HYPERELASTIC HOLD-DOWN
FOR CROSS-LAMINATED TIMBER SHEAR WALLS

by

Hosein Asgari

B.Sc., Azad University, Tehran, Iran, 2011
M.Sc., Amirkabir University of Technology, Tehran, Iran, 2014

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
November 2020

© Hosein Asgari, 2020

Abstract
Cross-laminated Timber (CLT) is increasingly being used in tall buildings. However,
there are some challenges when designing high-rise CLT structures, amongst them the need
for novel hold-downs (HD), for shear walls. While commonly used HDs behave as a
dissipative connection, the current Canadian Standard for Engineering Design in Wood
recommends designing HDs as a non-dissipative connection. As hyperelastic material, an
elastomer (rubber) is capable to carry high loads without inelastic deformation. This thesis
presents experimental studies at material- and component-levels using a hyperelastic rubber
HD solution for CLT walls. A total of 53 quasi-static monotonic and cyclic tests were
performed. The HDs exhibited high strength and deformation capacity without any residual
deformation after unloading. The shape factor and loaded area of rubber layers were found as
the main effective factors on the rubber HD’s response, and an empirical load-displacement
relation was also developed based on these parameters.

i

Table of Contents

Abstract: ........................................................................................................................................... i
Table of Contents ............................................................................................................................ ii
List of Figures ................................................................................................................................ vi
List of Tables ...................................................................................................................................x
Acknowledgments ......................................................................................................................... xi
1

2

Chapter 1: Introduction ............................................................................................................1
1-1

Background .......................................................................................................................1

1-2

Hold-downs .......................................................................................................................2

1-3

Objectives .........................................................................................................................3

1-4

Thesis organization ...........................................................................................................4

1-5

Scope and limitations ........................................................................................................5

Chapter 2: Literature Review...................................................................................................6
2-1

Wood and engineered wood products as structural materials ...........................................6

2-2

Cross-laminated timber .....................................................................................................7

2-2-1

Overview....................................................................................................................7

2-2-2

Relevant material properties and failure modes ........................................................8

2-3

Lateral load-resisting systems for wood structures .........................................................10

2-3-1

Light-frame shear walls ...........................................................................................10

2-3-2

Hybrid solutions and braced frames ........................................................................12

2-4

CLT lateral load resisting systems ..................................................................................13

2-4-1

Overview..................................................................................................................13

ii

2-4-2

Research on CLT shear walls ..................................................................................14

2-4-3

Vertical joints in CLT shear walls ...........................................................................17

2-4-4

Post-tensioned CLT walls ........................................................................................18

2-5

3

Hold-down solutions .......................................................................................................19

2-5-1

Conventional timber structure hold-downs .............................................................19

2-5-2

Dowel-type connections ..........................................................................................20

2-5-3

Slip friction connectors ............................................................................................21

2-5-4

Steel tube-type connectors .......................................................................................23

2-5-5

Perforated steel plate connections ...........................................................................24

2-5-6

X-RAD connector ....................................................................................................25

2-5-1

CSA-O86 design provisions for CLT shear walls ...................................................26

2-6

Elastomeric bearing applications ....................................................................................27

2-7

Summary of literature review .........................................................................................30

Chapter 3: Rubber material tests............................................................................................32
3-1

Materials and methods ....................................................................................................32

3-1-1

Compression tests ....................................................................................................32

3-1-2

Simple shear test ......................................................................................................35

3-1-3

Volumetric test.........................................................................................................38

3-2

Hyperelastic coefficients .................................................................................................40

3-3

Finite element analyses ...................................................................................................42

3-4

Results .............................................................................................................................44

3-4-1

Compression tests ....................................................................................................44

3-4-2

Double-lap shear tests ..............................................................................................47

3-4-3

Volumetric tests .......................................................................................................49

3-5

Evaluation of hyperelastic coefficients ...........................................................................52

3-6

Summary of rubber tests .................................................................................................54

iii

4

5

Chapter 4: CLT tensile strength tests.....................................................................................56
4-1

Specimen description ......................................................................................................56

4-2

Materials .........................................................................................................................57

4-3

Methods ...........................................................................................................................59

4-4

Results .............................................................................................................................59

4-5

Discussion .......................................................................................................................64

4-6

Summary of CLT material tests ......................................................................................67

Chapter 5: Hold-down tests ...................................................................................................69
5-1

Specimen description ......................................................................................................69

5-2

Materials .........................................................................................................................71

5-3

Test series overview ........................................................................................................72

5-4

Test set-up and loading protocols ...................................................................................76

5-5

Hold-down tests to determine the strength of the assembly ...........................................77

5-6

Results of monotonic stiffness tests ................................................................................79

5-7

Cyclic test results ............................................................................................................84

5-8

Hold-down strength test results ......................................................................................91

5-9

Discussion of monotonic tests ........................................................................................95

5-10 Discussion of cyclic tests ................................................................................................96
5-11 Empirical description of the load-displacement behavior...............................................97
5-12 Discussion of strength tests .............................................................................................99
5-13 Summary of hold-down tests ........................................................................................101
6

Chapter 6: Conclusion and outlook .....................................................................................103
6-1

Conclusion ....................................................................................................................103

6-2

Future research ..............................................................................................................104

References....................................................................................................................................107

iv

Appendix A.1-Rubber compression tests ....................................................................................117
Appendix A.2- Rubber simple shear tests ...................................................................................125
Appendix A.3- Rubber volumetric tests ......................................................................................128
Appendix A.4- Monotonic and cyclic tests of rubber HD ...........................................................129

v

List of Figures
Figure 1-1 Sketch of rubber hold-down (HD) assembly for CLT shear walls. ........................ 3
Figure 2-1 (a) Brock Commons tall timber building, Vancouver, BC [30], and (b) Treet tall
timber building, Bergen, Norway [31]. .................................................................................. 12
Figure 2-2 (a) common CLT shear wall connections; (b) different CLT connection categories
................................................................................................................................................ 13
Figure 2-3 Dowel-type connection with the internal steel plate: (a) sketch, (b) test set-up, and
(c) spacing of the fasteners on the bottom side of the test specimen [47] “With permission
form ASCE”............................................................................................................................ 20
Figure 2-4 Resilient slip friction (RSF) joint: (a) main components, (b) assembly, (c)
hysteresis loop [52] “With permission form ASCE”. ............................................................. 22
Figure 2-5 (a) proposed steel tube connector inside of the CLT panel and (b) the loaddisplacement curves related to three different tube sizes [55]. ............................................... 24
Figure 2-6 (a) plain elastomeric bearing (b) reinforced elastomeric bearing. ........................ 27
Figure 3-1 Rubber geometry under compression load. .......................................................... 33
Figure 3-2 Compression test on the rubber specimens under different boundary conditions:
(a) 4 sides free (b) 2 sides free - 2 sides restrained (c) 4 sides restrained. ............................. 35
Figure 3-3 Simple shear test: (a) specimen sketch (b) 13.5 mm-thick (c) 20.5 mm-thick
rubber ...................................................................................................................................... 36
Figure 3-4 Simple shear test: (a) set-up (b) loaded sample. ................................................... 37
Figure 3-5 Volumetric tests: (a) sketch, (b) components of the fixture, and (c) rubber
samples. .................................................................................................................................. 39
Figure 3-6 FEM model: a) double-lap shear test; b) volumetric test. ..................................... 43
Figure 3-7 Compressive load-displacement of specimen R4 under different boundary
conditions. ............................................................................................................................... 44

