of Moderate and High Seismic Hazard (De Sensi
2008) .......................................................................................................................5
Figure 2.7 - Bamboo Pull Out Tests: a) Improved; b) Conventional (Ghavami 2005).....24
Figure 2.8 - Shear Transfer Devices for Bamboo (Subrahmanyam 1984) .......................25
Figure 2.9 - Comparison of the Deformation Behaviour of Beams with Steel and
Bamboo Reinforcements (Subrahmanyam 1984) after (Youssef 1976)...................27
x
Figure 2.15 - Equilibrium of Sectional Forces at Flexural Failure of a Masonry Wall
(Drysdale & Hamid 2005)......................................................................................37
Figure 2.19 - Typical Lateral Displacement Time Histories (Tomazevic and Lutman
1996) .....................................................................................................................43
Figure 2.22 - (a) Evaluation of Energy Input in One Loading Cycle, (b) Evaluation of
Dissipated Hysteretic Energy in One Loading Cycle (Tomazevic et al. 1996).........49
xi
Figure 3.10 - Initial Bond Test Specimens and Apparatus ..............................................65
Figure 4.1 - Hysteresis Loops and Crack Pattern for Wall QS.B.1..................................80
Figure 4.2 - Hysteresis Loops and Crack Pattern for Wall QS.B.2..................................82
Figure 4.3 - Hysteresis Loops and Crack Pattern for Wall QS.B.3..................................84
Figure 4.5 - Hysteresis Loops and Crack Pattern for Wall QS.S.1 ..................................85
Figure 4.7 - Hysteresis Loops and Crack Pattern for Wall QS.B.5..................................87
Figure 4.8 - Hysteresis Loops and Crack Pattern for Wall QS.B.4..................................88
Figure 4.9 - Hysteresis Loops and Crack Pattern for Wall QS.B.6.L ..............................89
Figure 4.10 - Vertical Load Fluctuation for Cyclic Test Walls .......................................91
Figure 4.13 -Typical Cracking and De-bonding at Horizontal Bamboo Reinforcement ..95
xii
Figure 4.16 - Idealized Experimental Envelopes ..........................................................101
Figure 4.19 - Hysteresis Loops from Experimental Unreinforced Wall (Sadek &
Lissel 2008) .........................................................................................................107
Figure 4.22 - Vertical Load Comparison for QS.B.2 & QS.S.1 ....................................110
Figure 4.23 - Compression Strut Anchored By Bearing Plate and Strut (CSA 2004c)...117
Figure B.1 - Vertical Load Fluctuation for Walls QS.B.1, QS.B.2, QS.B.4 &
QS.B.6.L..............................................................................................................149
xiii
List of Symbols and Abbreviations
d` distance between wall edge and outermost wall vertical reinforcing steel
(mm)
xiv
Ediss hysteresis dissipation energy (N·mm)
xv
L length of bond test specimen; wall length (mm)
xvi
angle between compression strut and horizontal axis of wall
μ coefficient of friction
xvii
1
In developing regions, affordable, safe housing is not readily available. The process of
social progress, including the improvement of education, literacy, and health. However,
with the rise of rapid urbanization in developing regions, a handful of problems, such as
the development of high-density ‘slum’ areas inevitably occur. Housing in these areas
not only suffers from unsanitary conditions leading to the proliferation of disease, but is
also generally unsafe in terms of providing shelter from storms and seismic events.
Governments and individual municipalities in these regions are faced with a huge task to
find a balance between economics and safety for future housing projects.
It is predicted that, by the year 2020, almost three quarters of the world’s population
living in urban areas will be living in developing countries. This is a huge increase from
the estimated 50% in 1980 (United Nations 1996b). This leads to extraordinary stress on
the complete infrastructure of the urban areas in these countries. It is estimated that
during the 2020s, a total of 77 million people will be added to the urban areas of the
world, with the bulk (about 74 million people) being added in developing countries
(UNCHS 1999).
2
In 1996, the United Nations published a report in which the estimation from 1990 was
that:
Not only is there the need for safe, affordable, reliable housing for the influx of urban
migrants in developing countries, but also the need to provide safe living conditions for
the population that is already inhabiting the urban areas. It was estimated that as of 1993,
only 73% of all housing structures in developing regions were permanent structures, and
only 63% of all those structures were in compliance with their respective building
regulations (UNCHS 1998). Housing problems in developing regions have also been
exacerbated by choices made that were lacking in the use of the abundant unskilled
infrastructure, and more will be needed with the expected population explosion. As can
be seen from Table 1.1, estimates on the need for housing in these regions are in the order
of 35 million units annually from the year 2000 to 2010, and 39 million units annually
approximately 500 – 800 million people are living in so called ‘income-poverty’ urban
areas (UNCHS 1999). This term refers to the income of an individual as less than
US$1.00 per day. With poverty constantly increasing in urban areas, the challenge of
providing safe, reliable housing becomes even more difficult. The economic price to pay
for these developing regions is staggering. The housing that must be provided must be
as affordable as possible, so that the population in the so-called income-poverty zone can
On the twelfth of May, 2008, a 7.9 Mw magnitude earthquake struck the Chinese
province of Sichuan. The official death toll of this tragic event is upwards of 67,000
people, with over 20,000 people missing. Since the earthquake occurred, a total of over
five million people were left homeless (BBC News 2008a). From two aftershocks, more
than 420,000 homes were destroyed. Shown in Figure 1.1 is a photograph of some of the
damage done by this earthquake in the town of Wenyaun, located near the epicentre of
the earthquake. Casualties and destruction is unfortunately not an uncommon trend, with
earthquakes in Java, Iran, Pakistan, Indonesia, Algeria, Turkey and Afghanistan being
some of the most deadly and destructive over the last ten years (BBC News 2008b).
Data from Table 1.1 indicates that not only a large proportion of existing housing in
developing countries is unsafe, with many of these houses not even meeting building
5
codes, but also that a staggering proportion of new housing still needs to be constructed.
This need for housing goes beyond basic structures that keep out the elements, but
buildings that are able to withstand higher loads and deflections that are present in
seismic areas. In Figure 1.2, the distribution of seismic hazard around the world is
shown. A large portion of these hazards is present in developing countries, where much
of the population simply cannot afford the high cost of using conventional materials to
approximately 50% of the population live in earthen buildings, such as common adobe
Figure 1.2 - World Distributions of Moderate and High Seismic Hazard (De Sensi
2008)
6
Not only are conventional materials like steel and cement already expensive for these
countries, but the costs are likely to continue to rise. In 2008, the cost of raw steel in
Canada doubled from $60 to $120 per tonne, over a short period of only four months
(Banks 2008). However, part of this increase is due to consolidation within the Canadian
steel industry, and the closing of old and inefficient facilities, rather than material
availability. In China, it was the government’s mandate to provide clean air to athletes
for the 2008 Summer Olympics. Part of the solution for this situation was the closure of
steel manufacturing in the Olympic event areas, effectively shutting down their steel
supplies to the rest of the world. Price adjustments for steel from several manufacturers
have ranged from an increase of 8% all the way up to 20%, and while many industry
consultants feel the price will come back down once steel production is back online, they
earthquake and storm-proof housing. However, even at the low point of the fluctuating
cost of steel, its use can be unaffordable in developing nations. Therefore, research of
other effective, low cost materials becomes extremely important to provide safe housing.
This thesis focuses on the use of bamboo as a replacement for steel reinforcing in
masonry structures. It should be recognized that this is only a possible solution where
bamboo is a native material and that there may be many other solutions. Each region will
have its own unique culture and resources, both of which will influence the materials and
manner in which safe housing could be constructed. The housing must not only be safe,
7
but it must use technologies, aesthetics, and construction techniques that are embraced by
This thesis specifically deals with the topic of shear walls using bamboo reinforcement as
an alternative to steel reinforcement. A test program was set up in which seven concrete
block masonry walls were tested for in-plane shear strength, using cyclic loading to
represent earthquake loading. One control wall was built using steel reinforcement with
seismic zone. The remaining six walls were internally reinforced using the bamboo
native species of China known for its use in fly fishing rods.
This thesis is organized into five chapters. Chapter Two presents relevant work found in
the literature, providing a solid background to the research that was undertaken. The
review starts with previous low cost housing research, followed by a comprehensive
use as reinforcement in cement matrices and other construction projects. The chapter
then focuses on the use of masonry as a structural material in seismic areas, followed by a
review of shear walls and their role in resisting seismic loading. The chapter concludes
Chapter Three is focused on describing the experimental procedures that were used in the
laboratory testing. The chapter not only discusses the methodology for the cyclic testing
of the seven masonry walls, but the procedure and results for the individual material
properties of the masonry, steel and bamboo used in the walls. The results from the
cyclic testing are given in Chapter Four, along with an analysis and discussion. Finally,
2.1 Introduction
With the tremendous need for housing in developing countries, there have been many
previous development projects and research projects implemented. There are four main
concerns when implementing a low cost housing program: the political, economic, social
and technological aspects of the construction. All of these issues can affect the future of
low cost housing programs, and therefore technologies and planning need to be dynamic
as these aspects are not only different from nation to nation, but can change quickly in a
single nation.
Masonry is one of the most common materials for use in housing around the world.
Therefore, much research regarding the topic of masonry in seismic zones has been and
will continue to be carried out, as the world is constantly facing disasters such as the
related to safe, low cost housing, including the need for reinforcement in masonry in
seismic areas. The use of bamboo as a potential steel reinforcement replacement is also
discussed, followed by in-plane shear test methods and their analysis, which can be
Masonry is one of the most common building materials, and is used throughout the
world. Although constructing masonry is labour intensive, the materials are relatively
since it can be constructed by nearly anyone. In low cost housing applications, however,
the high cost of steel reinforcement results in it being left out in the construction. It has
long been reported and accepted in literature that the use of unreinforced masonry (URM)
construction in seismic areas (Bruneau 1994). With the absence of reinforcement, these
types of structures tend to exhibit a brittle, or non-ductile behaviour, and therefore are
often unable to accommodate large inelastic deformations. This can lead to sudden,
Different types of masonry, however, have better resistance to seismic events than others,
as shown in Figure 2.1. This figure illustrates the relatively new European Macroseismic
Scale (EMS), which has been suggested as a harmonization of seismic intensity scales.
As shown in this diagram, masonry such as rubble stone and adobe are some of the most
vulnerable types of URM masonry, while massive stone masonry has some of the best
resistance.
11
Due to the poor performance of URM in seismic areas in the past, modern day design
standards all contain provisions for minimum reinforcement in seismic areas. Sections
4.6 and 10.15.2 of CSA S304.1- 04 Design of Masonry Structures provide requirements
for minimum reinforcement to be used in masonry walls in seismic areas (CSA 2004b).
seismic hazard index, (SHI) which is defined by the National Building Code of Canada as
(NRCC 2005):
I E Fa S a = SHI (2.1)
acceleration based site coefficient, and finally Sa is the 5% damped spectral response
Even though most of Canada is at low risk for seismic activity, only when the product of
these three values is less than 0.35 can the requirement for minimum reinforcement be
waived.
Many of the developing nations around the world are in areas prone to seismic action,
and because of this, unreinforced masonry alone is not a solution towards safe, low-cost,
economically out of reach for these nations, and therefore it seems prudent to research a
variety of replacement materials in order to find a balance between the safest, and most
As early as the late 1970s, it was recognized that “The social cost of urbanization ha[s]
been grossly underestimated in the past, particularly when viewed in terms of changing
social values.” (Luytjes 1979) Even at that time, both economists and sociologists
technologies. The key to any housing project is the choice of a set of appropriate
technologies for the specific region. For example, while the structural soundness of
reinforced concrete is widely accepted, producing and using these conventional materials
requires skilled labour, along with educated supervision, and the cost of producing or
Appropriate technologies are therefore a product of the region that they are used in. In
13
the case of developing regions, it has been recognized that “it is intuitively obvious that
housing production should evolve into a labour intensive industry with products which
meet the effective demand for housing, using indigenous building materials.”
(Moavenzadeh 1979) Technologies for producing building materials in this case should
have high labour intensity with limited specialized machinery, be locally manufactured
and have a minimum of engineering input, once past the initial research phase. While
this is not always possible, this has to be the ultimate goal of all low cost housing projects
and research. Schumacher (1973) also identified the need for simple, yet effective
technology for use in local and small-scale production settings. Taking advantage of the
abundance of local labour and resources can also stimulate the local economy.
Different structural materials that are fraction of the cost of conventional materials have
been explored in past research. For example, different materials have been investigated
as cement replacement for mortar, including rice husk, rice ash, lime, gypsum fly ash,
and furnace slag (Ali 1979, Kumar 2002, Cook et al. 1977, Savastano et al. 2003).
Earthen block is used widely in the world, but can suffer from low durability, shrinkage
and low strength. It has been proposed to stabilize earthen block with cement to increase
the strength, (Ali 1979, Williams 1979) or use fibres such as barley straw to reinforce
provide satisfactory resistance to seismic events, and some form of strengthening, such as
internal reinforcement, is usually required. Confinement of masonry has also been shown
14
to increase seismic resistance (Blondet 2005, Brzev 2007). This is the case for masonry
infilled concrete frames where a traditional steel reinforced concrete frame structure is
With regard to internal reinforcement, investigations have been carried out using sugar
protection (Blondet and Villa Garcia 2004) and a guide was created to aid the populace in
constructing an earthquake resistant homes (Blondet et al. 2003). Other work showed that
having it externally attached to the structure, as shown in Figure 2.2. (Dowling et al.
2005).
Results showed significant cracking early on in the seismic simulation that specimens
without internal bamboo reinforcement did not exhibit. It was thought that the flexible
15
bamboo compared to the stiff adobe caused a discontinuity in the structure, leading to
these cracks. In contrast, the externally mounted bamboo reinforcement, along with a
timber ring beam at the top of the specimens performed well under these testing
conditions.
As can be seen, there are many different types of technologies and materials that have
been employed for use in low cost housing projects. Bamboo is but one of the materials
that can be used in place of steel reinforcement, yet it holds great potential, and therefore
To date, there are approximately 1000 known species of bamboo around the world.
Antarctica, most of Australia, Canada, Europe, North Africa & Western Asia are the only
places in the world where bamboo is not natively found. Bamboo is a plant that is
already extensively used as a food source, garden plant and building material, including
Some examples where bamboo has been used in housing construction include the use of
Guadua bamboo as a roofing material in Columbia, (Hadjri 2005) and plastered bamboo
panels, (Kalita et al. 1978) and building trusses have been constructed and tested, as a
replacement for timber (Abang and Aband 1983). Bamboo has also been the subject of
16
research dealing with soil stabilization (Datye et al. 1978, Aziz and Ramaswamy 1981,
Mehra et al. 1951). A huge cost savings compared to traditional materials is possible.
One study dealing with bamboo based ferrocement roofing elements found a savings of
30 to 50 percent compared to the use of steel and concrete (Venkateshwarlu & Raj 1989).
Research on the use of bamboo in masonry applications is rather limited relative to its use
in concrete. However, the interaction between bamboo and grout should be similar to the
interaction that it has with concrete since the main difference is only the aggregate size.
With smaller aggregate, the bond strength could even be improved in grout compared to
concrete.
One of the first major studies on the use of bamboo in a cement matrix came as early as
1914 (Chu 1914). With the short supply of steel during World War II, bamboo as well as
many other materials were researched for their construction potential. In this wartime
period, “…every available material was being investigated for its possible uses as a
Clemson Agricultural College of South Carolina produced some of the first major
research in the United States that looked at the viability of bamboo reinforcement in
Portland Cement Concrete, contracted by the United States War Production Board. The
(a) “To establish reliable design principles for load carrying members
constructed of bamboo reinforced concrete.
(b) To establish and recommend approved construction procedures.
(c) To establish the practicability of using bamboo as reinforcement in
concrete.
(d) To compare the cost of construction of structures of bamboo
reinforced concrete with those of standard construction,
(e) To establish the physical properties of bamboo.”
(Glenn 1950)
This 171 page bulletin became a foundation on which much of the future bamboo
research was based, with many other documents summarizing the research completed
(Anon 1951).
Bamboo is a wood-like plant that is part of the grass family, consisting of a cylindrical
hollow shoot, or culm. This culm is covered with a waxy surface, which prevents
moisture from escaping. At intervals, the culm has raised ridges called nodes, from
which branches will offshoot. The plant grows up from a throng of underground stems
and roots, called ‘rhizomes’. Some species can grow to a height of up to 30.5 meters,
with a diameter as great as 305 mm. An interesting property of bamboo is that the
diameter of the shoot that grows out of the ground is the greatest diameter it will ever
grow to. As one of the world’s true “rapid” renewable resources, bamboo can have an
extremely high growth rate, with some species growing up to 600 mm per day.
Unfortunately, despite this high speed of growth, it still takes four to five years for the
bast fibres, or so called “wood” fibres to mature (Glenn 1950). The general physical
Much of the research on the use of bamboo has been to use it as a replacement for
conventional materials, such as steel and timber. Abang & Aband (1983) summarized
Table 2.1. Bamboo, much like timber, has a high variation in its physical properties from
species to species. Different species, along with different seasoning lengths can
drastically affect the tensile properties of the material (Abang & Aband 1983, Glenn
Table 2.1 - Bamboo vs Other Construction Material Properties (Abang & Aband
1983)
Material Specific Ultimate Tensile Ultimate t/c Modulus of
Gravity Strength, t Compressive Elasticity, E
(N/mm2) Strength, c (N/mm2) (kN/mm2)
Mild 7.8 480 - 1.0 210
Steel
Concrete 2.4 - 25 - 55 0.1 10 – 17
Timber 0.4 – 0.8 20 - 110 50 – 100 1.1 8 – 13
Bamboo 0.8 – 1.4 180 - 440 38 - 65 4.8 – 7.1 7 - 20
19
In tensile testing of bamboo, two methods are usually used: 1) straight tensile test with
the bamboo placed directly into grips, and 2) casting concrete on the ends of the
specimens and applying grips to the concrete, so as not to crush the specimen in the grips.