vi

Figure 3-8 Compressive load-displacement of specimen R4 for different rubber layers layout.
................................................................................................................................................ 45
Figure 3-9 Compressive load-displacement of specimen R4 under different load speeding. 45
Figure 3-10 Maximum displacements vs. shape factor. ......................................................... 47
Figure 3-11 Load-displacement (six cycles) of the simple shear test and the shear stressstrain curves: (a) and (b) specimen #1, (c) and (d) specimen #2, and (e) and (f) specimen #3.
................................................................................................................................................ 48
Figure 3-12 (a) Load-displacement curve and (b) compression-volumetric ratio curve of
rubber specimen #5 under 400N pre-load and different displacement rates. ......................... 51
Figure 3-13 (a) Load-displacement curve and (b) compression-volumetric ratio curve of
rubber specimen #5 under high displacement rate. ................................................................ 51
Figure 3-14 (a) Load-displacement curves and (b) compression-volumetric ratio curves of all
six rubber specimens together; under the low displacement rate and 200N pre-load. ........... 51
Figure 3-15 Curve fitting for the average shear stress-strain data. ......................................... 53
Figure 3-16 Shear stress-strain based on FEA and experiments of the specimen #1. ............ 54
Figure 3-17 Experimental compression-volumetric ratio curve and vs. FEM results. ........... 54
Figure 4-1 Schematic of CLT specimens; (a) side view, (b) top view, and (c) definition of the
parameter x. ............................................................................................................................ 56
Figure 4-2 (a) schematic of the test set-up (b) the real test set-up. ......................................... 59
Figure 4-3 (a) Row shear failure and (b) load-displacement curve of specimens in group 1. 61
Figure 4-4 (a) Row shear failure and (b) load-displacement curve of specimens in group 2. 61
Figure 4-5 (a) Row shear failure and (b) load-displacement curve of specimens in group 3. 61
Figure 4-6 (a) Row shear failure and (b) load-displacement curve of specimens in group 4. 62
Figure 4-7 (a) Row shear failure and (b) load-displacement curve of specimens in group 5. 62
Figure 4-8 (a) Row shear failure and (b) load-displacement curve of specimens in group 6. 62
Figure 4-9 (a) Row shear failure and (b) load-displacement curve of specimens in group 7. 63

vii

Figure 4-10 (a) Row shear failure and (b) load-displacement curve of specimens in group 8.
................................................................................................................................................ 63
Figure 4-11 (a) Row shear failure and (b) load-displacement curve of specimens in group 9.
................................................................................................................................................ 63
Figure 4-12 Row shear failure and (b) load-displacement curve of specimens in group 10. . 64
Figure 4-13 Row shear failure and (b) load-displacement curve of specimens in group 11. . 64
Figure 4-14 Three different failure mechanisms; (a) no intersection, (b) symmetrical
intersection, and (c) unsymmetrical intersection between the opening and the lamellas. ...... 65
Figure 4-15 Average strength of CLT specimens vs. the loaded end distance (aL). .............. 66
Figure 4-16 Average strength of CLT specimens vs. the unloaded edge distance (a c). ......... 67
Figure 5-1 CLT panel for HD test; (a) sketch of the face view, (b) sketch of the top view, and
(c) photo. ................................................................................................................................. 69
Figure 5-2 (a) Sketch of rubber pads and steel plates, (b) steel plate, and (c) rubber pad. .... 70
Figure 5-3 (a) Technical sketch of the steel rods and (b) real steel rod.................................. 71
Figure 5-4 CLTs’ layers thicknesses (a) group A and (b) group B ........................................ 73
Figure 5-5 Rubber layers; (a) one-layered and (b) two-layered rubber pads for group A, and
(c) one-layered and (d) two-layered rubber pads for group B. ............................................... 73
Figure 5-6 Geometry of Rubber layers with a centric hole under compression. .................... 75
Figure 5-7 (a) Sketch of test set-up; the experimental test set-up for (b) group A and (c)
group B. .................................................................................................................................. 76
Figure 5-8: Component level loading protocols: a) monotonic and b) cyclic loading. .......... 77
Figure 5-9 Load-displacement curves from monotonic tests: (a) 100 mm-depth, (b) 175 mmdepth, (c) 120 mm width , and (d) 140 mm width rubber HDs, and (e) one-layered and (f)
two-layered rubber HD in terms of the shape factor. ............................................................. 81
Figure 5-10 140-2-1 rubber HD with partial failure. .............................................................. 82