The strength of any species of bamboo is said to increase drastically when it is seasoned
found that green bamboo possessed only 60% of the tensile, and between 30-35% of the
compressive strength compared to seasoned bamboo, which has a much lower moisture
content (Youssef 1976). Kankam adds to this in a later paper stating that “it is generally
agreed that bamboo obtains its greatest strength at an age of about 3 years, that the node
is weaker in tension than the internode and that seasoned bamboo has a higher tensile
strength than unseasoned bamboo.” (Kankam et al. 1986a) This research showed an
even greater variability in the tensile strength of bamboo compared with Table 2.1, being
between 63 - 392 MPa, with the modulus of elasticity being as high as 30 kN/mm2.
In addition to research involving only the tensile properties of various bamboo species, a
small number of studies have also investigated the compressive and bending stresses of
bamboo. Shui (1990) tested three different species of bamboo, gigantochloa levis,
gendrocalamus asper and gambusa blumeana, finding tensile strengths for the specimens
between 242 - 269.2 MPa, with a modulus of elasticity of between 12.3 – 15.3 kN/mm2.
In compressive tests, the strengths of the three species ranged between 45.5 – 58.1 MPa.
specimens of bamboo. Tensile strength ranged from 48.0 – 170.6 MPa, while
compressive strengths of the species ranged from 20.3 – 27.8 MPa. The flexural
20
strengths were found to be from 41.8 – 141.3 MPa. A typical stress strain diagram for
both tension and compression is shown in Figure 2.4. Research has also shown that
bamboo has a low shear strength, being only about eight percent of its compressive
strength (Abang & Aband 1983). This is relatively weak compared with other structural
strength.
Creep and fatigue of bamboo have not been well researched. It has been purported that
bamboo acts much like wood in a time dependant manner. In the long-term, wood can
only carry about 60% of its short-term strength, therefore due to the limited research it is
recommended for design purposes that only 50% of the strength of bamboo be taken
(Subrahmanyam 1984). Cox and Geymayer (1969) reported a creep coefficient (ratio of
creep strain to elastic strain) of 0.4 for the two stress levels that they considered, 27.6 and
21
56.9 MPa. They also concluded that changes in temperature and humidity would induce
loading, or the creep of specimens in compression or bending has been carried out.
As shown in the previous section, the tensile strength of bamboo can be as high as
392 MPa, equivalent to the strength of some types of steel. With such relatively high
strength, the bond to the cement matrix becomes increasingly important so that the tensile
Like timber, the engineering properties of bamboo are highly sensitive to different
Research has shown that bamboo may absorb up to 100% of its dry weight in water,
(Mehra et al. 1951). The moisture absorption tends to be quite high initially, which
causes the material to swell until it reaches its fibre saturation point. The horizontal
expansion ranges from 2 – 5 %, and the longitudinal expansion is around 0.05%. After
When bamboo is placed in fresh concrete, the bamboo in its wet environment will absorb
water, and start to swell. The strength of the concrete over the first few days of curing is
not strong enough to prevent the bamboo from swelling, hence, the expanding material
will crack the concrete. Once the available water in the cement matrix is used up, the
22
bamboo will again adjust its moisture content to the drier environment, shrinking in
volume during the process. This leaves a void in the cement matrix larger than the
volume of bamboo that is present, leaving very little physical bonding of the bamboo to
The use of seasoned culms as opposed to green culms will reduce the severity of cracking
and shrinkage, while the use of pre-soaked culms will negate any cracking, but will have
a loss of bond when the bamboo dries in the cement matrix (Subrahmanyam 1984). In
saturated with a waterproof material. Materials that have been used in the past include
asphalt emulsion (Glenn 1950), bitumen coatings (Kankam & Perry 1989, Narayana et al.
1962) Negrolin and metallic wire (Ghavami 1995, 2005), sulphur (Fang & Fey 1978),
anti-termite protective coating (Kumar & Prassad 2005) and varnish (Youssef 1976)
23
among others. Each material has produced varying results, with treatments such as
waterproofing materials can be a concern, with Ghavami (2005) stating that while the
treatment of bamboo with epoxy and fine sand was effective, it was an expensive
treatment. He suggested that cheaper treatments such as asphalt paints, tar based paints,
and bituminous materials would meet the requirements for making bamboo impermeable.
Subrahaman (1984) summarized studies by Youssef (1976) and Narayana & Rehman
With the variation of different waterproofing agents as well as different node placements
in the cement matrix, bond test results from previous research have had a high degree of
variance. Kankam & Perry (1989) tested 42 samples ranging from samples with only a
roughened untreated surface, to specimens coated with bitumen and dusted with sand.
24
Tested bond stress ranged from as low as 0.33 N/mm2 for untreated samples to as high as
2.6 N/mm2 for samples treated with bitumen and dusted with sand. The low bond strength
for untreated bamboo has been confirmed by others, with Subrahmanyam (1984) calling
it “unreliable” with stresses ranging from 0.29 N/mm2 to 1.18 N/mm2. Using an anti-
termite coating and then protective coating ‘Top Coat’, the bond strength of bamboo was
found to be somewhat low, with a range of 0.42 to 0.81 N/mm2 (Kumar & Prassad 2005).
Ghavami (2005), however, argues that the pull-out stress from these type of tests give a
non-uniform shear stress, and made an improvement on the pull out test to give an evenly
Figure 2.7 - Bamboo Pull Out Tests: a) Improved; b) Conventional (Ghavami 2005)
With this method, bond strengths were found to range from 0.52 to 0.97 N/mm2
(Ghavami 1995). The bonding shear stress was calculated from the following equation:
F
= (2.2)
L S
25
In this equation, is the bonding shear stress, F is the applied pulling load, S is the
Since bamboo does have a relatively low bond strength, research has also been conducted
on shear transfer mechanisms. Work has been done on simple integral connectors, in
which notches are cut in intervals along the length of the bamboo, as well as the use of
trapezoidal steel sheets, embedded into saw cuts along the bamboo (Narayana & Rehman
1962). While the steel did not perform well, the integral connectors obtained an
impressive maximum shear of 6.4 N/mm2. The drawback with this method was that a
large amount of labour was needed to cut all the notches, and the effective area of
bamboo that would be used as reinforcement decreased significantly. Bulb and wedge
type shear connectors have also been proposed for concretes with a low modulus of
elasticity (Datye et al. 1978, Datye 1976). The research showed that up to half the tensile
strength of bamboo can be taken up by the bulb type connectors. In Figure 2.8,
Subrahmanyam summarises the different types of shear connectors that have been used.
Limited testing on the bond between bamboo and mortar has been carried out. One study
reported a maximum bond stress of just over 1 MPa for a sample embedded in mortared
brick (Iyer 2002). Samples that were embedded in mortar alone produced a bond strength
of just over 0.8 MPa. Unfortunately, the samples with mortar alone were cured for only
36 hours, and samples with brick were only cured for 7 days. This combined with a lack
of information about the preparation of the bamboo leads to somewhat unreliable data.
While most of the engineering properties of bamboo combined with its low cost and rapid
renewability seem very attractive, there are some drawbacks that are likely to prevent its
reported by different sources, (Glenn 1950, Kurian & Kalam 1977) limits the use of
bamboo. As discussed above, swelling of the bamboo can lead to a cracked cement
matrix, as well as loss of bond. Since the modulus of elasticity of bamboo is almost the
same as that of concrete, from a theoretical point of view bamboo could never prevent or
reduce initial cracking in flexure. This has been confirmed in many flexural testing
programmes (Glenn 1950, Ghavami 1995, Kankam et al. 1986b). The load deflection
beam after the steel has entered the plastic phase. A general view of the difference in
load-deflection curves is shown below in Figure 2.9, where the more heavily reinforced
bamboo section has a much greater deflection for the applied load. These factors may
limit the use of bamboo as reinforcing in high capacity structures, yet the advantages
bamboo provides over unreinforced members in addition to the low cost makes it a highly
27
desirable material for use in low cost housing applications. Aziz & Ramaswamy (1981)
stated that “bamboo technology for appropriate applications in various low cost
construction works can definitely lead to the full utilization of bamboo resources of many
developing countries.”
Figure 2.9 - Comparison of the Deformation Behaviour of Beams with Steel and
Bamboo Reinforcements (Subrahmanyam 1984) after (Youssef 1976)
Much of the research reported in literature has focussed on comparing bamboo reinforced
total of 32 rectangular and T-beams. The bamboo reinforcing used in the beams included
study the effect of different treatments on beam behaviour. The percentage of bamboo
reinforcing for all test beams was between 1.5 – 6%. Results showed that the addition of
bamboo reinforcing added up to four to five times the ultimate capacity of the sections
28
percentage of three to four percent. As the reinforcement percentage grew, the ultimate
strength would increase, until the optimum percentage was reached. After this, no
Ghavami (1995, 2005) also found results similar to those of Glenn. Using bamboo
reinforcement coated with Negrolin and wire, flexural specimens were shown to have an
also agreed that the optimal reinforcement level, was about three percent. In contrast to
Glenn and Ghavami’s results, Kumar & Prassad (2005), as well as Akejua & Faladea
(2001), found much lower values of improvement in bamboo reinforced members versus
unreinforced members. Capacities of only 2.4 and 1.34 times the values of unreinforced
specimens were found respectively. Kumar & Prassad used an anti-termite coating and a
protective coating (TOP COAT), while Akejua & Faladea used plain seasoned bamboo,
compared to the bitumen and Negrolin-wire used in the other test programmes.
These differences in flexural strength found in literature can be traced to the high
variability of bamboo treatment for use in cement matrices. Glenn (1950) found in his T-
beams that with the optimum reinforcement ratio, the bamboo reinforcement was capable
of carrying a tensile stress of up to 55 - 69 MPa, which was much less than the tested
tensile strength of the bamboo used. The bond seemed to be the limiting factor in
developing the tensile strength of the bamboo. The differences in bond not only limit the
tension that develops in the bamboo, but also the overall structural response. Figure 2.10
29
shows work done by Cox and Geymayer (1969), in which the load responses of different
bamboo reinforced beams are shown. While Beam A has a lower percentage of
reinforcement compared to the other two beams, the epoxy sand coating gives a better
bond, and provides a stiffer element. This translates to a much higher failure load than
the other beams. Brink & Rush (1966) also suggested that “experience has shown that
split bamboo performs better than whole culms when used as reinforcing.”
There has been much less research on the shear behaviour of bamboo reinforced concrete
members than on flexural capacity. Glenn (1950) summarized failure from his
rectangular-beam testing program, stating that most of the specimens failed by diagonal
tension. The research found that bending up the upper rows of longitudinal
reinforcement proved to be beneficial to the ultimate load. Brink and Rush (1966) also
longitudinal reinforcement should be bent up, stating in one design example that “the top
two rows should be bent up randomly in the outer one-third sections of the beams to
assist the vertical stirrups in resisting diagonal tension.” They also state that bamboo
may be bent either wet or dry if the bamboo is heated while constant pressure is applied.
With this property, bamboo is suitable for c-shaped stirrups, and hooks for dowels.
Several investigations have also been carried out on one and two way slabs. These
sources suggest the viability of using bamboo as reinforcement for slabs, with similar
results to those found for flexural beams (Kankam, et al. 1986b, Kowalski 1974, Shui
1990, Ghavami 2005, Narayana & Rehman 1962, Singh & Jain 1977). Kankam (1986b)
tested several two way slabs loaded with a concentrated force at the center, and found the
failures to be generally from a combination of both moment and shear. There were no
instances of sudden shear failure, which led to the recommendation that bamboo could
safely be used as a replacement for steel. As with the beam testing, however, the
deflections of bamboo-reinforced slabs were greater than those of steel reinforced slabs.
Therefore, it was recommended that the span to total depth ratio not exceed 23. It was
conversely shown in another study that the use of integral shear connectors can be
employed to improve the load capacity and stiffness of one way slabs (Narayana &
Rehman 1962).
While the compressive strength of bamboo is greater than that of normal strength
concrete, the elastic modulus of bamboo is roughly the same or even lower than that of
concrete. Taking this into account, the ultimate axial load resistance of a concrete
31
column reinforced with bamboo should theoretically be less that that of an unreinforced
column. Brink and Rush (1966) state that, “bamboo reinforcement in columns serves to
resist a compression load equal to that taken by the concrete.” However, it was shown to
be slightly less than equal strength in laboratory testing, as illustrated in Figure 2.11
(Cook et al. 1978). In this figure, it can be seen that while the ultimate axial load of the
performs better.
Cook also compared the interaction diagrams of bamboo and steel reinforced columns as
shown in Figure 2.12. From these diagrams, it is shown that for two columns with an
to 40 percent of the ultimate load of the steel reinforced column. Ghavami (2005) also
concluded that a 3% reinforcement ratio of bamboo treated with Sikadur 32-Gel would be
32
construct several field buildings using bamboo reinforced concrete. One of the buildings
built in 1944 was the press box at the Memorial Stadium, in Clemson, South Carolina: a
18.2 x 4.0 m three storey structure, including slabs, rectangular and T beams, load
bearing walls, and columns. Bamboo was used as reinforcement in all elements.
Approximately five years after construction, a detailed inspection was carried out
revealing several cracks that had appeared in the structure at various locations, however
the majority of these cracks were attributed to shrinkage of the concrete. Cracks also
33
appeared in the floor slabs on the second and third floors, however, since they were
parallel to the bamboo placed, the safety of the structure had not diminished.
Observation of the structure five years later showed that no new cracks had formed, and
the previously present cracking had not increased in size. Deflections of the structure
were also minimal. With these observations, it was recommended that more temperature
reinforcement be placed in these types of structures in the future. Some fifteen years after
construction when the structure was demolished, it was said to still be in good shape
(Subrahmanyam 1984). In contrast, a planar shop building and a five room residence that
were constructed in this project, exhibited major distress in the roof beams, causing the
condition after spending 15 years in the open atmosphere at the PUC-Rio University in
Brazil. The tensile strength of the aged bamboo sample was tested, showing only a slight
deterioration in tensile capacity. In India, a bamboo reinforced concrete roof slab was
constructed in 1962 using integral shear connectors. Initially, a number of cracks formed
in the structure, but were sealed twelve years later with mortar. At that time, the
A few attempts have been made to develop guidelines for constructing bamboo-
reinforced structures. With the experience from the experimental buildings, Glenn
developed some construction principles to follow (Glenn 1950). Brink and Rush (1996)
developed a more complete design guide providing suggested values for material
construction details, such as splicing. More recent design examples and equations that
can aid designers in a bamboo reinforced construction project have also been reported
Some of the typical types of damage that occur in masonry structures subjected to seismic
forces include in-plane shear failure, out-of-plane bending, or in-plane flexure, all of
which are shown in Figure 2.13. In addition, other failure mechanisms, such as lack of
anchorage / anchor failure, diaphragm related failure, and pounding between adjacent
In Figure 2.13, the walls that are parallel to the seismic action serve to resist the lateral
load induced in the structure since they are much stiffer than those perpendicular to the
load. These shear walls are particularly important structural members in resisting seismic
The research presented in the previous section showed that bamboo performs well in
cement matrices, and because of the similarities between reinforced grouted masonry and
concrete, similar results are expected for load resisting masonry members. However,
there are many factors that affect the response of a masonry structure in the case of a
seismic event including the type and quality of units and mortar, the presence of
reinforcement, and the layout (including irregular plans and openings) and size of the
using a less robust masonry unit or type of reinforcement can potentially change the
overall seismic response of the building (Tomazevic et al. 2004, Tomazevic et al. 2006).
In light of the importance of shear walls in resisting a seismic event, this thesis
There are three main types of failure that occur in shear walls, as shown in Figure 2.14. In
the 2004 CSA S304.1 standard, equations are provided for each type of failure mode
based on empirical representations of observed shear behaviour and are for the ultimate
limit state (CSA 2004b). They are therefore not applicable to elastic behaviour of walls
(Drysdale & Hamid 2005). Shrive and Page (2008) also note that “methods of
36
determining shear strength of masonry elements also produce widely varying capacities.
The variation reflects the lack of confidence that any single method is correct in modeling
For proper design of shear walls, designers should actually try to achieve the more ductile
flexural type of failure, with the vertical reinforcement yielding before the more brittle
diagonal or sliding failures occur (NRCC 2005, Drysdale and Hamid 2005). For this
failure mode, the in-plane response of a shear wall can be treated much like a beam in
2.15.
37
Diagonal shear, or tension, is the mode of failure in which cracking occurs in the wall in
a diagonal direction, or in the case of two-way loading, an “X” pattern. The diagonal
shear resistance of masonry walls is something that has been heavily investigated, with
several different models formulated to quantify the resistance to loading (CSA 2004b,
Anderson and Priestley 1992, NEHRP 1997, Voon and Ingham 2007) .
Many of these shear models or equations do not take into account the dowel action of
vertical reinforcement. Tomazevic (1999) explains how the shear resistance of masonry
walls depends on many different mechanisms, including truss and arching action of the
38
masonry, tension in the vertical reinforcement, as well as the dowel action of the vertical
and horizontal reinforcement in the wall. Figure 2.16 shows the basic mechanism of
diagonal failure of a shear wall reinforced with both vertical and horizontal
reinforcement.
Priestley and Bridgeman (1974). The amount of shear that can be carried due to this
mechanism by each dowel was determined by Tomazevic (1999), based on the previous
In this equation, drv is the diameter of the vertical reinforcing bar, fm is the compressive
strength of the embedding mortar or grout and fy is the yield strength of the reinforcing
steel. The mechanism on which Equation 2.3 is based is shown below in Figure 2.17.