viii

Figure 5-11 Hold-down deformation mechanism under different arrangements of rubber
pads: a) 120-1-1, b) 140-1-1, c) 120-2-1, d) 140-2-1, (e) 100-1-D, (f) 175-1-D, (g) 100-2-D,
and (h) 175-2-D. ..................................................................................................................... 84
Figure 5-12 Load-displacement curves of 100 mm-depth rubber HD system: (a) 100-1-S, (b)
100-1-D, (c) 100-2-S, and (d) 100-2-D under monotonic and cyclic loading. ....................... 86
Figure 5-13 Load-displacement curves of 175mm-depth rubber HD system: (a) 175-1-S, (b)
175-1-D, (c) 175-2-S, and (d) 175-2-D under monotonic and cyclic loading. ....................... 87
Figure 5-14 Load-displacement curve of 120 mm width rubber HD system: (a) 120-1-0.75,
(b) 120-1-1, (c) 120-2-0.75, and (d) 120-2-1 under monotonic and cyclic loading. .............. 88
Figure 5-15 Load-displacement curve of 140 mm width rubber HD system: (a) 140-1-0.75,
(b) 140-2-0.75, (c) 140-1-1, and (d) 140-2-1 under monotonic and cyclic loading. .............. 89
Figure 5-16 (a) Partial failure on the CLTs’ face layers (b) Cracks at the rounded corners of
the slot; (c) Delaminated lamella, and (d) lack of complete support for the rubber layers. ... 91
Figure 5-17 Load-displacement of HD under failure tests (a) group A and group B 120 mm
width slot; (b) group B 140 mm width slot. ........................................................................... 92
Figure 5-18 Failed specimens: (a) S1-120-2-0.75, (b) S2-120-1-0.75, (c) S3-120-1-1; (d) S1140-1-1, (e) S2-140-1-1, (f) S3-140-1-1, (g) S4-140-1-1. Local bearing on (h) transverse and
(i) longitudinal layers of CLT panels, and (j) squeezed rubber. ............................................. 94
Figure 5-19 Variation of rubber layers’ displacement vs. the shape factor; (a) maximum
displacement at F=50kN and (b) maximum displacement at F=100kN. ................................ 96
Figure 5-20 Estimated load-displacement curves for (a) one-layered, and (b) two-layered
rubber HDs.............................................................................................................................. 99
Figure 5-21 Slot location to the lamellas; (a) S1-120-2-0.75, (b) S2-120-1-0.75, (c) S3-120-11, (d) S1-140-1-1, (e) S2-140-1-1, (f) S3-140-1-1, (f) S4-140-1-1. ..................................... 101

ix

List of Tables
Table 3-1 Mechanical properties of Masticord [76]. .............................................................. 32
Table 3-2 Features of four different rubber specimens used in the compression test. ........... 34
Table 3-3 Simple shear test series .......................................................................................... 37
Table 3-4 Compression test results for four rubber specimens. ............................................. 46
Table 3-5 Volumetric test results. ........................................................................................... 49
Table 3-6 Hyperelastic coefficients of the rubber. ................................................................ 53
Table 4-1 Strengths properties and modulus of elasticity of CLT panels [MPa] [88]. .......... 57
Table 4-2 Test series overview ............................................................................................... 58
Table 4-3 Results summary of CLT tension test. ................................................................... 60
Table 5-1 CLT prototypes dimensions for groups A and B [mm]. ........................................ 71
Table 5-2 Strengths properties and modulus of elasticity of CLT panels [MPa] [88]. .......... 72
Table 5-3 Mechanical properties of Masticord [76]. .............................................................. 72
Table 5-4 Overview of tested hold-down arrangements......................................................... 75
Table 5-5 Monotonic test results for all rubber HDs .............................................................. 80
Table 5-6 Cyclic test results of all rubber HDs. ..................................................................... 85
Table 5-7 Results summary of rubber HD failure test using. ................................................. 93

x

Acknowledgments

I owe a particular gratitude to my parents and all beloved ones; without their support I would
not be able to stand here.
First of all, it has been a great honor for me to work with Dr. Thomas Tannert; immense
amount of his time and effort was dedicated to the whole process of my study as both my
academic supervisor and personal mentor. He always provided thoughtful recommendations
and invaluable ideas for all parts of my research including the preparation of experimental
test setup, performing the experiments, analysis of results, and writing the thesis.
Undoubtedly, this thesis could not be successful without his brilliant supervision and perfect
support. Furthermore, I want to say a special thanks to Dr. Mehdi Ebadi for his contribution
in some parts of this research study. The other appreciation goes to the committee members,
Dr. Shahria Alam, Dr. Asif Iqbal, and Dr. Cristiano Loss, for the review of my thesis and the
kind and thoughtful comments.
Apart from the aforementioned contributions, the generous assistance from the graduate
students and technicians in the Wood Innovation and Design Center, and the Wood
Innovation Research Laboratory can never be forgotten. Special thanks to my colleague
Alison Conroy for her valuable help and suggestions on my academic writings. And last but
not least, the gentle supports by the senior lab instructor Maik Gehloff and the labtechnicians Michael Billups and Ryan Stern are greatly appreciated.

xi

1 Chapter 1: Introduction
1-1

Background

Since the early 20th century, the number of high-rise buildings has been rapidly growing
throughout the world with concrete and steel as dominant construction materials. However,
due to environmental concerns and high population growth especially in urban areas, it
appears essential to use alternative materials like wood for non-residential and tall buildings
because of the advantages such as carbon sequestration, sustainability, and faster
construction [1].
The development of mass timber products like Cross-laminated timber (CLT) made it more
feasible to use the wood in high-rise construction. Due to the structural properties of mass
timber products, mass timber structures – compared to light-frame wood structures – are
more resistant against different load cases because of the higher stiffness and strength along
with their structural performance are improved in terms of dimensional stability and fire
protection [2].
There are examples of tall (hybrid) wood buildings in Canada like the 18-storey Brock
Commons in Vancouver and the 13-storey Origine in Quebec City [3], and the maximum
height of wood buildings was recently increased to 12-storeys from 6-storeys [4, 5].
However, there are some challenges regarding tall timber buildings. Compared to steel and
concrete high-rise structures, tall wood buildings are lighter, and as a consequence, they are
more vulnerable against deflections and vibrations caused by lateral loads such as strong
winds. Further, since wood fails in brittle modes in tension and shear, wood structures rely

1

entirely on their connections to provide ductility and absorb energy when exposed to extreme
earthquakes.
Shear wall systems are one of the common lateral load resisting systems (LLRS). CLT
structures are stiffer and stronger than the light-frame wood structures; but because of high
rigidity, CLT walls have a negligible deformation under lateral forces. Hence, connections of
CLT shear wall systems act as ductile structural fuses to absorb energy when exposed to
earthquakes. Due to the higher load demand for the CLT shear wall system in tall wood
building applications and lower deformation capacity of CLT panels, common connections
adopted for light-frame wood systems are not suitable [6].
1-2

Hold-downs

Hold-downs (HD) are wall to foundation/floor connections; they provide the shear wall
system with resistance against uplifting. To adopt tall CLT-based buildings, it is required to
find alternative solutions rather than the common hold-down employed in low-rise lightframe buildings. A proper hold-down facilitates the CLT walls to rotate around their corner
(rocking). This rocking kinematic motion is preferred to dissipate the energy generated by
lateral forces through energy-dissipate connections between coupled CLT panels [7].
According to the Canadian Standard for Engineering Design in Wood CSA-O86 (2019) [2],
CLT shear walls need to be designed to move dominantly in the rocking mode under seismic
loads. Moreover, even though combined rocking/sliding was considered permissible
according to the 2016 version of CSA-O86 [8], it is now recommended to minimize or even
eliminate any sliding of CLT shear walls. The new design provisions further stipulate that

2

HDs should not participate in the energy dissipation [2]. Hence, HDs should be designed
with high deformation capacity to allow CLT walls to rock.
For this purpose, hyperelastic HDs based on the concept of elastomeric bearings seem a
viable option. Elastomeric bearings are used for both concrete and steel bridges as a link
between the foundation and bridge girders and decks. These bearings allow for large
displacements caused by environmental and traffic loads without structural damage [9]. In
the application as part of a CLT shear wall system, elastomeric bearings because of the
hyperelastic behavior, can carry high lateral loads and withstand large displacements without
permanent deformation after unloading. The proposed HD concept is shown in Figure 1-1.

Figure 1-1 Sketch of rubber hold-down (HD) assembly for CLT shear walls.

1-3

Objectives

The main objective of this study is to evaluate the structural performance of a hyperelastic
HD solution for CLT shear walls. The specific objectives are to determine the:
1) Rubber’s material properties, specifically its hyperelastic coefficients.
2) Required connection parameters to avoid brittle failure modes in the CLT.

3

3) Required geometry of the elastomeric HD solution for given target performance.
To achieve these objectives, experimental tests are conducted first at the material level
(rubber and CLT) and then on the component level to determine the proposed prototype’s
load-displacement behavior under quasi-static monotonic and cyclic loading for a given
target load. The effects of the number of rubber layers, the rubber geometry, and two
different configurations on the hold-down stiffness are assessed.
1-4

Thesis organization

In Chapter 2, a comprehensive literature review about the connection systems of wood
buildings is presented from common fasteners and connection systems for the light-frame
buildings to innovative connections for mid- and high-rise mass timber buildings. In Chapter
3, the material test to evaluate the properties of the elastomers including the simple shear
test, the volumetric test, and the compression test are presented. The hyperelastic coefficients
are estimated using the results of the material tests. In chapter 4, the experimental results of
the tensile strength of CLT panels are reported; the brittle failure on the CLT is avoided on
the design of the rubber hold-down system. Based on the net section tension strength of CLT
panels and its parameters, the rubber hold-down system can be designed for target loads. In
chapter 5, the rubber HD test results are presented. The influence of different variables such
as the geometry of rubber blocks on the performance of the rubber hold-down system is
evaluated to provide enough data for optimum design of the novel rubber hold-down for CLT
shear walls. In chapter 6, the results achieved in this study are summarized.

4

1-5

Scope and limitations

This research is focused on the novel elastomeric HD for CLT shear walls. The proposed
prototype was tested under quasi-static monotonic and cyclic loading. The experimental data
regarding the tensile strength of CLT panels can be used to design any type of internal
bearings or connections for CLT-based structures. The rubber material tests provides
guidance in designing any type of rubber bearings for different applications; and it can also
be employed for further numerical analysis of rubber behavior. The elastomeric properties of
rubber HD characterized in this study can be used in designing the novel hybrid bearing
assemblies for mass timber structures.
Full-scale CLT shear wall systems using the elastomeric HD as an elastic connection with
high load-carrying capacity can be designed based on the outcomes of this experimental
study. Furthermore, the experimental results can be used to verify the new design procedures
of CLT shear wall system equipped with distinct nonlinear connections compared to the
traditional ones. This future work was outside the scope of this research as was a cost
analysis as well as fire and acoustic insulation of the rubber HD. Other considerations such
as constructability and required on-site installation provisions were not part of this research
study’s aims.

5

2 Chapter 2: Literature Review
2-1

Wood and engineered wood products as structural materials

Wood as an orthotropic material has different properties in different directions and its
properties may be influenced by different effective variables. Commonly, mechanical
properties of wood are estimated in two different directions; 1) parallel to the grain and 2)
perpendicular to the grain. The strength properties and stiffness of wood are highly distinct
for parallel and perpendicular to the grain directions; for instance, the tensile strength
perpendicular-to-grain might only be equal to 2-6% of that in the parallel to the grain
direction [10]. Moreover, exposure to different environments considerably affects the wood
functions, because wood properties as a hygroscopic material highly rely on the moisture
content. The lower moisture content usually improves the strength properties of wood, so the
moisture content of wood structures should be kept around 5% if the highest strength is
desired [11].
Wood as a building material exhibits multiple benefits. Besides architectural features, wood
possesses an even higher strength-to-weight ratio than steel. Low heat conductivity and high
electrical resistance make wood a promising material for insulation purposes. Wood is
renewable as a natural resource and it often makes constructions faster than other options.
Moreover, the increasing environmental crises require using more environmental-friendly
and sustainable materials like wood for construction purposes; as the carbon footprint of
wood is considerably lower than that of the other alternatives due to low embodied energy
[12].

6

To satisfy new demands, engineered wood products such as mass-timber products are
becoming more popular in constructions. In addition to typical wood advantages, the
mechanical properties and strength of mass timber products are enhanced compared to
traditional wood structures; as natural defects like knots are reduced. The adhesive is
employed to create a strong bond between the wood members used to construct the
engineered wood products. Thus, not only the structural stability and strength of engineered
wood products are improved but this novel manufacturing process also provides them higher
resistance to natural destructive agents such as fungi and insects. Engineered wood products
can be produced in high variety shapes and dimensions; they do not depend on the size of
supplied lumbers. In other words, it is not required to use the large trees harvested from oldgrowth forests to manufacture beams and panels for large-scale applications [13].
2-2

Cross-laminated timber

2-2-1 Overview
Cross-laminated timber (CLT) panels as an engineered wood product provide many benefits.
CLT is a laminated composite with usually an odd number of layers; with each layer made of
wooden boards (lamellas) glued together orthogonal to the next layer. This arrangement
makes CLT a dimensionally stable panel with high stiffness and strength in two directions.
For this reason, the CLT panels can compete with concrete slabs to be used as the floor
system.
The high population growth, especially in rural areas, necessitates more high-rise buildings
[14], and environmental concerns require the use of sustainable building practices. CLT is a
wells-suited material to meet these needs as floors and walls. CLT as prefabricated wood