Experimental research and actual earthquake damage assessments have shown that walls
having only vertical reinforcement will fail in diagonal shear as opposed to flexure
(Tomazevic 1999) even with the additional dowel action. Despite this, improvements to
the performance of shear walls have been noted with uniformly distributed reinforcement
including better resistance to sliding, more effective control of diagonal cracking and a
higher compression splitting resistance at the toe of the wall (Drysdale and Hamid 2005).
The complexity of the dowel action mechanism, combined with few attempts to
The last of the three failure modes, sliding shear, occurs when part of the wall slides
along a horizontal plane, usually along a mortar joint. This mode of failure is more
common when shear walls are built with low strength mortar, have low aspect ratio, or
low levels of vertical load. Both diagonal tension and sliding shear failures are usually
considered brittle failure modes (Hatzinikolas and Korany 2005), as both usually have a
sudden decrease in load capacity with limited deformation capacity after the ultimate load
has been achieved. Since the three modes of failure occur for varying deformations and
loads, it is important that any wall be designed taking all of these failure modes into
account.
In Clause 10.16.3.3 of the Canadian Standard S.304.1-04 it is stated that for seismic
design of ductile walls that “walls shall be designed to resist a shear force not less than
the shear that is present when the wall develops a plastic hinge mechanism.” Since the
three types of failure modes exhibit varying degrees of ductility, with flexure usually
being the highest, and sliding shear being the lowest, in order to ensure that the most
Anderson and Brzev (2008) stated that the objective of the capacity design approach “is
to force the structure to yield in a ductile manner without failing at the expected
deformations and to design the rest of the structure to be strong enough not to fail when
the system forms a ductile mechanism.” For the case of shear walls, the practice has been
41
to design the wall such that the diagonal shear strength is greater than the flexural
strength of the wall for any given lateral loading. If this condition is met, the wall will
experience a large plastic deformation while still being able to resist large loads. If the
opposite is true, the wall will never be able to resist loading in which the more ductile
flexural failure would occur, and will instead fail in the more brittle diagonal shear
manner.
Construction of low cost housing introduces additional factors that may influence the
ductility compared to traditional construction. For example, in order to achieve the more
ductile flexural failure in conventional construction, steel dowels will be placed in the
wall, and embedded in a reinforced concrete slab or foundation, allowing them to reach
yield, and the wall to deform when lateral loads are applied. For low cost housing, there
may not be a reinforced concrete foundation to dowel into, steel dowels may be too
expensive, or the use of bamboo dowels may provide different results than steel dowels
do. Without any anchorage to the slab or foundation, the most ductile mode of failure
may not be flexural failure. The wall may be most ductile in diagonal shear, or perhaps
the wall is most ductile when it starts “rocking” in flexure due to the lack of anchorage.
This is only one of the special conditions that can affect the ductility of a low cost
housing project and it is up to the designer to assess the unique conditions that may occur
and failure in masonry (Shrive and Page 2008) and there is also no accepted standard in-
plane shear test method. One common method of evaluating the in-plane shear resistance
when it is fixed to a strong floor. In many cases, the masonry wall is doweled into a
concrete foundation beam, and that foundation in turn is fixed to the floor to prevent
movement of the base. A concrete beam can also be used at the top of the wall for
applying both vertical, and horizontal loads into the masonry wall. This type of testing is
The racking or in-plane load can be applied to the wall in several patterns, including
Figure 2.19 shows some of the typical time-history displacement patterns that are applied
43
to masonry walls. It has been noted in several sources (Tomazevic 1999, Benedetti, and
Castellani 1980, Tomazevic et al. 1996) that the response of masonry structures can be
Figure 2.19 - Typical Lateral Displacement Time Histories (Tomazevic and Lutman
1996)
hysteresis loop, which is a plot of the wall deflection versus the applied load that the wall
undergoes during testing. Typical hysteresis loops from testing on a masonry wall are
shown in Figure 2.20 (a), and the difference that is obtained in the hysteresis envelopes
The difference in ultimate load, as well as ultimate deformation is quite apparent from the
higher load and deflection. This illustrates the importance of using some type of cyclic
testing for masonry to assess its properties when the intended use is for seismic areas.
44
Boundary conditions in experimental tests can also significantly influence test results,
therefore it is important that the proper boundary conditions are chosen. For example,
testing has shown that an increase in axial load can increase the shear resistance of a wall
(Voon and Ingham 2006). While others have also reported this increase, Fehling and
Stuerz (2008) also showed in testing on calcium silicate and lightweight concrete units
that while the shear capacity increased, the deformation capacity of the walls decreased.
Other boundary conditions were discussed by Fehling and Schermer (2008), who showed
that slabs in multi-storey structures can have a significant restraining effect on shear
walls. Work by Magenes and Morandi (2008) tested two boundary conditions for the top
of the wall; a “cantilever” system in which free rotation at the top was allowed, and a
“double fixed” system, in which rotation was restrained. Results for tests on calcium
silicate masonry shear walls showed a decrease in both load and deflection for the walls
Previous testing has also examined the response of masonry walls with different
parameters, subject to the same boundary conditions. Several sources have reported a
large increase in shear resistance for lower aspect ratios (Voon and Ingham 2006,
Magenes and Morandi 2008, Magenes and Calvi 1997). This can, however, also lead to a
change in the mode of failure. Studies on unreinforced masonry walls have shown that a
rocking failure can provide a stable, non-linear response, and significant lateral
agreed that the addition of horizontal reinforcement in masonry walls has a significant
effect, enhancing the shear capacity, post cracking performance and ductility (Drysdale
and Hamid 2005, Voon and Ingham 2006). Testing conducted at the University of
Colorado showed that walls with adequate horizontal and vertical reinforcement had
ductile performance after the yielding of the flexural reinforcement (Shing et al. 1990a,
1990b, 1991). Walls that failed in diagonal shear and did not have yielding of the
flexural reinforcement had limited ductility. This study also showed that the maximum
horizontal reinforcement in a masonry wall is not a fixed quantity but rather, depends on
the wall geometry, compressive strength, and other attributes such as lateral confinement
testing of shear walls is the bilinear idealization of the hysteresis envelope (Magenes and
Morandi 2008). These models are generally based on simplified elastic theory, but are
modified by taking observations and results found from experimental testing and
46
incorporating them into the equations. There are also more complex tri-linear models
available, but the use of bilinear models make the idealization of hysteresis envelopes and
the respective calculations fairly simple. Figure 2.21 shows two examples of bilinear
idealization.
As shown, the two examples of bilinear idealization are nearly identical. Tomazevic
(1999) explains that in obtaining an idealized envelope, during actual testing, load and
deflections at three main epochs must be observed, the first being the initial cracking, dcr.
Note that in Figure 2.21 (b), this initial cracking is also defined as 0.7Hmax. The lateral
resistance at this point is denoted as Hcr. At this point the slope of the envelope should
start to change. The second epoch is the maximum lateral resistance of the wall, denoted
as Hmax, The final epoch is the maximum displacement or the ultimate state. The ultimate
deflection, du is defined as the point at which the line of the 0.8Hmax intersects the
47
experimental curve. Voon and Ingham (2006) also defined this point of wall failure in
their testing.
The initial slope of the curve, Ke is the effective stiffness of the wall, and is the ratio
between the wall’s resistance and the deformation at first cracking, given by:
H cr
Ke = (2.4)
dcr
In a hysteresis idealization, the area under both the experimental envelope, Aenv, and the
idealized curve should be equal each other. To accomplish this, the ultimate idealized
resistance, Hu, has to be evaluated taking into account the energy dissipation of real and
idealized walls. If the value of the initial or effective stiffness of the wall, Ke,, is known,
2A
2
H u = K e dmax dmax env (2.5)
Ke
Testing done on more than sixty masonry walls showed an average value for Hu of
The ultimate ductility factor, μu, of a wall can be defined by the equation:
du
μu = (2.6)
de
where du is the ultimate deflection at 0.8Hmax,, and de is the deflection at the end of the
elastic stage of the bilinear idealization. This last term can be evaluated by dividing the
Tomazevic (1999) notes that while it is quite possible for experimental testing to give
high values of ductility, he recommends that for seismic resistance verification the values
are limited to 2.0-3.0 for unreinforced, 3.0-4.0 for confined, and 4.0-5.0 for reinforced
individual masonry walls, such that excessive damage to structural walls is avoided.
From experimental hysteresis loops, it is possible to calculate both the cumulative energy
input, Einp, as well as the dissipated energy, Ediss. The input energy for one cycle, Einp is
defined as the sum of the shaded areas shown in Figure 2.22 (a). The dissipated
hysteresis energy for one cycle is defined as the area encompassed by the hysteresis loop
Figure 2.22 - (a) Evaluation of Energy Input in One Loading Cycle, (b) Evaluation
of Dissipated Hysteretic Energy in One Loading Cycle (Tomazevic et al. 1996)
both strength and stiffness. Previous testing has shown that the measured lateral force for
the same deformation of walls that have undergone cyclic loading compared to
monotonic loading is significantly lower (Tomazevic et al. 1996). Figure 2.23 shows that
different types of masonry walls. The three curves in the figure represent data from
In Figure 2.23, the y-axis is the ratio of the actual wall stiffness, K, over the effective
wall stiffness, Ke. The x-axis is the ratio of the displacement of the wall, d, over the
displacement of the wall at the ultimate lateral load, dHmax. Given the similarities
between the unreinforced, reinforced and confined masonry curves, it is anticipated that
the curve for bamboo reinforced masonry walls would be similar, however no such data
2.7 Summary
With increasing steel prices, and the need for reinforcing of masonry structures in seismic
areas, bamboo, a less expensive and potentially more sustainable material, has great
potential to provide not only strength, but also ductility in seismic areas. While no
research has been performed on bamboo-reinforced masonry as of yet, there have been
numerous projects in which bamboo has successfully been used in concrete. In order to
51
compare bamboo to conventional materials in masonry, this thesis will examine the
behaviour of shear walls subjected to cyclic in-plane testing, which was also discussed in
this chapter. How bamboo or any other alternative material performs as a reinforcement
3.1 Introduction
A quasi-static cyclic testing procedure was developed to evaluate the benefits of using the
from China, it was chosen since it was available for purchase in Western Canada. Seven
walls were constructed and tested in an in-plane, cyclical manner, with the main variable
being the amount and arrangement of bamboo reinforcement. In addition to these main
tests, material properties for each component of the specimens were determined
experimentally. The arrangements of these test methods as well as the material property
Seven concrete block masonry walls were constructed for the purpose of in-plane quasi-
static shear testing, using several configurations. Each wall was constructed using
standard 200 mm (190 x 390 x 190 mm nominal dimensions) block seven courses high,
and four blocks in length. This gave overall wall dimensions of 1400 mm (high) by
1600 mm (long) and an aspect ratio (height divided by length) of 0.875. Each wall was
constructed by the same master mason, in order to maintain the quality of workmanship
for all specimens. Six of the walls used normal strength concrete block, while the
seventh wall used a low strength block. In all of the bond beams, a knock out block was
53
used. This type of block allowed the webs to be easily removed, such that horizontal
reinforcement could be placed inside, and the course could be fully grouted. Type N
specimen.
Configuration of the reinforcement and grout in the walls is shown in Figure 3.1. A wide
variety of configurations was used such that the benefits or drawbacks of each type of
wall could be compared. The hatched areas indicate grouted cores, and dashed lines
indicate reinforcement.
In all of the walls, grouting was required where vertical or horizontal reinforcement was
placed. For grouted masonry in Canada, the CSA A179 (CSA 2004a) has a required sand
to cement ratio of 3:1 by volume. Since the goal in low cost housing applications is to
54
reduce the amount of expensive materials such as cement, a ratio of 4:1 was used. A high
water to cement ratio of 1:1.62 was also used. Since normal strength concrete block was
used, the water cement ratio used in the grout could not be increased too significantly, as
this would create a large disparity between the relative strengths of the block and grout
potentially causing the wall to act less homogenously during testing, with a single
The high water cement ratio was used to promote two main factors: workability and
consolidation. With higher water content, the grout could be mixed and poured into the
masonry with greater ease. Furthermore, with relatively low viscosity, the need for
vibration or rodding of the grout to prevent honey combing and air voids was eliminated.
Nevertheless, using this grout did come with a drawback. When the grout was poured
into the walls, some grout would flow into adjacent cores through the small openings,
leading to material waste. However, the workability outweighed the small amount of
The control wall, QS.S.1, was reinforced using conventional steel reinforcement, with the
minimum amount of horizontal and vertical steel for seismic conditions as required by
the CSA S304.1-04 Design of Masonry Standard. (CSA 2004b) The remaining six walls
used a combination of vertical and horizontal whole bamboo culm reinforcement with an
average diameter of 25 mm. The bamboo reinforcement was coated with a spar (marine)
varnish to waterproof the bamboo reinforcement. As shown in Figure 2.6, varnish has
was purchased for under $15 CAD, and provided coating for all of the bamboo tested.
With a total length of 65.6 meters of bamboo used in the walls, and assuming all the
varnish was used, this gives an estimated cost of CAD$0.22 per meter of waterproofed
bamboo. In actuality, costs would be less than this given that firstly, not all of the varnish
was used and secondly, a retail price was paid for the varnish. After the varnish coating,
the bamboo was also dusted with sand to improve the bond with the grout. A summary
3.2.2.1 Masonry
Twelve masonry prisms were constructed to determine the strength of the masonry (f`m)
used in the wall specimens. For the normal strength block used in six walls specimens,
six prisms were constructed, of which three were grouted. These prisms were tested at
28 day strength of the mortar and grout. An equal number of prisms was constructed for
the low strength block wall and were tested the same day as the wall, QS.B.6, at 38 days.
The compressive properties of all masonry prisms were tested according to the CSA
Standard A165.1-04 (CSA 2004d). The results from these tests are shown in Table 3.2,
with the apparatus used for these tests shown in Figure 3.3.
Individual material testing was also conducted on the strength of the units for the
“normal” block, as shown in Table 3.3, as well as on the compressive strength of grout
cylinders from the grout used in the walls, shown in Table 3.4
The grout behaved as desired in that no honey-combing, or air voids were present in the
cores, even with no consolidation or vibration being used. However, it was found that
variation of more than 20% from the compressive testing performed on the grout alone,
as shown in Table 3.4. With the segregation present in the grout, it was easy for one
sample cylinder to accumulate more water than another, which is a likely cause for this
high variation in strength. The grouted prism compressive tests shown in Figure 3.2,
however, showed much less variation with COVs of 1.7 and 10.3%. This is likely due to
the concrete block being able to draw excess moisture from the grout, resulting in more
The steel reinforcement used in wall QS.S.1 was tested for tensile strength. Loading was
applied to the rebar using steel grips in a Tinius Olsen tensile testing apparatus. An
extensometer was used to measure the strain. The results of the tensile tests including the
Because of the hollow nature of the bamboo, it could not be tested for tension in the same
manner that the steel reinforcing was since the grips of the testing machine would crush
the ends of the bamboo samples. Therefore, the bamboo was grouted in 50 mm diameter
SikaGrout® 212HP, was used to ensure adequate bond to the bamboo and steel piping.
The ends of the bamboo test specimens were treated with varnish and sand in the same
way as in the full wall specimens. Notches were also cut on two sides of the bamboo
culm at mid-length, to ensure failure at that point. The preparation of these specimens is
Unfortunately, before these specimens failed in tension, they experienced pull out from
the bamboo to the grout, as well as bond failure between the grout and the steel pipe.
Therefore, an alternative test method was sought. Each end of the hollow bamboo
specimens was filled with an epoxy, so that the same grips used for steel could be used,
60
and the ends of the bamboo would not be crushed. Again, this test method failed to
provide accurate tensile properties, as the bamboo split at the grips, as shown in Figure
3.5.
It was found that if the bamboo was split into sections, and epoxy was applied to the ends
of the specimens to give a flat surface for the grips to take hold of, the specimens could
be tested for their actual tensile properties. The specimens used, along with a close up
view of the notches cut into the bamboo are shown in Figure 3.6.
These specimens were placed in steel grips, and tested in a Tinius Olsen tensile testing
A total of 10 specimens were tested, in which three had notches cut into the culm (T.C.1-
3), three had notches cut in the nodal area (T.N.1-3), and one was left without any
notches (T.N.A). The remaining three specimens (T.ME.1-3) had strain gauges attached
to them, such that stress-strain diagrams and the modulus of elasticity could be found as
discussed in Section 3.2.2.4. No notches were cut into these specimens. Table 3.6 shows
All the tensile specimens appeared to fail by a combination of both splitting, and tension,
with the splitting occurring perpendicular to the notched area at the centre of the
specimens. A photo of the typical failure that occurred on the bamboo splints can be seen
Due to the brittle nature of the splitting failure, the results obtained may actually
represent a lower bound for the tensile strength of this type of bamboo. The values
obtained appear to be in the lower range compared to results reported in literature. Some
of the lowest tensile strengths have been reported in literature were in the range of 48.0 –
170.6 MPa (Ghavami 1995), while strengths as high as 440 MPa were reported by Abang
and Aband (1983). Due to the notching of the specimens, and the micro-fracture that
would occur, it can be reasonably assumed that the values reported are lower than the
actual failure stress of Tonkin cane. It can still be seen, however, that the test results
presented here agree with literature in that the node tensile strength (84.7 MPa) is less
63
than the culm tensile strength (117.6 MPa). Unfortunately, none of the values reported in
literature are for the Tonkin cane species of bamboo, and therefore a direct comparison
cannot be made.
The strain gauges for the three samples T.ME.1 to T.ME.3 were all placed on the culm, at
the midpoint between nodes. The stiffness of the node material may be slightly different
than the culm material, however since the node material only accounts for a fraction of
the total length of bamboo, it was decided that the node surface would have a much lower
influence on the total modulus of elasticity than the culm portion. The stress-strain
relationships for the three specimens are plotted in Figure 3.9. As can be seen, an almost
perfect elastic linear behaviour is shown for all three specimens. Each specimen failed
by splitting near the end of the steel grips, and then pulling apart at that location in a
brittle manner.
The modulus of elasticity for each of these specimens was also calculated from the stress-
strain diagrams, and listed in Table 3.7. The stress-strain diagrams, shown in Figure 3.9
really highlight the sudden failure that the bamboo experienced during these tests, since
the behaviour is nearly perfectly linear, with no significant curvature or plastic behaviour
at the end of the test. While some stress-strain diagrams found in literature, such as that
shown in Figure 2.4, show somewhat similar behaviour of bamboo tensile specimens, a
small amount of curvature is observed compared to the straight lines that were found for
these Tonkin cane specimens. The values determined for the modulus of elasticity
ranged from 16.5 – 19.2 GPa with a mean of 17.6 MPa, corresponding well with values
reported in literature (Abang & Aband 1983, Shui 1990). For example Abang & Aband
(1983) report a range of approximately 7-20 GPa, putting the value for Tonkin cane at the
The bond strength of bamboo to grout was first tested in a manner similar to the first
attempt for testing the tensile strength. For every specimen, each end of the bamboo was
65
coated with a spar varnish, and then dusted with sand, as was done in the full-scale wall
specimens.
In the first bond test attempt, one end of each test specimen was embedded in 50 mm
diameter piping with high performance grout, in exactly the same manner as the first
tensile tests. On the other end, a 100 mm diameter steel pipe, with a depth of 100 mm
was used as a container for grout in which the bamboo was embedded 50 mm. The grout
used on this end had the same mix proportions as the grout used in the full scale wall
specimens. Six specimens were tested, of which three had a node embedded in the wall
grout end. The other three of these specimens had no node embedded in the wall grout
end, but the natural culm only. These specimens are shown in Figure 3.10.
These specimens were then placed in the same Tinius Olsen apparatus used in testing the
tensile properties of the bamboo. Unfortunately, in all specimens but one, failure
66
the high performance grout to the steel piping. For the one specimen that failed at the
desired end, a maximum average bond stress of 5.33 MPa was observed, which includes
the mechanical anchorage of a node embedded in the grout for that specimen. In the
remaining tests, the minimum average bond stress was 1.10 MPa, for the specimen that
The test arrangement was changed to incorporate the successful gripping of specimens
that had occurred in the tensile tests, and was also modeled on the research done by
Ghavami (2005). Bamboo splints were embedded a total of 165 mm in grouted cylinders,
with 50 mm on each end covered with wax paper to ensure that no bond would occur
between this surface and the grout. The 65 mm center portion of the embedded specimen
was treated with spar varnish, and dusted with sand. The intent of this method was to
produce a uniform shear bond stress, as opposed to a non-uniform stress, which was
discussed and shown in Figure 2.7. Four of the eight specimens shown in Figure 3.11
contained a node in the 65 mm treated center, while the other four were absent of a node.
On the opposite ends of all specimens, epoxy was used in the same manner as the
bamboo tensile testing, such that steel grips could pull the bamboo out of the grout
cylinder.
67
These eight specimens were all tested in the same Tinius Olsen tension testing apparatus
as the previous bamboo tensile tests. Figure 3.12 shows a specimen being tested, as well
as a tested specimen in which the bamboo splint has been pulled out of the grout cylinder.
One sample from the specimen group that had nodes embedded, as well as one from the
group that was absent of embedded nodes was tested at 7 days from the casting of the
grout, while all remaining specimens were tested at 28 days from the casting of the grout.
All results from this series of bond testing are given in Table 3.8. The bonding shear
The bond strength of varnish and sand treated Tonkin cane in grout was highly variable
with an extremely large coefficient of variation for the splints with no node embedded
(almost 110 percent). In contrast to this, when a node was embedded, the coefficient of
variation was only 10.6 percent. Therefore, it appears that the mechanical anchorage of
the node not only increases the bond strength, almost doubling the value, but also reduces
much of the variance of the bond strength of bamboo to grout. Given the high variability,
The results for the bond between Tonkin cane and grout are on the high end of the values
reported in the literature for bamboo specimens embedded in concrete. The average
tested value of 2.16 MPa for Tonkin cane with an embedded node is nearly on par with
specimens tested with a bitumen coating and dusted with sand by Kankam & Perry
(1989) which achieved bond strengths of up to 2.6 MPa. Unfortunately, while the
embedded bamboo specimens in Kankam and Perry’s work are similar to the Tonkin
cane bamboo, there are a few major differences that make a direct comparison difficult.
69
First, the effect of the small aggregate size in grout compared to concrete is unknown.
The second difference is that Kankam and Perry’s tests most likely produced non-
uniform stress, as described in Figure 2.7. While the values are a good start for
comparison, more tests would need to be completed with the same conditions to make a
direct comparison.
It is also difficult to translate the results from the Tonkin cane bond tests to the actual
behaviour of the cane in the wall specimens since only splints were used in the bond
tests, while in the test walls, full culms were used. In the original bond stress test setup,
an attempt to use full culms was made, however only one of the tests in this series had a
proper pull out. Obviously, one test value is not enough to indicate a true value of bond
stress, but the one value that was obtained was significantly higher than any of the tests
that used splints even though there was no embedded node. The results could also be
affected by the fact that the splints had the varnish-sand treatment applied on both the
outside and the inside of the culm. Further research is needed to verify the validity of the
With some modifications, the testing arrangement in which both ends were encased in
grout is most likely to show the true tensile strength and mechanical properties of
bamboo. With increased embedment length of the bamboo, the chances of splitting
would be decreased. Though the test methods used here are not ideal, they still provide
valuable data that could be used to improve the testing arrangements for tensile and bond
70
strength. Further tests of the tensile properties of Tonkin cane could then be carried out
As discussed in Chapter 2, many previous tests have included a concrete foundation beam
for cyclic testing. It was decided that since materials such as steel and concrete are too
expensive for the construction of walls for low cost housing, then the likelihood of
having a reinforced concrete foundation would be low. Therefore, this was left out to try
and simulate actual boundary conditions that may occur in a low cost housing project.
The initial testing frame for the quasi-static testing is shown in Figure 3.13. Additional
photos are shown in Appendix A. The masonry wall specimen was placed between two
upright W-section columns supporting a 500 kN actuator, which was used to apply a
vertical load to the top of the wall. The wall was held stationary laterally by two braces
bolted to the concrete floor, and a hand-pumped hydraulic jack was used to ensure that
the wall would not slip along the concrete floor. Spherical seats were placed at the base
of each side of the wall to ensure that there were no eccentric forces being transferred
The cyclic horizontal force was applied to the wall specimen by means of a 250 kN
actuator mounted to a braced W-section column, East of the wall specimen. This actuator
had a total stroke of 250 mm, allowing it to push or pull the wall to a maximum
deflection of 125 mm each way. At the top of each side of the wall (East and West), steel
plates were suspended, and tightened to each other through the use of four steel threaded
rods, in effect clamping the top of the wall. One of the steel plates was directly attached
to the actuator. The centre of the bearing area of these plates was located at the top bed
joint, which had the effect of lowering the aspect ratio of the walls. In determining the
aspect ratio, the height to the bottom edge of the plate (1100 mm) was used and the
Each test specimen was moved into the testing frame by means of a welded steel frame,
as shown in Figure 3.14. One side of the moving frame has a hinge that applies a
distributed clamping force on each side of the wall when it is lifted by crane. This
during transport, so walls could be built at any place in the lab, and transported to the
A laser level was used to align the wall with the horizontal and vertical actuators. Plaster
of Paris was also placed on the floor to ensure that the wall was level and that no stress
concentrations would occur from any minor deformities of the concrete block, or
Fibreboard (tentest board) was placed directly on top of the wall to account for any
irregularities in the top of the wall surface. A composite steel W-section was placed
directly on the fibreboard and rollers perpendicular and parallel to the plane of the wall
73
were placed on top to ensure the vertical load applied to the wall would be evenly
distributed. The two plates on the East and West side of the wall were then attached to
each other via threaded rods and tightened. A 1 kN load was then placed on the wall by
the horizontal actuator, which was regulated by force control. This allowed the hydraulic
jack at the base of the wall to be tightened, without shear forces greater than 1 kN being
Metal targets were glued to each side of the test specimen to provide a flat surface for the
Linear Strain Converters (LSCs). As shown in Figure 3.15, eight LSCs were used to
record the displacements of the wall: four on the West end and four on the East end.
During loading, these displacements, along with the stroke and force of the horizontal
and vertical actuators, were fed to a main computer and electronically recorded.
When each test specimen was aligned and levelled, an initial vertical load of 100 kN was
applied. This load was used to pre-compress the fibreboard at the top of the wall so that
the load would not fluctuate because of crushing of the board. The vertical load from the
actuator was then reduced to a constant 60 kN. When this load was reached, the actuator
was put into stroke control, keeping the position of the actuator locked. The racking
(horizontal) load was then applied to the specimen, also in stroke control. This load was
applied at a constant rate of 1 mm/s, with two cycles per deflection step. The step-
displacement history that was used for the cyclic testing is shown in
Figure 3.16.
These displacements were applied to the wall until failure, which was the point at which
it was deemed based on damage to the wall to be unsafe to continue, or when the peak
load had decreased by at least twenty percent from the ultimate load obtained.
During the first quasi-static test, specimen QS.B.1 rotated out of the East-West base
supports due to minute eccentricities in the alignment of the wall and horizontal actuator.
This occurred on the 10 mm stroke cycle with a horizontal load of approximately 95 kN.
The spherical seat at the end of the horizontal actuator twisted out of alignment due to the
To prevent these small, unavoidable eccentricities from causing the wall to kick out, two
modifications were made to the testing frame. Firstly, bracing was added to the side of
the actuator to ensure it would stay aligned with the rest of the testing apparatus.
Secondly, bracing was provided at the base of the wall on the side closest to the actuator
76
(East side) to prevent any rotation of the wall. These two bracing elements are shown in
Figure 3.17.
Specimen QS.B.2 failed at just below the horizontal actuator’s rated capacity of 250 kN.
Since the remaining test specimens were stronger walls (more grouted cores), a higher
capacity actuator was needed. While an actuator with a capacity of 500 kN would have
been sufficient, the only actuator available had a rated capacity of 1 MN. This bigger
actuator, on the other hand, required a new column with new bracing in order to support
it.
The revised testing frame that was used for the remainder of the tests is shown in Figure
3.18, with additional photos shown in Appendix A. At the back of the actuator, two W-
section columns were bolted to the floor and braced together just above the height where
77
the actuator was bolted to the first column. The tops of these two columns were also
braced together for added stability. New steel plates were also used at the horizontal
bearing area at the top of the wall. These plates were thicker, longer and wider than the
previous plates used, nonetheless, the bearing area applied to the wall was kept the same
such that the effective aspect ratio was still 0.69. On the East side of the frame, two HSS
sections were tack welded to the plate attached to the actuator to act as the new bearing
surface on the wall. The same threaded rods used in the initial test setup were again used
to tighten the plates on each side of the wall and clamp the top of the wall. With the
additional weight of the steel plates, as well as the new actuator, overhead cranes were
In addition to the structural modifications to the testing frame, an LSC was added to
measure the deflection of the W-section column at the back of the horizontal actuator.
This LSC, along with the existing LSC at the West base of the wall were used as
feedback to control the stroke of the horizontal actuator. These two LSCs recorded the
movement of the frame and any sliding of the wall. With the input of these two
parameters, real-time corrections were made in the stroke such that the control step
displacement was the actual displacement applied to the top of the wall as opposed to
being the stroke of the horizontal actuator as in previous tests. This led to more
symmetrical results in the wall specimens, as roughly the same top of wall deflection
The final modification to the testing frame was made after testing two walls using the
improved frame. While loading the second wall tested in this test setup (the first being
QS.S.1), QS.B.5, the wall rotated out of plane at a horizontal load of approximately
100 kN in a similar manner to the test specimen QS.B.1. To ensure that this would not
occur again, a second set of braces at the base of the wall was added on the West side, as
This final quasi-static testing frame was used for the remainder of the tests, QS.B.4,
4.1 Introduction
This chapter focuses on the results and observations that were obtained from the cyclic
tests on the seven wall specimens. The results were analysed, including the comparison
of the performance of bamboo against steel reinforced walls. Several shear models are
also presented, and comparisons of experimental to theoretical wall resistance values are
made.
All seven experimental walls were subjected to cyclic loading using the test setup
described in Chapter Three. The results from the cyclic tests are presented in the same
order in which the walls were tested. In this section, the hysteresis loops, crack pattern
and overall test results are presented for each of the walls. In addition to the results
presented, supplementary photos for each of the walls are provided in Appendix A.
The first of the walls tested was QS.B.1, a regular strength concrete block wall reinforced
with two vertical bamboo poles. The hysteresis loops, and the cracking pattern are
presented below in Figure 4.1. As can be seen by the crack pattern, the wall failed in a
diagonal tension manner, with most of the diagonal cracking splitting through block, as
Figure 4.1 - Hysteresis Loops and Crack Pattern for Wall QS.B.1
An interesting observation is made from the hysteresis loop for this wall at the ultimate
capacity of the wall. The hysteresis loop reaches its ultimate load, and as the deflection
increases, the load starts to decrease, followed by a sudden increase in load. This
occurred not only in the loops for this wall, but also in walls QS.B.4 and QS.B.6.L.
There are several possible explanations for this behaviour. It was observed that the
pressure of the hand pumped hydraulic jack fluctuated slightly, and oil leaked from it.
This could lead to a slight unloading of the wall if it was unable to maintain constant
resistance to the applied load. Once the wall reached its ultimate capacity, load would
start to transfer more rapidly to the internal reinforcement. There is the possibility that
the bamboo reinforcement could have started to slip, before the mechanical anchorage of
the nodes allowed more load to be developed in the bars. Another possibility is shifting
of the frame. If the frame shifted a small amount suddenly, or “jumped”, it could
decrease the load, and then build up again. This became apparent in wall QS.B.5, and is
discussed in more detail in Section 4.2.5. Finally, this tail of the hysteresis loop could be
81
due to the crushing of masonry at the toe of the wall. Even a small amount of material
being loosened from the toe would allow the wall to shift slightly, decreasing the load.
It was noted that the loading and deflection curves for the push (positive deflection) and
pull (negative deflection) cycles, were not equal each other. This very large difference
was noticed in both walls QS.B.1 and QS.B.2, which were both tested using the initial
testing frame in which the actuator was programmed solely on stroke control, or actuator
displacement, and not top of wall deflection. This disparity between the push and pull
cycles comes from the fact that the stroke of the actuator was not necessarily the same as
the top of wall deflections, with greater discrepancy on the pull cycle. This discrepancy
came from several possible sources, the first being the movement of the frame itself. On
the pull cycles, there were two things that were different than the push cycles, one being
that the steel rods holding the top of wall bearing plates were in tension and experiencing
small elongation, and the second being that the in-plane wall support was the steel
bracing, as opposed to the hydraulic jack on the other side. When the test setup was
changed to include the 1 MN actuator, an LSC was added to the frame, and this
measurement, along with the measurement of the LSC at the base of the wall on the west
side was fed back to the main computer. These measurements were used to correct the
stroke to more accurately relate the stroke to the top of wall deflection. As can be seen in
the results from later walls tested on the new frame setup, the discrepancy between the
Wall QS.B.2 had four vertical cores reinforced with bamboo poles, and was the last wall
that was tested using the 250 kN horizontal actuator and initial testing frame. A strange
dip in the hysteresis loops shown in Figure 4.2 can be seen, right after the wall achieves
its ultimate load. The horizontal actuator actually reached its maximum capacity, but this
maximum load was continually applied until the target deflection was reached. The test
program was paused at this point. This sustained load eventually caused the failure of the
wall which caused a more sudden decrease in load than occurred for any of the other
specimens. Again, like wall QS.B.1, the hysteresis loops of the push and pull cycles did
Figure 4.2 - Hysteresis Loops and Crack Pattern for Wall QS.B.2
The crack pattern in this wall seems to indicate that the wall failed in diagonal tension,
however, since the loading was not equal for the push and pull cycles, this is most evident
on one side of the wall. This wall also exhibits what appears to be a compression strut in
the crack pattern, which also indicates that this wall may have the influence of a partial
83
flexural failure. Without a base for the wall to be doweled into, flexural failure can only
occur by crushing of the toe, with a diagonal compression strut, as opposed to yielding of
It was also discovered that wall QS.B.2, due to mis-calibration of the control system, had
an initial vertical load of 20 kN, as opposed to the 60 kN that was placed on all other
walls. This was not noticed until the results from the data were analysed. It has been
shown (e.g. Kranzler and Graubner 2008) that with lower axial loads, in plane shear
With the capacity of the horizontal actuator maxed out on wall QS.B.2, the test setup was
Wall QS.B.3 had all cores fully grouted and reinforced with vertical bamboo poles. It
was noted during the 4 mm cycle of the test that the testing frame had started to shift,
with a creaking noise resonating from the setup. As the horizontal loading reached
approximately 200 kN, the test frame started to slide, or ‘jump’ and the whole frame
would slide millimetres at a time. At this point, the test was stopped, and the bolts
holding the frame to the lab strong floor were re-tightened. The test was then restarted,
and the results and crack pattern of this wall are shown in Figure 4.3.
84
Figure 4.3 - Hysteresis Loops and Crack Pattern for Wall QS.B.3
As can be seen from the crack pattern, the wall did not fail in diagonal tension or sliding,
but by a flexural mode with crushing of the toes of each side of the wall. A photo of the
crushed toe on the west side of the wall is shown above in Figure 4.4. During testing,
uplift of the wall was also noted on the so-called tension side, with the heel lifting
approximately 25 mm above the floor level. As can be seen from the hysteresis loops in
Figure 4.3, the test was stopped after the cycle in which it reached ultimate load. This
85
was done for safety reasons, as it was feared that the crushing of the toe could lead to
lateral instability. With the extremely high loading placed on the wall, it was feared that
the wall could suddenly slip out-of-plane and cause damage to equipment and personnel.