7

panels makes construction projects faster compared to concrete and steel due to the ease of
panel installation. Regarding fire safety, CLT panels have higher fire resistance than lightframe structures; the mass timber panels form an insulation layer against the fire through
charring. However, rolling shear failure on the cross layers of CLT panels can occur and the
rolling shear strength specified in the Canadian Standard for Engineering Design in Wood
CSA-O86 is only 0.43, 0.50, or 0.63 MPa for five different stress grades CLT panels [15].
2-2-2 Relevant material properties and failure modes
CLT panels are usually made with specific layers’ thicknesses between 17 mm and 35 mm.
The width of lamellas (individual boards) varies from 30 mm to 240 mm while their common
width is 130 mm for Canadian CLT. CLT boards are made of stress-graded lumbers;
including visually stress-grade lumber (V grades) and machine stress-rated lumber (E
grades). In total, 5 different stress-grade lumbers are specified in CSA-O86 [15]. Each stress
grade consists of specific lumbers combinations for the longitudinal layer (major axis) and
transverse layer (minor axis); higher stress grades lumber is usually used for the major axis.
Hence, the strength properties in different load cases and modulus of elasticity E are defined
for both all layers. The shear modulus G, the transverse modulus of elasticity ET, and also the
rolling shear modulus of each layer can be estimated based on E [15].
To investigate the CLT panels’ behavior subjected to the in-plane loading like lateral loads,
the shear properties of CLT panels acting as walls and floors were studied [16, 17]. Besides
two different shear failure modes under out-of-plane loading, three different in-plane shear
failure modes were recognized for CLT panels; including 1) gross-shear, 2) net-shear, and 3)
torsion [18]. The CLT panels with narrow face bonding commonly fail under gross-shear
failure mode while the lack of bonding between the narrow face of CLT panels and the
8

existence of cracks or gaps between the narrow faces of lamellas pushes CLT panels to fail
under net-shear failure mode, eventually. As well as theses, torsion failure mode might
happen in the interface of CLTs’ layers which are glued together orthogonally [19].
The uplift-resistance of HDs installed at the bottom corners of CLT walls induces tension
stress on the CLT walls when they are subjected to the lateral loads [7]. Few studies were
conducted regarding the tensile strength of CLT panels, even though an empirical formula
was proposed [20] that take into account the tensile properties of lamellas used on the face
layer of CLT panels acting as the effective parameter on the tensile strength of CLT panels
loaded in the major strength axis direction.
Another experimental study was conducted using CLT panels made of Japanese Sugi [21].
Different CLT panels were subjected to the uniaxial tensile load while their width and lay-up
were considered as the main variables. The results showed that the failure at finger-joints or
knots in the outermost layer was the main failure mode for the CLT panels loaded in major
direction. The lamellas on the cross layers simply separated. Furthermore, it was found the
width of CLT panels had no significant impact on their tensile strength for the width range
considered.
In general, timber connections mainly rely on geometrical factors and the materials’
mechanical properties. While brittle failure commonly occurs for wooden structures
subjected to the shear or tension loading, the connection system equipped with ductile
members and fasteners can prevent brittle failure. Further, for CLT structures there are the
other effective factors on the connection design. The position of fasteners and connections’
members to the lamellas on both longitudinal and transverse layers is important to be
considered for the CLT connection system. According to CSA-O86 [15], it is only required

9

to consider the ductile failure of CLT connection systems using dowel-type fasteners; and
there is no need to check the CLT resistance against row shear failure and group tear-out
failure. However, these provisions are valid for common types of connections that employ
small diameter fasteners, and when it comes to the CLT panels with big openings, there is
not enough data available. Some experimental results showed that the size, the number, and
the position of fasteners to the CLT’s layers are significant factors to cause brittle failure in
the CLT connection system rather than expected ductile behavior [22, 23].
2-3

Lateral load-resisting systems for wood structures

All buildings are exposed to lateral forces resulting from wind and earthquake. High-rise
buildings are more susceptible to lateral forces compared to low-rises. A building’s structural
components to resist lateral forces are commonly referred to as Lateral Load Resisting
System (LLRS); this system makes buildings capable of transferring shear forces and
moments that act on the roof and all other floors to the foundation. LLRSs also provide
stiffness to control the building’s lateral drift to meet serviceability requirements and to
avoid collapse due to coupling second-order effects [24]. Thus, LLRSs should satisfy two
criteria; 1) desired resistance to the shear forces, and 2) required stiffness to control lateral
drifts. The common LLRSs for wood buildings include light-frame shear walls, braced
frame, steel/concrete-wood hybrid solution, and CLT shear walls.
2-3-1 Light-frame shear walls
The shear wall system in light-frame wood buildings is the most common LLRS in low and
mid-rise buildings in North America. Light-frame shear walls include two main structural
elements; horizontal floor systems, and vertical walls referred to as “diaphragm” and “shear

10

wall”, respectively. Shear walls transfer the lateral loads generated by seismic loads or wind
pressure in each storey to the diaphragm below. Light-frame shear walls include two main
connections; the sheathing-to-framing joints commonly made by nails, and the anchorages
and shear connectors connecting vertical walls to floors [25].
To characterize the mechanical performance of shear walls in the light-frame wood
buildings, a large body of previous research is available, e.g. [26]. The effect of different
parameters on the shear walls’ response to the lateral load was investigated. According to the
experimental results, the vertical load considerably affected the rocking kinematic of the
shear walls. If the sheathing-to-framing connection is weaker than the hold-down, the nail
spacing has an almost linear correlation to the wall lateral resistance. Besides this, the
sheathing panels also play an important role on the walls’ behavior, especially when they
carry out-of-plane loads.
Some full-scale shake table tests were conducted on the multi-storey light-frame timber
buildings to assess their seismic performance. The full-scale test results showed that the main
part of the energy released by lateral loads was absorbed by the sheathing-to-framing joint,
and this joint was deformed plastically while hold-downs and angle brackets remain almost
elastic and with no permanent deformation [27, 28]. Some research on the full-scale tests
focused on the interior and exterior finishing of walls showing that the seismic response
could be significantly improved through employing the gypsum wallboard as the interior
surface of sheathing panels [29].

11

2-3-2 Hybrid solutions and braced frames
There are some challenges regarding tall timber buildings. Compared to steel and concrete
high-rise structures, tall wood buildings are lighter, and as a consequence, they are more
vulnerable to lateral loads. To tackle this problem, timber hybrid structures are a viable
solution. Tall buildings like Brock Commons in UBC Vancouver combine a concrete core
acting as the LLRS with a timber frame structure carrying gravity loads [30], c.f. Figure 2-1
(a). Some pure tall wood buildings use other LLRS; as an example, the 14-storey Treet
building in Bergen, Norway uses braced frames to resist wind pressure. Concrete slabs are
installed at specific levels to improve the dynamic behavior of the building by increasing the
weight [31], c.f. Figure 2-1 (b).

(a)

(b)

Figure 2-1 (a) Brock Commons tall timber building, Vancouver, BC [30], and (b) Treet tall timber
building, Bergen, Norway [31].