Wall QS.S.1 was the control wall out of the seven walls, and the only wall that was
reinforced with steel. As mentioned in Chapter Three, this wall was reinforced in four
vertical cores, and contained 3 reinforced bond beams. The amount of reinforcing in this
wall was such that the minimum Canadian Code requirements for reinforcing in seismic
areas were met. The hysteresis loops and crack patterns found for this wall are shown
below in Figure 4.5. As can be seen, the wall exhibited a diagonal tension shear failure,
with a slight rocking, and compression strut failure at each of the toes. As in other walls,
this compression strut is likely due to an influence of a flexural failure. This wall, while
not developing a full “X” pattern associated with diagonal shear failure, does appear to
Figure 4.5 - Hysteresis Loops and Crack Pattern for Wall QS.S.1
86
Wall QS.B.5 was one of two test walls that included horizontal bamboo reinforced bond
beams. This specimen in particular had horizontal reinforcement in every course, as well
as vertical reinforcement in every core. It was found after curing of the wall specimen,
that cracking had occurred in the wall, attributed to expansion of the bamboo
reinforcement when placed in the wet grout. Figure 4.6 shows the cracking that had
occurred at the top and side of the wall. These cracks were marked with white paint to
distinguish them from other cracks that would develop during testing and are shown as
blue lines in Figure 4.7. The hysteresis loops and crack pattern are also shown in this
figure.
Figure 4.7 - Hysteresis Loops and Crack Pattern for Wall QS.B.5
This wall failed differently than all of the previous walls, possibly due to the cracking
that was already present in the wall. Similarly to wall QS.B.3, lifting, and somewhat of a
rocking behaviour was observed, crushing the toes at each end. It can be seen from the
hysteresis loops in Figure 4.7, that the ultimate loads reached during the push and pull
loading cycles were not equal, due to the damage that was already present in this wall.
This test was stopped as soon as the ultimate load started to decrease, as it was observed
that the wall had become out of plumb from the testing, and an out of plane failure
seemed imminent.
It can also be seen at the apex of the last two hysteresis loops that as the deflection
increases, the load decreases, increases, and decreases again. It was observed during the
testing of this wall that despite the re-tightening of the bolts, the higher loads in this test
caused the testing frame again to “jump”, or shift suddenly to take up the tiny amount of
space available in the bolts where the testing frame was connected to the floor. The
clamping force of the bolts on the concrete lab floor would be enough to resist the
88
movement of the small gaps, until the lateral force became great enough to overcome this
resistance. At this point, the frame would shift slightly, decreasing the load on the
actuator. With the decreased load, the clamping forces would be too great to resist the
load, and would stop shifting. As soon as the load increased, the process would continue
again, until all the slack in the bolt holes was taken up.
This wall specimen contained three horizontally bamboo reinforced bond beams and four
cores reinforced with bamboo. This specimen, like QS.B.5, had cracking that had formed
in the wall during the curing of the grout. This cracking was again attributed likely due
to the expansion of the bamboo reinforcement in the wet grout. Figure 4.8 shows the
hysteresis loops from the cyclic testing as well as the cracking pattern occurring in the
wall. The blue lines again indicate cracking that was present in the wall before the cyclic
Figure 4.8 - Hysteresis Loops and Crack Pattern for Wall QS.B.4
89
Like wall QS.B.5, the cracking pattern found on this wall was not typical of the normal
failure modes, which was attributed to the extensive pre-cracking. As can be seen in
Figure 4.8, the ultimate load only reached just above 120 kN. The damage to the wall
from the expansion of the bamboo resulted in the lowest lateral resistance out of all the
walls tested.
The last wall subjected to cyclic testing was wall QS.B.6.L. The only difference between
this wall and wall QS.B.1 was that instead of regular strength concrete block, low
strength block was used. Shown in Figure 4.9 are the hysteresis loops and crack pattern
Figure 4.9 - Hysteresis Loops and Crack Pattern for Wall QS.B.6.L
The cracking pattern is somewhat similar to the pattern for wall QS.B.1, but as can be
seen by the hysteresis loops, the loading was more symmetric between the push and pull
90
loading cycles. This led to more of an “X” pattern suggesting that diagonal tension
It was observed during testing of all walls, that there was a significant variation in the
vertical load placed on the walls. This was of course the result of the fact that the initial
vertical load was placed on the wall using force control, and then the actuator was locked
in position using displacement control in an effort to prevent uplift of the wall during the
cyclic loading. The tendency for the wall to want to rotate due to the horizontal load
resulted in increased axial force on the wall from the vertical actuator. The stiffer the
wall was, or the more grout that was used, the greater this fluctuation became, as there
was a higher lateral load resistance and therefore a higher rotational moment applied.
The vertical loadings on the three stiffest walls tested are shown below in Figure 4.10.
As can be seen from Figure 4.10, walls QS.B.3 and QS.B.5 both reach a vertical load of
just under 500 kN, which is the maximum capacity of the vertical actuator. Once its
maximum capacity was reached it would have been unable to hold position, allowing
The deflection reported by the LSCs at the top of wall would therefore be due to the shear
deflection as well as the additional deflection due to the rotation as shown in Figure
4.11b).
The rotation of the wall not only affected the perceived deflections of the wall specimens,
but the increased vertical load would ultimately have had an effect on the wall failure
mode, as well as the failure strength. In testing of URM walls, Kranzler and Graubner
(2008) mention that with an increased normal load on walls, diagonal tension failure of
discussed in Chapter Two, an increased axial load is also known to increase the shear
In all the wall specimens, the predominant shear failure modes were diagonal tension
shear and flexural failure characterized by compressive strut toe crushing with slight
rocking. In many cases, a combination of these two modes was somewhat apparent, with
the emergence of diagonal shear failure first. No sliding shear failure occurred. At first
glance, the rocking / compression strut mode found predominantly in QS.B.3 and QS.B.5
gave the highest deformation out of any of the walls. However, as discussed above, the
rotation increased the measured deflection because of the orientation and geometry of the
tilted wall. In addition, the crushing of the toe also served to increase the measured
deflection above the actual deflection. In the case of the two stiffest and fully grouted
walls, QS.B.3 and QS.B.5, there was significant spalling and damage to the corners of the
93
wall allowing for more rotation, and for the base of the wall to slide. From the hysteresis
loops, it can be seen that there is no significant decrease in load after the ultimate load, as
in all the other tests. While the other walls show large hysteresis loops after ultimate load
capacity, and a significant decrease in load with increased deflection, the fully grouted
walls show a more sudden and brittle failure. While QS.B.4 was fully grouted, it did not
have similar behaviour to walls QS.B.3 and QS.B.5, as the pre-cracking of the wall
One of the things observed during the cyclic testing was the tendency for the test walls to
want to twist and kick out of the in-plane testing frame. Bracing was placed at the
bottom of the walls to prevent this out-of-plane movement however this highlights the
importance that boundary conditions can have on the ultimate failure of a test specimen.
While the test programme was set up to test walls in-plane, more research needs to be
There did not appear to be any pullout or bond failure in any of the tests with the
exception of where there was pre-cracking in the two walls with horizontal bamboo
reinforcement. Figure 4.12 shows the bond between the grout and bamboo in test wall
QS.B.6.L. The bond of the vertical reinforcement appeared to be more than sufficient in
all of the test specimens (save sections of QS.B.4 & QS.B.5 that had pre-cracking)
alleviating fears that poor bond between bamboo and grout could be a problem in
(2005) showed that the tensile capacity of bamboo encased in concrete for 15 years
showed only a minimal loss of strength. This, as well as other properties of bamboo-
reinforced masonry such as long-term bond strength, sustained loading, creep, flexural
As mentioned previously, wall specimens QS.B.4 and QS.B.5, did develop cracking a
few days after they were grouted. These were the only two that contained horizontal
bamboo reinforced bond beams. A total of 80 poles of Tonkin cane were originally
ordered for this research, much more than was needed for the walls constructed. Many of
these poles contained significant horizontal cracking, and were not used for construction.
95
The vertical bamboo bars were the first to be selected and cut from the material available,
and therefore the highest quality bamboo, which contained the least cracking, was used.
Some of the bamboo culms used as horizontal reinforcement, unfortunately did contain
some small longitudinal cracking between nodes. Despite the varnish waterproofing the
small cracks allowed moisture absorption and expansion of the bamboo causing cracking
in several places along the two walls, as shown in Figures 4.6, 4.7 & 4.8. After testing
the walls, the face shell was removed and they were more closely inspected showing that
along sections of the horizontal bamboo reinforcement, some de-bonding had occurred.
Despite the presence of a longitudinal crack in the grout along much of the length of the
bamboo and a few vertical cracks, most of the surface area of the horizontal
reinforcement still had some bond. The typical de-bonding, as well as cracking in the
Compared to the other walls, the two with this pre-cracking exhibited significantly lower
load carrying capabilities, and therefore proper treatment of the bamboo to prevent such
96
cracking is crucial for safe walls. There are three possible solutions for preventing the
moisture from being absorbed by bamboo that contains small cracks. The first is to
simply use bamboo splints instead of culms such that each part of the bamboo could be
properly treated despite the cracking. Unfortunately, this would also increase the labour
needed, as all the bamboo would have to be split, but this may not be an issue in
the cracks after the specimen has been treated. The varnish itself is not a thick enough
substance to fill a crack of more than about a millimetre or two, however, a hard
waterproof substance such as wax, or other similar substance, could span the crack, and
ensure that no moisture makes its way to the inside of the bamboo culm. Finally, the
third way would be to ensure that full culms used have no cracks present, which
recommended that in future testing, the bamboo be split. This would eliminate swelling
due to longitudinal cracks in whole culms, and as suggested by Brink and Rush (1966),
could increase the performance of the bamboo compared to using whole culms.
In any case, the treatment and prevention of moisture to the bamboo reinforcement is
extremely important. As shown here, even a few small cracks along the bamboo can be
enough to allow moisture ingress, and induce cracking and loss of strength of a structural
member.
97
The boundary conditions of experimental arrangements can play a huge part in the results
that the tests produce. In the first two cyclic tests, it was immediately noticed that the
ultimate load on the pull side was much lower than the push side. Possible reasons for
this were assumed to be sliding and bending of the test frame, and/or sliding of the wall
specimen. To combat this, the actuator was programmed to compensate for any
movement at the base of the wall, as well as any frame deflection. This proved to
improve the discrepancies between the push and pull loads, but there still was a slight
difference. The remaining difference could be due to the threaded steel rods used to
connect the bearing plates on the wall to the actuator. Rough calculations showed that
the deflection in the rods due to the loads placed on them during the pull cycle could be
as high as a few millimetres. While this isn’t an extremely large deflection difference, it
could have been enough to give the differences seen in the hysteresis loops. This too is
just a small part of the test setup that had an effect on the outcome of the wall tests.
The fluctuating vertical load discussed above likely also affected the test outcomes. In
force control mode, the actuator would have allowed more vertical movement of the
walls but the vertical load would have remained constant and very different results could
have been obtained. The walls would have likely experienced more rocking, and more of
them would have succumbed to a compression strut failure as was seen in some of the
stiffer walls. If two vertical actuators were used instead of one, the behaviour of the
walls would also be affected since the vertical movement would be better prevented and
the total vertical load on the wall could be kept constant (Fehling and Schermer 2008).
98
The lack of a concrete base into which the vertical reinforcement could be dowelled
obviously prevented a flexural failure in tension from occurring. The mode of failure
may have turned from a diagonal tension to a flexural tensile failure. Even the toe-
crushing mode of failure may have been different, had a concrete base been used.
However, as explained previously, the lack of a concrete base was intended to best
represent the possible boundary conditions that would be present in a low cost housing
In all types of experimental testing, the boundary conditions will affect the results, and
therefore need to be tailored for the type of result that the testing is meant to investigate.
Here, an attempt was made to use the most realistic boundary conditions possible and
investigate specifically the more brittle diagonal tension type shear failure. However, as
a result, other types of failure and properties of bamboo-reinforced masonry could not be
investigated. Further research is required to look at these different failure modes and see
combinations, the envelopes of the cyclic test hysteresis loops were plotted, and placed
on a single graph, shown below in Figure 4.14. The envelopes were taken from the push
side of the hysteresis loops, as they showed the greatest resistance in all wall tests and are
These envelopes for the load and top of wall deflection for each of the walls serve not
only to make a direct comparison of the wall results, but also were used to make further
analytical comparisons.
Section 2.7.2 in Chapter Two described how load-deflection envelopes can be idealized
into a bilinear relationship. This methodology was applied to the envelopes shown in
Figure 4.14. All of the key values that were taken from the envelopes, as well as
The critical points for each bilinear idealization were also plotted graphically on the
envelopes of each wall, with a typical plot shown above in Figure 4.15. Plots for the
remaining walls can be found in Appendix B. All of the bilinear idealizations are shown
in Figure 4.16.
101
From the experimental envelopes, it was found that it was extremely difficult to
differentiate the change in slope once the wall cracked. Therefore, the initial cracking
deflection, Dcr, was taken at the value of 0.7Hmax, as recommended in literature (Ötes and
Löring 2003). Values of De and Du were both determined at the point at which the load
equalled to 0.8Hmax as shown in Figure 2.21. In the case where the load did not drop
significantly enough after reaching the ultimate load to reach eighty percent, the same
value as Dmax, or the maximum displacement was used for Du. The area under the
envelope, Aenv, was obtained by transforming area under the curve into strips and
approximating the strips as trapezoids. This area was then used in Eq. 2.5 to obtain the
With the exception of QS.B.4, all of the walls had a calculated value for ductility, μ,
around two (Table 4.1). In the case of QS.B.4, this departure from the relatively close
102
values for ductility is attributed to the extensive pre-cracking caused by the expansion of
bamboo during the curing period. Although QS.B.5 was also pre-cracked, the damage
was not as significant as in wall QS.B.4. In terms of ductility and the case of diagonal
shear failure mode, the use of bamboo as reinforcement in masonry walls looks
promising. Compared with the steel reinforced wall, all of the walls showed somewhat
equivalent ductility under the given test conditions. However, as discussed previously,
walls QS.B.3 & QS.B.5 had inflated values for their ultimate deflection, and by looking
at their hysteresis loops, it can be seen that their failure mode is more brittle than that of
It has also been noted by Tomazavic (1999) that experimental testing has the potential to
give high values for ductility and that the values would have to be limited in seismic
design to avoid excessive damage to structural walls. It is interesting to note that all the
test results, except QS.B.4, did not reach his suggested limiting value of 4.0 – 5.0 for
reinforced masonry. Three out of the seven walls did not even reach his suggested
limitation for unreinforced masonry, 2.0 – 3.0. The mean ratio of Hu/Hmax = 0.9 from
tests on over sixty walls (Tomazevic 1999) corresponds well with the mean value of 0.87
for the experimental walls tested here. The values coming from wall specimens with
horizontal reinforcement, QS.S.1 and QS.B.5, were particularly close to literature values,
being 0.89 and 0.91 respectively. Other values were in the range of 0.79 -0.98. For the
values calculated for walls QS.B.3 and QS.B.5, if the inflated deflection of the wall did
not occur, the ratios would decrease, as the area under the hysteresis envelope would be
From the hysteresis loops found for each of the walls tested, the cumulative dissipation
energy was calculated by finding the area of the loops between consecutive displacement
peaks. Since there was a discrepancy between the maximums of the push and pull cycles,
the dissipation was calculated for the positive push cycle only. For each of the walls, a
table was constructed, which shows the displacement for each hysteresis loop calculated,
the maximum deflection of the loop, as well as the cumulative dissipation of energy at
that point of testing. These tables are found in Appendix B. Table 4.2 below shows the
As can be seen, the wall with the most dissipated hysteresis energy is wall QS.B.5, yet as
discussed before, both this wall and QS.B.3 had inflated top of wall deflections, which
would have increased this value substantially. The high value for dissipation of energy
for the steel reinforced wall, QS.S.1 indicates that steel reinforcement still is the premier
material for use in resisting seismic loads. Both walls QS.B.1 and QS.B.2, however, show
a fairly high dissipation which implies that bamboo is a viable reinforcing material for
As discussed in Chapter Two, previous testing has shown that different types of masonry
walls such as reinforced, unreinforced and confined masonry experience similar stiffness
analysis was completed. The slopes for each of the experimental hysteresis envelopes
were first determined at intervals of 0.5 mm of the actual top of wall deflection, and
divided by the effective stiffness, Ke. The deflection at these intervals was divided by the
top of wall deflection that occurred at the maximum resistance of the wall, dHmax. This
procedure normalizes both the stiffness, as well as the deflection of each of the walls,
making a direct comparison between walls possible. The results were plotted for each of
the seven test walls, and are shown below in Figure 4.17.
Since the slope points were taken at intervals of 0.5 mm, the plots in Figure 4.17
fluctuated quite a bit. To give a more even representation of the slope degradation, trend
lines were fitted to the data points. These curves are shown below in Figure 4.18.
As can be seen from the experimental hysteresis loops for each of the walls in
Section 4.2, there are sometimes great increases or decreases in the maximum load from
one cycle to the next. As can be seen in the time displacement plot in Figure 3.16 and the
peak load for most of the walls. There are therefore fewer data points available to
construct the hysteresis envelope in this region. Smaller deflection steps in this region
would have allowed not only a more accurate envelope, but also more accurate points for
The stiffness degradation curves in Figure 4.18 do not exhibit the same similarities as
those in Figure 2.23. There is more variation in the values for the ratio of K/Ke compared
to the curves shown in Figure 2.23. It is interesting to note that wall QS.B.4, which had
severe damage from pre-cracking, had one of the smoothest curves, and was most
comparable to curves found in Figure 2.23. This analysis shows that the stiffness of the
walls was not significantly changed by using bamboo as opposed to steel reinforcement,
which is especially noted with the actual normalized points in Figure 4.17. This
conclusion, however, is based mainly upon two modes of failure. With future testing, the
degradation curves could be compared with other boundary conditions to confirm this
result.