12

2-4

CLT lateral load resisting systems

2-4-1 Overview
Unlike light-frame wood shear walls, the deformation of CLT shear walls is negligible under
in-plane lateral loadings due to the high rigidity of CLT panels, so all deformation and
energy dissipation capacity must be provided by their connections [32, 33]. The CLT shearwall with its common connections is shown in Figure 2-2 (a). Three different connection
categories are defined: wall-to-wall or floor-to-floor connections, the wall-to-foundation
connections, and the wall-to-floor connections [19], c.f. Figure 2-2 (b).

(a)

(b)

Figure 2-2 (a) common CLT shear wall connections; (b) different CLT connection categories

There are two main systems to construct CLT buildings: i) platform-type and ii) balloon-type
construction. In the balloon-type system, continuous CLT walls are erected along the height
of the building, and the floor systems of each storey are connected to it. In contrast, the
platform-type system uses the floor system of each storey as the base of the wall system,
resulting in a large number of walls being employed in the LLRS. More research studies are

13

conducted on the performance of CLT shear wall systems utilizing the platform-type system
[2].
2-4-2 Research on CLT shear walls
Based on the desired kinematic mode, the connections of CLT shear walls are designed as
dissipative and non-dissipative. The characteristics of connection systems on CLT walls can
affect the kinematic behavior of CLT walls under lateral loading, as the number of fasteners
employed in vertical joints between wall segments as well as their layout can push CLT
walls to have the desired rocking deformation mode. Experimental results showed that higher
vertical loads on CLT walls and higher height-to-length aspect ratio of CLT wall segments
can facilitate the rocking mode; these factors also contribute to the self-centering
characteristic of CLT panels after unloading [34].
CLT walls’ behavior under monotonic and cyclic loading protocols was studied by Ceccoti et
al. [35] as a part of the SOFIE project. Different configurations of CLT walls were tested to
estimate their seismic performance; with the main factors considered being the anchorage
system, the wall openings, and the effect of inter-storey connection systems. The
configurations of connection systems like the types and numbers of fasteners were found to
highly affect the behavior of CLT walls. According to the obtained results, the connections
were only acting as dissipative energy members while the CLT panels were almost behaving
as a rigid body.
To investigate the seismic performance of CLT shear walls, a series of monotonic and cyclic
tests were conducted by Popovski et al. [32], with three different configurations of walls’
aspect ratio along with four different types of steel brackets acting as angle brackets being

14

considered. The connection systems were equipped with various fasteners like nails and
rivets to figure out their effects on the lateral resistance of the walls. Besides these, the step
joint was considered as the screwed wall-to-wall connection in the multi-walls
configurations. Simpson Strong-Tie HDs were installed for some connection systems
between walls and the floor to estimate its influence on the load-carrying capacity. It was
found that the CLT walls which were equipped with the steel brackets were efficient enough
and that employing HDs at both ends of the wall improved it slightly. The shear walls with a
higher aspect ratio could reach higher load-carrying capacity. Moreover, the step joints on
the two CLT panel configurations made it possible to increase the walls’ contribution in
deformation of the shear wall system, and as a consequence, a lower stiffness of the shear
wall system and higher ultimate deformations were achieved.
Besides experimental research, analytical studies have been performed as well. Shahnewaz et
al. [36] formulated the lateral resistance of CLT shear walls under different configurations of
connections. This analytical formulation was governed for CLT shear walls used under
platform-type construction. In general, five configurations of connection systems on the
single and coupled CLT panels were considered. It was assumed that the hold-down could
only resist against uplifting forces while brackets could take deformations in both the sliding
and rocking deformation mechanisms. The lateral resistance of CLT shear walls under
different configurations was separately evaluated against sliding, rocking, and combination
of sliding-rocking. The ratio of sliding to rocking deformation for each configuration of CLT
shear walls was estimated in terms of the CLT walls’ aspect ratio. In the final part, the
analytical results derived from the theoretical formulas were compared to the previous
experimental results.

15

Casagrande et al. [37] presented a comprehensive study regarding the kinematic models of
multi-panel CLT shear walls under lateral loading. Analytical equations of elastic stiffness
and load-carrying capacity were developed based on the minimum potential energy principle,
defined as a function of the stiffness of the hold-down and the wall-to-wall joints, and the
geometrical factors of the CLT panels. It was shown that the deformation mechanism of the
CLT panels depends on the stiffness of the hold-down and the vertical joints. The wall-towall connections with lower stiffness than that of the hold-down system (relatively stiff holddown case) push the multi-panel CLT walls system to rock. However, in the case that the
CLT panels are equipped with flexible hold-downs, the other factors including the vertical
load, and the numbers of CLT panels joined together act as effective parameters on the
kinematic of the multi-panel CLT walls.
Lukacs et al. [38] reviewed different established methods for evaluation of the load-carrying
capacity and the stiffness of CLT panels. It mentioned that most methods used the static
equilibrium equation to develop the analytical procedures while the CLT walls were
considered as the rigid panels, and the deformation of the CLT wall system was mainly
considered based on the connections’ properties. All methods were developed for the
deformation mechanism of a single CLT wall. All methods took into account the angle
brackets and hold-downs resisting only against sliding and uplifting, respectively; however,
the method developed by Gavric and Popovski was introduced as the only method which
considers the interaction of horizontal and vertical resistance in the angle brackets, even
though it was proved that the hold-down has no significant resistance against loads in the
shear direction [39].

16

2-4-3 Vertical joints in CLT shear walls
In terms of energy dissipation, CLT shear walls can be divided into two categories. Panels
with a low height-to-length ratio rely more on the wall-to-floor connection systems to
dissipate the energy induced by seismic loads. For systems with a high height-to-length ratio,
the vertical joints act as the main dissipative connection; because these can make the wall
structures more flexible, the higher energy level can be absorbed and dissipated by
deforming the vertical joint’s fasteners and also the relative displacements of CLT walls to
each other [7].
Self-tapping screw (STS) has been used as the common fasteners for the wall-to-wall and
floor-to-floor CLT connections; much research has been done and is still being completed to
evaluate the characteristic of screwed connections on the CLT panels. Gavric et al. [40]
performed an experimental study of screwed connections under monotonic and cyclic tests.
Different configurations of CLT panels as floors and walls were considered, including
parallel wall-to-wall connections, wall-to-wall orthogonal connections, wall-to-floor
connections, and floor-to-floor connections. To create the wall-to-wall connections two
different types of joints comprising the lap joint and the spline joint were provided on the
CLT panels while the floor-to-floor connections were only created through the lap joint. The
obtained results showed that the CLT panels connected by the spline joints had higher loadcarrying capacity with higher displacement, even though the stiffness of lap joined panels
was higher than that of the spline joined panels. Moreover, the panels subjected to the
withdrawal loads were dealing with the screw head penetration more than pulling the screw
out of the panels.