In the same time frame that the experimental testing for this thesis was being completed,
investigate the use of GFRP and Geogrid bed joint reinforcement to strengthen ungrouted
masonry walls. (Sadek & Lissel 2008) In this experimental program, an ungrouted,
unreinforced control wall was built by the same master mason, using the same
dimensions, and approximately the same specified strength block (f’m = 12.7 MPa) as in
this thesis. The wall was also experimentally tested in the same testing frame and manner
that was described in Chapter 3. From this testing, hysteresis loops were produced, and
Figure 4.19 - Hysteresis Loops from Experimental Unreinforced Wall (Sadek &
Lissel 2008)
As can be seen from the hysteresis loops, the ultimate shear capacity for this wall was 60
kN, while the corresponding top of wall displacement was only 5 mm. It can be seen that
the unreinforced wall experiences a sudden failure, and the maximum load corresponds to
the maximum deflection. It is apparent that the addition of bamboo reinforcement not
only increases the ductility of unreinforced masonry, but also the ultimate shear capacity.
The two walls QS.B.1 and QS.B.6.L were similar in construction, except for the use of
low strength block instead of regular strength in the latter. Compressive testing of both
grouted, and ungrouted prisms, showed mean strengths for the regular strength block
were approximately thirty percent stronger than their lower strength counterparts. It is
therefore interesting that the maximum ultimate load, Hmax obtained for the regular
strength wall, QS.B.1, was also just under thirty percent higher than QS.B.6.L. The
108
maximum axial load for QS.B.1 was 248.1 kN, while the maximum for QS.B.6.L was
only 198.8 kN. It is possible that the ultimate load capacity could have been slightly
higher for QS.B.6.L had the same higher axial load been reached.
The similarities between the behaviour of the two walls can be more clearly seen from
Figure 4.20 which shows the two experimental envelopes, along with their bilinear
idealizations. Each has a relatively similar peak, and post ultimate load performance,
however, the initial slope of the low strength specimen is slightly lower, having a Ke
value of 26.8 kN/mm as opposed to 29.1 kN/mm. The ductility was also very similar
between the two, with the low strength wall being slightly more ductile.
109
With only two specimens to compare, it is difficult to draw any definitive conclusions
about the performance of low strength versus normal strength block though there does
appear to be good potential for the use of low strength block in low cost housing. From
the limited data, however, it appears that the main difference in using a high strength
Of all the walls that were tested, the bamboo reinforced wall QS.B.2, which had four
vertical cores reinforced, has the most comparable performance to that of the steel
reinforced control wall, QS.S.1. Although they both had the same number of reinforced
cores, the bamboo reinforced wall did not have the two horizontal reinforced bond beams
that were present in QS.S.1 (the third horizontal bond beam was in the top course, where
the load was applied, and therefore was ineffective). For a direct comparison, the
hysteresis envelopes, along with the bilinear idealizations for both walls are plotted in
Figure 4.21.
110
While it may be assumed that the higher ultimate load capacity in the steel reinforced
wall is due to the stronger vertical, and additional horizontal reinforcement, one first has
to also compare the vertical loads that were applied to the two specimens. The vertical
As mentioned in Section 4.2.2, wall QS.B.2 was the one wall that experienced only part
of the intended initial vertical load due to calibration errors, as can be seen in Figure 4.22.
In addition to the initial loading, it can be seen that the loads imposed on the steel
reinforced wall were much greater than those imposed on the bamboo reinforced wall.
shear wall specimens leads to an increase in shear capacity. This begs the question of
what the capacity of the bamboo wall would have been, had the maximum axial load
placed on the wall been the same as QS.S.1. Comparing the points of peak axial load, the
steel reinforced wall has almost three times the axial stress placed on it.
The ductility of both walls was similar as well, with the steel wall being 2.0 and the
bamboo being slightly less ductile at 1.94. This is somewhat apparent in the hysteresis
envelopes, in which the bamboo-reinforced wall has a steeper post-peak slope compared
to the steel reinforced wall. This observation is also reinforced by the values of Hu/Hmax
for each wall, where the steel is the higher of the two at 0.89 compared to 0.8.
It is unfortunate that wall QS.B.4 was damaged from the pre-cracking, as it had the same
reinforcement layout to QS.S.1 and a proper comparison of using horizontal bond beams
with bamboo and steel reinforcement could also have been made. However, even with
the absence of horizontal reinforcement, from the results obtained for QS.B.2 it appears
to have performed almost as well as QS.S.1. While it cannot be said based on one test
112
that bamboo is a sufficient replacement for steel reinforcement for low cost housing, it
This series of tests only looked at the effectiveness of bamboo as reinforcement, in the
case of a diagonal tension or compression strut failure. The case in which bamboo
performed nearly as well as steel may not be so if a flexural tensile failure mode is
introduced. From the perspective of using bamboo in a shear wall, other boundary
conditions such as axial loads, positioning of the horizontal loads, and the use of footing
and capping beams are all parameters that need to be tested. Other factors such as the
species of bamboo available, type of waterproofing and bond agents, block type and
strength, and geometry of the walls are other factors that should be studied to provide
more insight to the strengths and weaknesses of using bamboo as internal reinforcement
in masonry.
There are numerous shear models available to calculate the shear strength of masonry
walls. Calculations were carried out using three different equations to compare the
various models to the experimental results. Diagonal tension failure as well as flexural
and sliding shear failures were considered for each of the walls. Although none of the
selected equations were intended for use with bamboo as a reinforcing material, the tested
properties of bamboo were substituted in place of steel properties. Since bamboo did not
exhibit a yield plateau, the average failure stress of the strain-gauged specimens was used
113
as the yield strength in the shear equations. In the following sections, the models used
The equation for the diagonal shear resistance of a reinforced shear wall is given in
dv
Vn = m (v m bw dv + 0.25Pd ) g + (0.60 s Av f y ) 0.4 f m bw dv g (4.1)
s
In this equation, the factors m and s are material resistance factors. The term g is the
grouting factor for walls. For partially grouted walls g equals the ratio of the effective
cross-sectional area divided by the gross area, but may be no greater than 0.5. For fully
grouted walls, this factor is equal to 1.0. Equation 4.1 also contains the term vm, which is
M
v m = 0.16(2 ) f m (4.2)
Vdv
M
where (2 ) = a value not less than 0.25, nor greater than 1.0
Vdv
114
For a low aspect shear or squat wall, defined as having a height to length ratio (hw/lw) less
than one, the maximum allowable shear is increased by a factor equal to 2 - (hw/lw). This
The use of Equation 4.3 only applies if “care [is] taken to ensure that the shear input to
the wall is distributed along the entire length of the wall and will not lead to a failure of a
Research from the Federal Emergency Management Agency and the National Earthquake
earthquake damage (NEHRP 1997). This equation however, does not apply to masonry
shear strength in the plastic hinge regions, which would give an inelastic response instead
M
Vn = 0.0834.0 1.75 An f m + 0.25 n An + 0.5Ah f yh L /sh (4.4)
VL
This equation, like the Canadian model, has limitations on the maximum shear allowable.
However, in this case, the maximum amount of shear resistance is dependent on the term
Between the values of 0.25 and 1.0, a straight-line interpolation can be used to find the
maximum allowable shear. Both this equation and that in the Canadian Standard include
a shear friction term, allowing for an increase in shear capacity with axial stress, and a
One of the most recent models proposed for the in-plane diagonal shear capacity of
masonry comes from work done by Voon and Ingham (2007). They proposed for
practical design calculations, that the shear strength of masonry be calculated from the
following equation:
Deff
Vn = 0.8k(Ca + Cb )An f m + 0.9N * tan + Ah f yh 0.33An f m (4.6)
sh
In this expression, the value of M/VL is limited to be between 0.25 and 1.0, similar to
S304.1-04 and NEHRP’s equations. The term 0.9N*tan represents the enhancement of
shear strength that comes from the axial compression and resulting diagonal shear strut.
The angle is the angle between the centres of flexural compression at the top and
bottom of the wall and the axis of the wall. For single bending, this is from the heel of
the wall, to the centre of the top of the wall. Finally, the term Deff = L-2d’ -Ldh, where Ldh
should be taken as 20db for reinforcement with a yield strength of 300 MPa, and 35db for
In contrast to the first two models presented, this model considers the effect of
longitudinal reinforcement, and ductility of the wall. The additional shear strength
provided by the vertical reinforcement is accounted for by the term Ca, while the term k is
a factor that takes into account the diminishing effects that these bars have in the case of
None of the above sources provide equations for the determination of the flexural shear
in Figure 2.15. This approach, conversely, does not work for the experimental walls as
there is no tension reinforcement that is doweled into the foundation. Instead, a diagonal
compression strut forms on the experimental walls, stretching from the application of
force, to the toe of the wall. While the CSA S304.1-04 masonry code does not give
provisions for the calculation of compression struts, the CSA Concrete Design Standard
117
A23.3-04 (CSA 2004c) does give guidelines to calculate the strength of the compression
The term Acs represents the cross sectional area of the compressive strut. This area can
be calculated as seen below in Figure 4.23, in which a bearing plate and another
compressive strut anchor the strut. In this figure, the term lb is the length of the bearing
plate, and da is the length of the compression strut. The angle is the angle of the strut,
which is found between the line connecting the toe of the wall to point of application of
Figure 4.23 - Compression Strut Anchored By Bearing Plate and Strut (CSA 2004c)
The term fcu in Equation 4.7 is the crushing strength of the concrete, or effective strength,
given by:
118
f c
f cu = 0.85 f c (4.8)
0.8 + 1701
In this equation, the usable strength of concrete is decreased due to the tensile strain of
the strut, with the term 1. If the tensile strains of the compression strut in the walls are
assumed to be a minimum, fcu can simply be taken as 0.85f`c. If all these terms in
Equations 4.7 and 4.8 are adjusted for masonry, the length of the confining compression
strut is taken as the thickness of the wall, tw, and assuming that half of the axial load, Pd is
transferred into the strut, this gives us a proposed equation for nominal flexural shear
Pd
Vn = (lb sin + t w cos ) t w 0.85 f m (4.9)
2
In the case of partially grouted walls, a linear interpolation can be used between the
properties of a fully grouted section and an ungrouted section. Note that the material
For sliding shear that occurs in reinforced masonry, the following equation is provided in
Vr = m μP2 (4.10)
For this equation, μ is a friction coefficient with a value of 1.0 for masonry to masonry.
P2 is a term that includes the compressive force in the masonry acting normal to the
sliding plane, the dowel action of the vertical reinforcement, as well as the diagonal strut
forces acting on the wall. Note that to calculate nominal resistances, the material
resistance factor, m was removed, and the full axial and compression strut loads were
used. No further models for sliding were considered as it is not a dominant failure mode
in reinforced masonry.
The nominal resistance using each of the aforementioned equations was calculated for
each of the seven test walls. For the Canadian Standard S304.1-04, previous research by
Miller et al. (2005) described the standard’s provisions to be among the most
horizontal shear reinforcement factor in the second part of Equation 4.1 from 0.6 to 1.0,
the theoretical results became less conservative by over 10 percent. In addition to being
conservative on the shear reinforcement, the limiting value of 0.5 placed on the factor g
Therefore, three separate calculations were performed using the S304.1-04 equation to
compare to experimental data. The first used Equation 4.1 with only the material
120
resistance factors set to 1.0, as to obtain a nominal resistance. In addition to setting the
material resistance factors to 1.0, the second set (Case 1) removed the conservatism in
partially grouted walls by removing the limit of 0.5 for the factor g . Finally, the last set
of calculations (Case 2) had the same conditions as Case 1, but also modified the shear
reinforcement factor from 0.6 to 1.0, so the last part of Equation 4.1 was equal to
dv
1.0Av f y . All other equations were used as previously shown. The tabulated
s
The experimental ultimate resistance value, Hmax, as well as the bilinear idealized
resistance, Hu, are also included for comparison. All of the calculations can be found in
Appendix C. For an additional visual reference, all of the calculated nominal resistances
typical plot for wall QS.B.6.L is shown below in Figure 4.24. The graphs for the
To give a better comparison of the different shear resistance equations, the ratio of the
calculated nominal resistance over the experimental resistance, Vn/Hmax, was calculated
for each result. As none of the experimental walls experienced sliding shear failure, the
ratios for this type of failure were not included. The tabulated results are presented below
in Table 4.4.
122
As can be seen by Tables 4.3 and 4.4, the theoretical equations provided varying results.
It is clear that due to the extensive pre-cracking, there was a large loss of strength in wall
QS.B.4, since all of the equations predicted a much higher strength than was achieved in
the test. The nominal sliding shear failure values were well above all other values, which
was to be expected, since no sliding shear failure was observed during experimental
testing. The proposed equation for flexural (compression strut) failure, Equation 4.9,
provided mixed results. On the one hand, for walls that had a predominant flexural
failure (QS.B.3 & QS.B.5), values of Vn/Hmax were below 1.0, and were also less than the
ratios for diagonal tension. From a theoretical point of view, this successfully predicts a
flexural failure over a diagonal tension failure. On the other hand, the calculated values
are quite conservative, ranging from 28 – 32 percent of the experimental result for walls
It can be seen that for wall QS.B.5, all the diagonal shear equations are at least close to
the maximum experimental load, and most exceed the value, indicating a flexural failure.
In the results, the most conservative equation is S304.01-04, when no modifications were
made to the original equations. Even with the modifications to reduce the
conservativeness in Cases 1 & 2 of the same equation, the results varied by quite a bit.
To compare this variation, the results from all of the diagonal tension equations were
taken for only the walls the showed a more dominant diagonal tension failure (QS.B.1,
QS.B.2, QS.B.6.L & QS.S.1), and are presented below in Table 4.5.
Table 4.5 - Shear Equation Comparison for Walls with Diagonal Tension Failure
Mean Coefficient of Variation
Equation Standard Deviation
(Vn/Hmax) (%)
S304.1-04 (2004) 0.60 0.17 28.5
S304.1-04 (2004) Case 1 0.72 0.19 26.5
S304.1-04 (2004) Case 2 0.77 0.29 37.7
NEHRP (1997) 1.0 0.18 17.9
Voon & Ingham (2007) 0.99 0.17 17.3
As can be seen, the values calculated from the NEHRP, and Voon and Ingham equations
were closest to experimental results but both equations still had fairly high coefficients of
variation. The COVs were, however, significantly lower than those for the S304.1-04. It
appears the equations from both NEHRP and Voon and Ingham show the greatest
potential for use when designing masonry shear walls reinforced with bamboo.
The comparison of experimental results and design shear models had a few purposes,
with the first being to see how applicable these equations would be if one simply
could look at modifying the best existing equations to develop one for specific use with
bamboo, perhaps in conjunction with some of the existing equations suggested for
bamboo reinforced concrete, such as those suggested by Brink & Rush (1966). The use
of different equations also illustrated the strengths and weaknesses of the equations, such
as the inability for S304.1-04 to effectively model partially grouted walls. None of these
equations were intended for use with bamboo as reinforcement, yet the equations from
NEHRP and Voon & Ingham were quite accurate in their predictions. In future research,
the results presented should be a valuable basis in formulating design equations for
5.1 Conclusions
masonry shear walls for cyclic loading conditions. The walls were reinforced with
• Test walls that were fully grouted showed much different hysteresis behaviour
than those walls that were not fully grouted. Fully grouted walls failed by
crushing of the toe, as opposed to diagonal shear, and exhibited a much more
abrupt drop off in load after reaching ultimate capacity. With the additional
material, lower ductility, and only marginal increase in shear capacity, the use of
fully grouted walls for low cost housing projects is not recommended.
126
behaviour to that reinforced with steel. The performance between wall QS.B.2,
which had four vertical cores reinforced with bamboo, and QS.S.1, which had
four vertical cores and three bond beams reinforced with steel were very similar
in terms of both ductility and ultimate resistance. The slightly lower ultimate
applied to it, along with the lack of bond beams. In addition, given the variability
in the materials the difference in performance for these two walls cannot be
considered significant.
• The use of low strength block compared to regular strength block did not
significantly affect the shear wall behaviour, other than an expected decrease in
construction.
This research is one of the first attempts to use bamboo as internal reinforcement in
grouted block masonry, and demonstrates that bamboo shows great promise for use in
future low cost housing applications. The relevant material properties of bamboo as
reinforcement were also considered throughout this research, with varying results. Some
• The tensile strength of bamboo from species to species is variable. For use in any
similar such that the specimens do not fail by splitting or the ends being crushed.
variable for different treatments used. The experimental testing carried out here
also showed significant variability for bamboo samples embedded in grout with a
varnish and sand waterproof treatment. It was found, however, that for the
specimens tested with an embedded node, the variability of results was decreased.
As long as no cracking was present in the bamboo poles, the use of spar (marine)
varnish was a simple and effective waterproofing treatment along with a dusting
observed in two of the walls constructed. This led to a reduction in the ultimate
strength of both of these walls, and could also lead to further long term
such that moisture cannot enter through the cracks to cause swelling.
128
be recognised that the experimental testing program described here looked at only
one set of boundary conditions, and as such, the relative performance between
• The analysis of the results and comparison to various shear failure models showed
that the diagonal shear equations from Voon & Ingham (2007) and NEHRP
(1997) predict the the shear strength of bamboo reinforced walls fairly accurately.
Though this research was the first of its kind, it permitted several conclusions regarding
described above. In addition, many areas that should be investigated in the future were
also revealed. Some of the key areas for future research in bamboo-reinforced masonry
include:
129
presented in this thesis. Since there were a limited number of tests performed,
and the materials themselves are quite variable, there is no statistical reassurance
that the results obtained are correct. There is a need for verification of the results
presented in this thesis, ideally with enough (at least 3) repetitions to produce
that experimental testing is quite costly and therefore there are limitations on the
conditions for in-plane shear testing. As discussed, the boundary conditions can
further testing not only on in-plane shear, but also on other aspects of building
connections and their detailing are paramount to determine the behaviour of full
bamboo reinforced concrete, there are only a few limited studies on long term
the long term effects on tensile capacity, bond strength and deterioration of the
humidity. To date, there has been no research carried out on the cyclic or fatigue
• Spar varnish was used in this testing to waterproof the bamboo reinforcement and
prevent swelling. Previous research has shown that there are many different
bamboo reinforcement impermeable. For developing regions, only the most cost
building materials for a respective nation or area would also be useful. This thesis
touched on the high and rising prices of steel and cemenr with the low cost
that each developing nation will have its own infrastructure, natural resources and
labour pool. The masonry units produced in a developing nation are not likely to
have the same high strength as the concrete block produced in Canada, which was
used in this research. In areas where bamboo is readily available and could
block.
131
• With only a slight difference in performance between walls QS.B.2 and QS.S.1,
the use of extensive bond beams becomes questionable. With the additional
labour and construction difficulties that are presented with the addition of bond
beams, it may be beams only at the top and bottom of the wall are necessary for
the results obtained are not reliable. In addition, the performance of a wall with
bond beams, but no vertical cores is something that needs to be investigated. This
could lead to sliding shear becoming a dominant failure mode. More research is
wall performance.
• As discussed, there are a few shear models that were fairly accurate in predicting
results, more effort should go into the development of design equations, with the
reinforced masonry. Such standards or guidelines need not only to provide simple
masonry for developing nations, such that safe housing can be built by anyone.
132
Bibliography
Ali, I. 1979, "New Materials and Techniques for Low Cost Housing Projects",
Proceedings of the International Conference on Housing, Planning, Financing and
Construction in North, Central, South American and Caribbean Countries,
Pergamon Press Inc., Miami Beach, USA, pp. 263-270.
Anderson, D.L. & Priestley, M. J. N. 1992, "In-Plane Shear Strength of Masonry Walls",
Proceedings of the Sixth Canadian Masonry Symposium, Department of Civil
Engineering, University of Saskatoon, Saskatoon, Canada, pp. 223-224.
Anderson, D. & Brzev, S. 2008, Guide to the Seismic Design of Low and Medium-Rise
Masonry Buildings in Canada, Canadian Concrete Masonry Producers Association,
Toronto, Canada.
Anicic, D., Fajfar, P., Petrovic, B., Szavitz-Nossan, A. & Tomazevic, M. 1990,
Earthquake Engineering Buildings (in Serbo-Croatian), 1990, Gradevinska Knjiga,
Beograd.
Anon 1972, The Use of Bamboo and Reeds in Building Construction, Publication No.
ST/SOA/113, Department of Economic and Social Affairs, United Nations, New
York, USA.
Aziz, M.A. & Ramaswamy, S.D. 1981, "Bamboo Technology for Low Cost
Construction", Appropriate Technology in Civil Engineering, ICE by Thomas
Telford Ltd., London, England, pp. 110-112.
Banks, K.G. 2008, The State of Steel - Letter to Customers, April 11, 2008, Blok-Lok
Limited, Weston, Canada.
133
BBC News 2008a, May 27, 2008-last update, Aftershocks Demolish China Homes
[Homepage of BBC News], [Online]. Available: http;//news.bbc.co.uk/go/pr/fr/-
/2/hi/asia-pacific/7422106.stm [2008, June 2,] .
BBC News 2008b, May 12, 2008-last update, History of Deadly Earthquakes [Homepage
of BBC], [Online]. Available: http://news.bbc.co.uk/2/hi/in_depth/2059330.stm
[2008, July 14] .
Blondet, M. (ed) 2005, Construction and Maintenance of Masonry Houses, 2nd edn,
Pontificia Universidad Catolica Peru, Lima, Peru.
Blondet, M. & Villa Garcia, G., 2004, "Earthquake Resistant Earthen Buildings", 13th
World Conference on Earthquake Engineering, Vancouver, Canada.
Blondet, M., Villa Garcia, G. & Brzev, S. 2003, Earthquake Resistant Construction of
Adobe Buildings: A Tutorial, Earthquake Engineering Research Center, Oakland,
USA.
Bouhicha, M., Aouissi, F. & Kenai, S. 2005, "Performance of Composite Soil Reinforced
with Barley Straw", Cement and Concrete Composites, no. 27, pp. 617-621.
Brink, F. & Rush, P. 1966, Bamboo Reinforced Concrete Construction, U.S. Naval Civil
Engineering Laboratory, Port Hueneme, California.
Canadian Standards Association (CSA) 2004a, CSA A179-04 Mortar and Grout for Unit
Masonry, CSA, 5060 Spectrum Way, Mississauga, Canada.
Cook, D., Pama, R. & Paul, B. 1977, "Rice Husk Ash-Lime-Cement Mixes for use in
Masonry Units", Building and Environment, vol. 12, pp. 281-288.
Cook, D.J., Pama, R.P. & Singh, R.V. 1978, "The Behaviour of Bamboo-Reinforced
Concrete Columns Subjected to Eccentric Loads", Mag. Conc. Res., no. 30, pp. 145-
151.
Cox, F.B. & Geymayer, H.G. 1969, Expedient Reinforcement for Concrete for Use in
Southeast Asia: Report I - Preliminary Tests of Bamboo, US Army Engineers
Waterways Experiment Station, Vicksburg, USA.
Datye, K.R., Nagaraju, S.S. & Pandit, C.M. 1978, "Engineering Applications of
Bamboo", Proc. Int. Conf. Materials of Construction for Developing Countries,
Bangkok, Thailand pp. 3-20.
Dowling, D., Samali, B. & Li, J. 2005, "An Improved Means of Reinforcing Adobe
Walls - External Vertical Reinforcement", SismoAdobe 2005PUCP, Lima, Peru.
Drysdale, R. & Hamid, A. 2005, Masonry Structures: Behaviour and Design, Canada
Masonry Design Center, Mississauga, Canada.
El Gawady, M., Hegner, J., Lestuzzi, P. & Badoux, M. 2004, "Static Cyclic Tests on
URM Wall Before and After Retrofitting with Composites", 13th International Brick
and Block Masonry Conference, Eindhoven University of Technology, Amsterdam,
Netherlands.
Fehling, E. & Schermer, D. 2008, "ESECMASE - Shear Test Method for Masonry Walls
with Realistic Boundary Conditions", 14th International Brick and Block Masonry
Conference, University of Newcastle, Callaghan, Australia.
Fehling, E. & Stuerz, J. 2008, "Shear Tests on Clay Unit Masonry Walls Under Static-
Cyclic Loading", 14th International Brick and Block Masonry Conference,
University of Newcastle, Callaghan, Australia.
135
Glenn, H.E. 1950, Bamboo Reinforcement in Portland Cement Concrete, No. 4 edn,
Engineering Experimental Station Bulletin, Clemson, USA.
Hadjri, K. 2005, "Experimenting with Hybrid Construction - Guadua Bamboo and Adobe
- For Housing in Rural Columbia", International Journal for Housing Science and its
Applications, vol. 29, no. 2, pp. 165-179.
Hatzinikolas, M. & Korany, Y. 2005, Masonry Design for Engineers and Architects,
Canadian Masonry Publications, Edmonton, Canada.
Kalita, U.C., Khazanchi, A.C. & Thyagarajan, G. 1978, "Bamboo-Crete Wall Panels and
Roofing Elements for Low Cost Housing", Proceedings of the International
Conference on Materials of Construction for Developing Countries, Bangkok,
Thailand, pp. 21-36.
Kankam, J.A. & Perry, S.H. 1989, "Variability of Bond Strength between Bamboo and
Concrete", ACI Materials Journal, vol. 86, no. 6, pp. 615-618.
Kowalski, T.G. 1974, "Bamboo Reinforced Concrete", Indian Concrete Journal, no. 48,
pp. 119-121.
Kranzler, T., & Graubner, C-A. 2008, "Integral Model for the In-Plane Lateral Load
Capacity of URM (Shear) Bearing Walls and Calibration with Test Results", 14th
International Brick and Block Masonry Conference, The University of Newcastle,
Callaghan, Australia.
Kumar, S. 2002, "A Perspective Study on Fly Ash-Lime-Gypsum Bricks and Hollow
Blocks for Low Cost Housing Development", Construction and Building Materials,
no. 16, pp. 519-525.
136
Kumar, S. & Prassad, M.M. 2005, "Bamboo Reinforced Flexural Members for Housing
Development in India", International Journal for Housing Science and its
Applications, vol. 29, no. 2, pp. 137-151.
Kurian, N.P. & Kalam, A.K.A. 1977, "Bamboo-Reinforced Soil-Cement for Rural Use",
Indian Concrete Journal, no. 1, pp. 21-36.
Lo, T., Cui, H.Z., Tang, P.W.C. & Leung, H.C. 2007, "Strength analysis of bamboo by
microscopic investigation of bamboo fibre", Construction and Building Materials, In
Press, Corrected Proof.
Luytjes, J.B. 1979, "Meeting the Challenge of World Housing", Proceedings of the
International Conference on Housing, Planning, Financing and Construction in
North, Central, South American and Caribbean Countries, ed. O. Ural, Pergamon
Press Inc., Miami Beach, USA, pp. 22-32.
Magenes, G. & Calvi, G.M. 1997, "In-Plane Seismic Response of Brick Masonry Walls",
Earthquake Engineering and Structural Dynamics, vol. 26, pp. 1091-1112.
Magenes, G. & Morandi, P. 2008, "In-Plane Cyclic Tests of Calcium Silicate Masonry
Walls", 14th International Brick and Block Masonry Conference, University of
Newcastle, Callaghan, Australia.
Mehra, S.R., Uppal, H.L. & Chadda, L.R. 1951, "Some Preliminary Investigations in the
use of Bamboo for Reinforcing Concrete", Indian Concrete, no. 25, pp. 20-21.
Miller, S.C., El-Dakhakhni, W., & Drysdale, R.G. 2005, "Experimental Evaluation of the
Shear Capacity of Reinforced Masonry Shear Walls", 10th Canadian Masonry
Symposium, The University of Calgary, Banff, Canada.
Mitchell, B. 2008, Steel Prices - Letter to Customers, May 21, 2008, Northern Metalic
Sales (F.S.J.) Ltd., Fort St. John, Canada.
National Research Council of Canada (NRCC) 2005, National Building Code of Canada,
NRCC, Ottawa, Canada.
Ötes, A. & Löring, S. 2003, Tastvesuche zur Identifizierung des Verhaltensfaktors von
Mauerweksbauten für den Erdbebennachweis; Abchlussbericht, Universität
Dortmund, Lehrstuhl für Tragkonstrucktionen, Dortmund, Deutschland.
Sadek, H., Lissel, S. 2008, “Seismic Performance of Masonry Walls with GFRP and
Geogrid Bed Joint Reinforcement”, 8th International Seminar of Structural Masonry,
Istanbul, Turkey.
Savastano, H., Warden, P.G. & Coutts, R.S.P. 2003, "Potential of Alternative Fibre
Cements as Building Materials for Developing Areas", Cement and Concrete
Composites, , no. 25, pp. 585-592.
Schumacher, E.F. 1973, Small is Beautiful, Harper and Row, New York, N.Y.
Shing, P., Noland, J., Spaeh, H., Klamerus, E. & Schuller, M. 1991, Response of Single-
Story Reinforced Masonry Shear Walls to In-Plane Lateral Loads, U.S.-Japan
Coordinated Program for Masonry Building Research, Report No. 3.1(a)-2,
Department of Civil and Architectural Engineering, University of Colorado, Boulder,
USA.
Shing, P., Schuller, M., Hoskere, V. & Carter, E. 1990b, "Flexural and Shear Response of
Reinforced Masonry Walls", ACI Structural Journal, vol. 87, no. 6, pp. 646-656.
Shrive, N.G. & Page, A.W. 2008, "In-Plane Cyclic Loading of Partially Grouted Masonry
- A Review and Assessment of Research Needs", 14th International Brick and Block
Masonry Conference, The University of Newcastle, Callaghan, Australia.
Singh, M.P.J. & Jain, S.K. 1977, “Use of Bamboo as Reinforcement in Concrete Slabs”,
Technical Note, Central Building Research Institute, Roorkee, India.
138
Sumintardja, D. 1983, "New Methods of Using Bamboo for Low Cost Housing", To
Build and Take Care of What We Have Built With Limited Resources, The 9th CIB
Conference, Natl. Swedish Inst. for Building Rsrch., Stockholm, Sweeden, pp. 36-
37.
Tomazevic, M., Bosiljkov, V. & Lutman, M. 2004, "Robustness of Masonry Units and
Seismic Behavior of Masonry Walls", 13th International Brick and Block Masonry
Conference, Amsterdam, Netherlands.
Tomazevic, M., Lutman, M. & Petkovic, L. 1996, "Seismic Behavior of Masonry Walls:
Experimental Simulation", Journal of Structural Engineering, vol. 122, no. 9, pp.
1040-1047.
Tomaevi, M., Lutman, M. & Bosiljkov, V. 2006, "Robustness of hollow clay masonry
units and seismic behaviour of masonry walls", Construction and Building
Materials, vol. 20, no. 10, pp. 1028-1039.
UNCHS (Habitat) (United Nations Center for Human Settlements) 1999, Basic Facts on
Urbanization, United Nations, Nairobi.
UNCHS (Habitat) (United Nations Center for Human Settlements) 1998, Global Urban
Indicators Database, United Nations, Nairobi.
UNCHS (Habitat) (United Nations Center for Human Settlements) 1996a, An Urbanizing
World: Global Report on Human Settlements 1996, United Nations, Oxford
University Press.
United Nations 1996b, World Urbanization Prospects: The 1996 Revision, United
Nations, New York, USA.
139
Voon, K.C. & Ingham, J.M. 2007, "Design Expression for the In-Plane Shear Strength of
Reinforced Masonry", Journal of Structural Engineering, vol. 133, no. 5, pp. 706-
713.
Voon, K.C. & Ingham, J.M. 2006, "Experimental In-Plane Shear Strength Investigation
of Reinforced Concrete Masonry Walls", Journal of Structural Engineering, vol.
132, no. 3, pp. 400-408.
Included in this Appendix are additional reference photos for the testing frame, as well as
testing photos for each wall specimen and their respective failures.
For clarity, the vertical load vs time was only shown for walls QS.B.3, QS.B.5 and QS.S1
in Chapter 4, as these were the walls that had the highest fluctuation in vertical load. The
Figure B.1 - Vertical Load Fluctuation for Walls QS.B.1, QS.B.2, QS.B.4 &
QS.B.6.L
150
In Chapter 4, tabular results for values calculated for the cyclic testing hysteresis
envelopes were shown in Table 4.1. The results were also plotted, with a typical graph
(Wall QS.B.1) shown in Figure 4.15. The graphs for the remaining walls are shown
below.
The calculated values of the hysteresis energy dissipation for the push cycles of each of
The tabular results for shear resistances based on the shear equations were presented in
Chapter 4 in Error! Reference source not found.. The results were also plotted, with a
typical graph (Wall QS.B.6.L) shown in Figure 4.23. The graphs for the remaining walls
Included in this Appendix are the calculations done to find the shear resistances given in
Table 4.2.
lw := 1590mm h e := 1100mm
Note - effective height of wall is taken at bottom of
he plates tranfering lateral load to specimen
aspectratio := aspectratio = 0.692
lw
5 2 2
b w := 190mm A g := lw b w A g = 3.021 10 mm A bamboo := 226mm d bam := 25.4mm
For all walls, the shear and moment shall be taken as the ultimate load that the wall reached
during cyclic testing. Note that all material factors have been removed, providing a nominal shear
resistance. The area of bamboo in wall specimens is an approximate value, and its yield strength
is taken as the ultimate failure load found in tensile testing.