17

Hossain et al. [41] presented an experimental study to investigate the performance of a novel
STS solution. CLT panels were connected together using the half-lap joint using three
different screw layouts; the two first layouts included the shear and withdrawal screw
configurations and the third one was a combination of a shear and withdrawal layout of
screws. Monotonic and cyclic tests were done using the small and large scale of CLT panels.
The experiments showed that the CLT panels with shear connections were highly ductile but
not so stiff; on the contrary, the CLT panels with withdrawal connections were highly stiff
but they failed before reaching the large displacements. Hence, the novel layout of screws
provided both advantages of the CLT panels’ performance under shear and withdrawal
loading: high stiffness as well as large ductility. The results showed that the number of
screws could impact both load-carrying capacity and stiffness of the STS withdrawal
connections; further, the group effect on the STS connections was investigated by the same
authors in another study [42].
2-4-4 Post-tensioned CLT walls
To increase the resilience of tall timber building against lateral loads, post-tensioned CLTbased structures were studied. Garney et al. [43] tested CLT wall systems equipped with
vertical post-tensioning bars in order to investigate the seismic behavior of the high resilient
CLT LLRS. Besides the post-tensioned (PT) bars in the middle of each CLT wall segment,
the U-shape flexural plates (UFP) were employed to connect the wall segment together.
Here, the PT bars and the UFPs were acting as hold-downs and angle brackets, and vertical
screwed connections for the self-centering CLT LLRS, respectively. The energy dissipation
was achieved through yielding of the UFP connections placed between the CLT wall
segments under cyclic loading. The respective limit states of self-centering CLT walls under

18

cyclic loading were yielding of UFPs, splitting of CLT walls, yielding of PT cables, and the
CLT crushing.
Chen et al. [44] evaluated the performance of PT CLT walls equipped with an extra energy
dissipation device. Tests were performed using the single PT CLT wall as well as PT CLT
shear wall systems. A steel fuse was used as another dissipator device, as it behaved as an
extra joint between CLT walls and the foundation. Besides the dissipator fuses, the UFPs
were installed between two walls of the CLT shear wall system to act as another dissipation
device. According to the experimental results, using the dissipator fuses improved the
mechanical properties of the CLT shear walls and also the single CLT wall. It was found that
the fuses yielded and buckled at the early step of the loading process along with local
crushing on the CLT panels corresponding to the maximum lateral drift of the CLT walls
equal to 2.5% or higher.
2-5

Hold-down solutions

2-5-1 Conventional timber structure hold-downs
Traditionally, L-shaped steel anchorages are used as hold-downs at the corner of light-frame
shear walls to support them against uplifting, with the vertical section is connected to walls
by nails while an anchorage bolt connects the horizontal side to foundations. Steel straps act
as the other anchorage systems for low-rise light-frame buildings, with the emerged section
joining the light-frame walls’ studs to the foundation. In addition to these solutions, it is also
common to connect the light-frame wall to the foundation directly using threaded rods or
foundation bolts embedded to concrete foundations. The Simpson Strong-Tie Strong-Rod is
using as a reliable and also common solution for multi-story wood buildings. This type of

19

connector, which is also called the continuous rod tie-down system, is able to stand out
against two various tension loads including shear wall overturning forces and uplift forces on
roofs [45]. Although some studies investigated the performance of conventional HDs for
CLT-based low/mid-rise buildings, it seems required to adopt novel solutions for tall CLT
buildings [46].
2-5-2 Dowel-type connections
Ottenhaus et al. [47, 48] characterized the performance of dowel-type connections as HD for
mass timer products under monotonic and cyclic loading. The connection was installed using
internal steel plates inserted into slots of CLT panels, c.f. Figure 2-3. Variables including the
spacing of the dowels and the loaded end distance were changed to find the effective
parameters on the failure mode and the mechanical properties of the connection system.

(a)

(b)

(c)

Figure 2-3 Dowel-type connection with the internal steel plate: (a) sketch, (b) test set-up, and (c)
spacing of the fasteners on the bottom side of the test specimen [47] “With permission form ASCE”.

20

It was found that the gap between the non-glued edges on the face layers of CLT panels had
the potential to considerably decrease the resistance and ductility because they were acting as
the potential shear failure planes. Similarly, the rolling shear strength of the middle layer of
CLT panels may adversely impact the strength of the connection. It was concluded that the
ultimate ductile strength of the designed connection system which was defined according to
the European Yield Model (EYM) must be higher than the ultimate brittle strength of the
CLT panels if the connection system with relatively high ductility range is targeted.
2-5-3 Slip friction connectors
Loo et al. [49] proposed a new design of slip-friction connectors as hold-downs. The steel
shims of common slip-friction connectors were replaced by an abrasion resistant plate. In the
common design, the friction between mild steel plates provides the desired energy dissipative
capability and hysteretic stability. The rocking motion of the rigid timber wall was achieved
by employing the slip friction connectors at both ends of the wall [50]. Regarding the
hysteretic behavior, there was no mitigation on the stiffness or the strength of mass timber
walls under reversed cyclic loading. It was also found that the new design of the shear key
for minimization of the wall sliding had no undesirable effect on the rocking movement
mode of the wall.
Based on the concept of slip friction connectors, an innovative connection called the resilient
slip friction (RSF) joint was developed [51]. Some critical factors were changed compared to
the slip friction connectors, including the replacement of the slotted center plate of the slip
friction connectors with two separated plates with a specific shape, resulting in the center
plates being able to ridge inside of the two outer plates. Along with conical spring washers,
high-strength bolts were used to keep the connection as tight as desired with high potential
21

flexibility in the perpendicular direction to the plane of the steel plates. The RSF joint and its
theoretical hysteresis loop are illustrated in Figure 2-4.

(a)

(c)
(b)
Figure 2-4 Resilient slip friction (RSF) joint: (a) main components, (b) assembly, (c) hysteresis loop
[52] “With permission form ASCE”.

Additional experimental studies [52, 53] assessed the seismic performance of the CLT wall
system equipped with the RSF joint. In addition, a novel shear key design was proposed. The
new shear key provided high resistance against sliding while it was minimally activated
against the uplifting force so that the desired rocking deformation mode was achieved. The
experiments showed that there was no damage to the CLT wall or the structure of the RSF
joint even under large displacements. The CLT wall was capable to self-center. Furthermore,
the reversed cyclic force-displacement graphs showed that high energy dissipation was
achieved through frictional sliding of the slotted-center plates of the RSF joint as designed.