Wall QS.B.1
2
A e := 166440mm V := 187.6kN M := V h e M = 206.36 kN m
2 6
f'm := 11.5MPa + 13.7MPa f'm = 13.15 MPa
8 8
M M M
Ratio := if < 0.25 , 0.25 , if > 1, 1, Ratio = 0.865
V d v V d v V d v
f'm
v m := 0.16( 2 Ratio) MPa v m = 0.659 MPa
MPa
Ae
g := min 0.5 ,
g = 0.5
Ag
166
f'm h e
Ae
g := g = 0.551
Ag
f'm h e
Vr = 122 kN
2 M
A n := A e A n = 166440 mm Ratio2 := Ratio2 = 0.692
V lw
Pd
n := n = 0.83 MPa
b w lw
f'm
(
Vnmaxi := 0.273 + 0.227Ratio2 A n
MPa
)
MPa Vnmaxi = 259.557 kN
f'm f'm
Vnmax := if Ratio2 0.25 , 0.5A n MPa , if Ratio2 1.0 , 0.33 A n MPa , Vnmaxi
MPa
MPa
Vnmaxi = 259.557 kN
f'm
(
Vn := 0.083 4.0 1.75 Ratio2 A n
MPa
)
MPa + 0.25n A n
(
Vn := min Vn , Vnmaxi ) Vn = 174 kN
Av
A v := 2 A bamboo v := v = 0.003 fyv := 123.7MPa
An
167
lw 0.5
N' := Pd N' = 250.63 kN tan := tan = 0.723
he
0.022 v fyv
Ca :=
MPa
Ca = 0.007 (
Cb := 0.083 4 1.75 Ratio2) Cb = 0.232
f'm
Vmax := 0.33 A n MPa Vmax = 199.175 kN
MPa
f'm
( )
Vn := 0.8 k Ca + Cb A n
MPa
MPa + 0.9 N' tan Vn = 249.539 kN
(
Vn := min Vmax , Vn ) Vn = 199 kN
he
:= atan
= 34.676 deg lb := 190mm d a := 190mm
lw
f'grouted := 11.5MPa f'ungrouted := 13.7MPa
2
A strut := b strut b w A strut = 50226.628 mm
Now if the limiting compressive stress is placed at 0.85 and the ungrouted is taken as 54% solid
Nc 8 Nc
Pstrut := A strut 0.85 f'grouted + 0.54 A strut 0.85 f'ungrouted Pstrut = 359.621 kN
8 8
Pd
Vflex := Pstrut Vflex = 234 kN
2
Wall QS.B.2
2
A e := 212240mm V := 240.6kN M := V h e M = 264.66 kN m
4 4
f'm := 11.5MPa + 13.7MPa f'm = 12.6 MPa
8 8
M M M
Ratio := if < 0.25 , 0.25 , if > 1, 1,
Ratio = 0.865
V d v
V d v V d v
f'm
v m := 0.16( 2 Ratio) MPa v m = 0.645 MPa
MPa
Ae
f'm h e
Vr := min
v m b w d v + 0.25 Pd g , 0.4
( ) MPa b w d v g 2
Vr = 92 kN
MPa lw
If the factor for partially grouting is not limited:
Ae
g := g = 0.703
Ag
f'm h e
Vr := min
v m b w d v + 0.25 Pd g , 0.4
( ) MPa b w d v g 2
MPa
lw
Vr = 130 kN
169
NEHRP
2
A n := A e A n = 212240 mm
M
Ratio2 := Ratio2 = 0.692
Pd V lw
n := n = 0.385 MPa
b w lw
f'm
(
Vnmaxi := 0.273 + 0.227Ratio2 A n )
MPa
MPa Vnmaxi = 323.985 kN
f'm f'm
Vnmax := if Ratio2 0.25 , 0.5A n MPa , if Ratio2 1.0 , 0.33 A n MPa , Vnmaxi
MPa MPa
Vnmaxi = 323.985 kN
f'm
( )
Vn := 0.083 4.0 1.75 Ratio2 A n
MPa
MPa + 0.25n A n
(
Vn := min Vn , Vnmaxi ) Vn = 195 kN
Av
A v := 4 A bamboo v := v = 0.004 fyv := 123.7MPa
An
lw 0.5
N' := Pd N' = 116.43 kN tan := tan = 0.723
he
0.022 v fyv
Ca :=
MPa
Ca = 0.01 (
Cb := 0.083 4 1.75 Ratio2) Cb = 0.232
f'm
Vmax := 0.33 A n MPa Vmax = 248.614 kN
MPa
f'm
( )
Vn := 0.8 k Ca + Cb A n
MPa
MPa + 0.9 N' tan Vn = 185.622 kN
(
Vn := min Vmax , Vn ) Vn = 186 kN
170
he
:= atan = 34.676 deg lb := 190mm d a := 190mm
lw
f'grouted := 11.5MPa f'ungrouted := 13.7MPa
2
A strut := b strut b w A strut = 50226.628 mm
Now if the limiting compressive stress is placed at 0.85 and the ungrouted is taken as 54% solid
Nc 8 Nc
Pstrut := A strut 0.85 f'grouted + 0.54 A strut 0.85 f'ungrouted Pstrut = 403.403 kN
8 8
Pd
Vflex := Pstrut Vflex = 345 kN
2
Wall QS.B.3
2
A e := 302100mm V := 365.4kN M := V h e M = 401.94 kN m
8 0
f'm := 11.5MPa + 13.7MPa f'm = 11.5 MPa
8 8
M M M
Ratio := if < 0.25 , 0.25 , if > 1, 1,
Ratio = 0.865
V d v V d v V d v
f'm
v m := 0.16( 2 Ratio) MPa v m = 0.616 MPa g := 1
MPa
f'm h e
Vr := min v m b w d v + 0.25 Pd g , 0.4
( ) MPa b w d v g 2
Vr = 269 kN
MPa lw
Ae
g := g = 1
Ag
f'm h e
Vr := min v m b w d v + 0.25 Pd g , 0.4
( ) MPa b w d v g 2
MPa lw
Vr = 269 kN
NEHRP
2
A n := A e A n = 302100 mm
M
Ratio2 := Ratio2 = 0.692
V lw
Pd
n := n = 1.585 MPa
b w lw
f'm
(
Vnmaxi := 0.273 + 0.227Ratio2 A n
MPa
MPa ) Vnmaxi = 440.568 kN
f'm f'm
Vnmax := if Ratio2 0.25 , 0.5A n MPa , if Ratio2 1.0 , 0.33 A n MPa , Vnmaxi
MPa MPa
Vnmaxi = 440.568 kN
f'm
(
Vn := 0.083 4.0 1.75 Ratio2 A n
MPa
)
MPa + 0.25n A n
(
Vn := min Vn , Vnmaxi ) Vn = 357 kN
172
Av
A v := 8 A bamboo v := v = 0.006 fyv := 123.7MPa
An
lw 0.5
N' := Pd N' = 478.83 kN tan := tan = 0.723
he
0.022 v fyv
Ca :=
MPa
Ca = 0.016 ( )
Cb := 0.083 4 1.75 Ratio2 Cb = 0.232
f'm
Vmax := 0.33 A n MPa Vmax = 338.075 kN
MPa
f'm
( )
Vn := 0.8 k Ca + Cb A n
MPa
MPa + 0.9 N' tan Vn = 463.775 kN
(
Vn := min Vmax , Vn ) Vn = 338 kN
he
:= atan
= 34.676 deg lb := 190mm d a := 190mm
lw
f'grouted := 11.5MPa f'ungrouted := 13.7MPa
2
A strut := b strut b w A strut = 50226.628 mm
Now if the limiting compressive stress is placed at 0.85 and the ungrouted is taken as 54% solid
Nc 8 Nc
Pstrut := A strut 0.85 f'grouted + 0.54 A strut 0.85 f'ungrouted Pstrut = 490.965 kN
8 8
173
Pd
Vflex := Pstrut Vflex = 252 kN
2
Wall QS.B.4
2
A e := 212240mm V := 124.6kN M := V h e M = 137.06 kN m
4 4
f'm := 11.5MPa + 13.7MPa f'm = 12.6 MPa
8 8
M M M
Ratio := if < 0.25 , 0.25 , if > 1, 1,
Ratio = 0.865
V d v V d v V d v
f'm
v m := 0.16( 2 Ratio) MPa v m = 0.645 MPa
MPa
Ae
2
g := min 0.5 , g = 0.5 A v := A bamboo A v = 226 mm fy := 123.7MPa
Ag
s := 400mm
dv f'm h e
Vr := min
v m b w d v + 0.25 Pd g + 0.60 A v fy
( ) , 0.4 MPa b w d v g 2
s MPa lw
Vr = 146 kN
174
dv f'm
h e
Vr := min v m b w d v + 0.25 Pd g + 0.60 A v fy
( ) , 0.4 MPa b w d v g 2
s MPa lw
Vr = 184 kN
If the horizontal reinforcement factor is changed from 0.6 to 1
dv f'm
h e
Vr := min v m b w d v + 0.25 Pd g + 1.0 A v fy
( ) , 0.4 MPa b w d v g 2
s MPa lw
Vr = 219 kN
NEHRP
2
A n := A e A n = 212240 mm
M
Ratio2 := Ratio2 = 0.692
V lw
Pd 2
n := n = 0.396 MPa A h := A v A h = 226 mm
b w lw
f'm
(
Vnmaxi := 0.273 + 0.227Ratio2 A n
MPa
)
MPa Vnmaxi = 323.985 kN
f'm
f'm
Vnmax := if Ratio2 0.25 , 0.5A n MPa , if Ratio2 1.0 , 0.33 A n MPa , Vnmaxi
MPa MPa
f'm lw
(
Vn := 0.083 4.0 1.75 Ratio2 A n )
MPa
MPa + 0.25n A n + 0.5 A h fy
s
(
Vn := min Vn , Vnmaxi ) Vn = 251 kN
lw 0.5
N' := Pd N' = 119.63 kN tan := tan = 0.723
he
175
for a ductility of approx 2 (this wall had a ductilty of over four, but the mean ductility for all walls
was taken:
k := 0.75
0.022 v fyv
Ca :=
MPa
Ca = 0.012 (
Cb := 0.083 4 1.75 Ratio2) Cb = 0.232
f'm
Vmax := 0.33 A n MPa Vmax = 248.614 kN s h := s s h = 400 mm
MPa
f'm Deff
( )
Vn := 0.8 k Ca + Cb A n
MPa
MPa + 0.9 N' tan + A h fy
sh
Vn = 250.046 kN
(
Vn := min Vmax , Vn ) Vn = 249 kN
he
:= atan
= 34.676 deg lb := 190mm d a := 190mm
lw
f'grouted := 11.5MPa f'ungrouted := 13.7MPa
2
A strut := b strut b w A strut = 50226.628 mm
Now if the limiting compressive stress is placed at 0.85 and the ungrouted is taken as 54% solid + 2
out of the 5 courses have bond beams
Nc 8 Nc 2
Pstrut := A strut 0.85 f'grouted +
A strut 0.85 f'ungrouted ...
8 8 5
8 Nc 3
+
0.54 A strut 0.85 f'ungrouted
8 5 Pstrut = 457.212 kN
Pd
Vflex := Pstrut Vflex = 397 kN
2
176
Wall QS.B.5
2
A e := 302100mm V := 347.0kN M := V h e M = 381.7 kN m
8 0
f'm := 11.5MPa + 13.7MPa f'm = 11.5 MPa
8 8
M M M
Ratio := if < 0.25 , 0.25 , if > 1, 1,
Ratio = 0.865
V d v
V d v V d v
f'm
v m := 0.16( 2 Ratio) MPa v m = 0.616 MPa
MPa
2
g := 1 A v := A bamboo A v = 226 mm fy := 123.7MPa
s := 200mm
dv f'm h e
Vr := min
v m b w d v + 0.25 Pd g + 0.60 A v fy
( ) , 0.4 MPa b w d v g 2
s MPa lw
Vr = 375 kN
177
dv f'm h e
dv f'm h e
2
A n := A e A n = 302100 mm
M
Ratio2 := Ratio2 = 0.692
V lw
Pd 2
n := n = 1.587 MPa A h := A v A h = 226 mm
b w lw
f'm
(
Vnmaxi := 0.273 + 0.227Ratio2 A n
MPa
)
MPa Vnmaxi = 440.568 kN
f'm f'm
Vnmax := if Ratio2 0.25 , 0.5A n MPa , if Ratio2 1.0 , 0.33 A n MPa , Vnmaxi
MPa
MPa
f'm lw
(
Vn := 0.083 4.0 1.75 Ratio2 A n )
MPa
MPa + 0.25n A n + 0.5 A h fy
s
(
Vn := min Vn , Vnmaxi ) Vn = 441 kN
lw 0.5
N' := Pd N' = 479.53 kN tan := tan = 0.723
he
0.022 v fyv
Ca :=
MPa
Ca = 0.016 (
Cb := 0.083 4 1.75 Ratio2) Cb = 0.232
f'm
Vmax := 0.33 A n MPa Vmax = 338.075 kN s h := s s h = 200 mm
MPa
f'm Deff
( )
Vn := 0.8 k Ca + Cb A n
MPa
MPa + 0.9 N' tan + A h fy
sh
Vn = 588.915 kN
(
Vn := min Vmax , Vn ) Vn = 338 kN
he
:= atan
= 34.676 deg lb := 190mm d a := 190mm
lw
f'grouted := 11.5MPa f'ungrouted := 13.7MPa
2
A strut := b strut b w A strut = 50226.628 mm
Now if the limiting compressive stress is placed at 0.85 and the ungrouted is taken as 54% solid
Nc 8 Nc
Pstrut := A strut 0.85 f'grouted + 0.54 A strut 0.85 f'ungrouted Pstrut = 490.965 kN
8 8
Pd
Vflex := Pstrut Vflex = 251 kN
2
179
Wall QS.B.6.L
2
A e := 166440mm V := 240.6kN M := V h e M = 264.66 kN m
4 4
f'm := 8.6MPa + 10.7MPa f'm = 9.65 MPa
8 8
M M M
Ratio := if < 0.25 , 0.25 , if > 1, 1,
Ratio = 0.865
V d v V d v V d v
f'm
v m := 0.16( 2 Ratio) MPa v m = 0.564 MPa
MPa
Ae
f'm h e
Vr := min
v m b w d v + 0.25 Pd g , 0.4
( ) MPa b w d v g 2
MPa lw
Vr = 93 kN
f'm
h e
Vr := min v m b w d v + 0.25 Pd g , 0.4
( ) MPa b w d v g 2
MPa lw Vr = 103 kN
NEHRP
2
A n := A e A n = 166440 mm
M
Ratio2 := Ratio2 = 0.692
Pd V lw
n := n = 0.665 MPa
b w lw
f'm
Vnmaxi := ( 0.273 + 0.227Ratio) A n MPa Vnmaxi = 242.648 kN
MPa
f'm
f'm
Vnmax := if Ratio2 0.25 , 0.5A n MPa , if Ratio2 1.0 , 0.33 A n MPa , Vnmaxi
MPa MPa
Vnmaxi = 242.648 kN
f'm
( )
Vn := 0.083 4.0 1.75 Ratio2 A n
MPa
MPa + 0.25n A n
(
Vn := min Vn , Vnmaxi ) Vn = 147 kN
Av
A v := 2 A bamboo v := v = 0.003 fyv := 123.7MPa
An
lw 0.5
N' := Pd N' = 200.93 kN tan := tan = 0.723
he
0.022 v fyv
Ca :=
MPa
Ca = 0.007 (
Cb := 0.083 4 1.75 Ratio2)
f'm
Vmax := 0.33 A n MPa Vmax = 170.622 kN
MPa
181
f'm
( )
Vn := 0.8 k Ca + Cb A n
MPa
MPa + 0.9 N' tan Vn = 204.809 kN
(
Vn := min Vmax , Vn ) Vn = 171 kN
he
:= atan
= 34.676 deg lb := 190mm d a := 190mm
lw
f'grouted := 8.6MPa f'ungrouted := 10.7MPa
2
A strut := b strut b w A strut = 50226.628 mm
Now if the limiting compressive stress is placed at 0.85 and the ungrouted is taken as 54% solid
Nc 8 Nc
Pstrut := A strut 0.85 f'grouted + 0.54 A strut 0.85 f'ungrouted Pstrut = 276.798 kN
8 8
Pd
Vflex := Pstrut Vflex = 176 kN
2
Wall QS.S.1
2
A e := 212240mm V := 124.6kN M := V h e M = 137.06 kN m
182
4 4 2
f'm := 11.5MPa + 13.7MPa f'm = 12.6 MPa A v := 100mm fy := 462.3MPa
8 8
M M M
Ratio := if < 0.25 , 0.25 , if > 1, 1,
Ratio = 0.865
V d v
V d v V d v
f'm
v m := 0.16( 2 Ratio) MPa v m = 0.645 MPa
MPa
Ae
dv f'm h e
Vr := min
v m b w d v + 0.25 Pd g + 0.60 A v fy
( ) , 0.4 MPa b w d v g 2
s MPa lw
Vr = 210 kN
dv f'm h e
Vr := min
v m b w d v + 0.25 Pd g + 0.60 A v fy
( ) , 0.4 MPa b w d v g 2
s MPa
lw
Vr = 260 kN
dv f'm h e
Vr := min
v m b w d v + 0.25 Pd g + 1.0 A v fy
( ) , 0.4 MPa b w d v g 2
s MPa
lw
Vr = 315 kN
183
NEHRP
2 M
A n := A e A n = 212240 mm Ratio2 := Ratio2 = 0.692
V lw
Pd 2
n := n = 1.167 MPa A h := A v A h = 100 mm
b w lw
f'm
(
Vnmaxi := 0.273 + 0.227Ratio2 A n )MPa
MPa Vnmaxi = 323.985 kN
f'm f'm
Vnmax := if Ratio2 0.25 , 0.5A n MPa , if Ratio2 1.0 , 0.33 A n MPa , Vnmaxi
MPa MPa
f'm lw
( )
Vn := 0.083 4.0 1.75 Ratio2 A n
MPa
MPa + 0.25n A n + 0.5 A h fy
s
(
Vn := min Vn , Vnmaxi ) Vn = 324 kN
lw 0.5
N' := Pd N' = 352.43 kN tan := tan = 0.723
he
for a ductility of approx 2 (this wall had a ductilty of over four, but the mean ductility for all walls
was taken:
k := 0.75
0.022 v fyv
Ca :=
MPa
Ca = 0.019 (
Cb := 0.083 4 1.75 Ratio2) Cb = 0.232
f'm
Vmax := 0.33 A n MPa Vmax = 248.614 kN s h := s s h = 400 mm
MPa
184
f'm Deff
( )
Vn := 0.8 k Ca + Cb A n
MPa
MPa + 0.9 N' tan + A h fy
sh
Vn = 401.613 kN
(
Vn := min Vmax , Vn ) Vn = 249 kN
he
:= atan
= 34.676 deg lb := 190mm d a := 190mm
lw
f'grouted := 11.5MPa f'ungrouted := 13.7MPa
2
A strut := b strut b w A strut = 50226.628 mm
Now if the limiting compressive stress is placed at 0.85 and the ungrouted is taken as 54% solid + 2
out of the 5 courses have bond beams
Nc 8 Nc 2
Pstrut := A strut 0.85 f'grouted + A strut 0.85 f'ungrouted
...
8 8 5
8 Nc 3
+ 0.54 A strut 0.85 f'ungrouted
8 5 Pstrut = 457.212 kN
Pd
Vflex := Pstrut Vflex = 281 kN
2