22

2-5-4 Steel tube-type connectors
Using a steel tube as a member of the connection system on engineered wood structures was
firstly introduced by Murty et al. [54], with a small diameter tube being welded to the steel
plate and all inserted into the timber panel. The proposed connection showed a highly ductile
response and its tube-type fasteners were mainly acting as the failed members. A hollow
steel tube embedded into CLT panels was proposed by Schneider et al. [55] to increase the
initial stiffness and ductility of mass timber panel connections. Three different tubes
diameters were mounted inside the CLT panel and subjected to both monotonic and cyclic
loading. High energy dissipation was achieved but it was required to increase its resistance. It
was also found that two important mechanical properties, including yield point and loadcarrying capacity, were dependent on the tube size, with the smaller tube diameter led to
higher load-carrying capacity. Moreover, according to the cyclic test results, the tube
connector was considerably able to dissipate energy if the appropriate size for the tube
geometry was chosen. The proposed steel tube connector inside of the CLT panels and the
force-displacement curves related to different series of monotonic tests are shown in Figure
2-5.
A numerical study was performed by Mpidi Bita and Tannert [56] to optimize the design of
the steel tube connector, as it was previously found that the diameter and the thickness of the
tube as well as the coupler diameter play a significant role on its load-carrying capacity.
Before the analysis of the original model with tube connectors, material-level investigations
using numerical and experimental procedures were completed validating the material
behaviors of the CLT panels and the steel tube. A specific configuration of tube connectors
was modeled based on the most optimized conditions to evaluate its performance in terms of

23

different unknown factors in the material properties and the connection geometry. It was
mentioned that all numerical simulations and optimization analyses were implemented in
terms of a 3-layer CLT panel; hence, it was predicted that the optimum design on the CLT
panel with more layers and higher thicknesses could improve the mechanical characteristics
of the connector. To consider the steel tube connector as a viable hold-down solution for
mid-rise CLT-based buildings, it was noted that it is feasible to employ two or three steel
tubes joined together in a group.

(a)

(b)

Figure 2-5 (a) proposed steel tube connector inside of the CLT panel and (b) the load-displacement
curves related to three different tube sizes [55].

2-5-5 Perforated steel plate connections
The proprietary Holz-Stahl-Komposit-System (HSK System) has been employed as holddown in some CLT-based structures like the Wood Innovation and Design Center in Prince
George, BC [57]. In this connection system, a perforated steel plate is inserted into a slot
created inside the CLT panels and then bonded with adhesive. The holes of the steel plates,
which are called adhesive dowels, act as the load-carrying elements. The steel links between
the holes behave as a flexible component to provide the desired ductility. According to

24

Bathon et al. [58], the HSK System is not only stiff and ductile but fails under ductile failure
mode regardless of the load cases when the appropriate design was selected. However, there
was buckling risk to the HSK system even though high stiffness was provided to the
connection system.
The performance of a modified HSK System was studied by Zhang et al. [59]. The reducedsection side plates were replaced with the steel profile without any reduction in its section.
Tests at the material scale, component mid-scale, and full-scale hold-down tests were
performed under quasi-static monotonic and reversed cyclic loading. As a novel design,
some rows of perforated steel plates were covered by duct tape to provide more flexibility in
comparison with the original HSK System. According to the results, the failure mode of the
perforated plate under tension was the yielding of SL while the failure mode of the inserted
plate into the timber was a fracture at the timber-adhesive interface, as well as a ductile
failure in the SL portions. It was found that the number of SLs plays a crucial role on the
connection properties including strength, stiffness, and ductility. Moreover, the obtained
results showed that the holes which were covered by the tape led to achieving higher
ductility and energy dissipation, and also the risk of buckling was considerably prevented.
2-5-6 X-RAD connector
Polastri et al. [60] studied the proprietary X-RAD systems as an alternative to the common
hold-down and angle brackets developed for the light-frame wood structures. This
connection is installed at the corners of CLT walls. Its components include fully-threaded
STS, hardwood inner core, thin steel envelope, and an internal steel plate. The different STS
installation angles help the X-RAD resist loads in different directions. To prevent the brittle
failure of the internal hardwood element and provide adequate ductility, two different
25

deformation mechanisms are activated: 1) the ductile steel envelope and the internal steel
plate which are deformed when the axial loads are transformed to the hardwood component
form the CLT walls, and 2) the transversal bolt which retains the steel envelope tight to the
hardwood core would bend, and so provide the high ductile capacity to the X-RAD
connector system.
Four different mid-scale test series were conducted, with the X-RAD connection system
loaded under tension, shear, shear-tension, and shear compression loading. The X-RAD
connection could be considered as a stiff connector with high load-carrying capacity.
Regarding the failure mechanism, although the ductile members like the transversal bolts and
the steel envelope deformed under highly plastic deformation mode, the connection system
ultimately collapsed due to the block-shear failure mode of the internal plate. Further to these
failure mechanisms, local brittle failure and splitting failure mode of the hardwood core was
also observed. The X-RAD energy absorbing capability depended on the loading direction
because failure mechanisms and inelastic deformation happened through different
components under different loading.
2-5-1 CSA-O86 design provisions for CLT shear walls
According to the 2019 version of the Canadian Standard for Engineering Design in Wood
CSA-O86 [15], CLT shear walls must be designed using wall segments with an aspect ratio
to generate the desirable rocking kinematic motion. Wall sliding should be eliminated and
cannot be accounted for energy dissipation. While previous provisions allowed discrete HDs
to act as a dissipative connection, the new standard requires designing HDs to not participate
in the energy dissipation. The vertical joints between walls should act as the main dissipative
components of CLT shear wall systems.
26

Consequently, hold-downs used in platform-type construction must avoid inelastic
deformation while satisfying the demands to carry a high load with high deformation
capacity. Hence, due to its hyperelastic behavior, elastomeric bearings can be considered as a
potential hold-down solution for CLT shear walls to satisfy the requirements of CSA-O86. It
is assumed that the novel elastomeric hold-down can provide an elastic hold-down with high
load-carrying capacity for CLT shear wall applications.
2-6

Elastomeric bearing applications

Using elastomeric pads as a structural element in bridge construction began in the mid 20th
century in North America based on AASHTO design provisions [61]. The elastomeric
bearings act mainly as the resilient structural member to transfer loads from the
superstructure of bridges to the substructure. Elastomeric bearings are mainly made of
neoprene and natural rubber in both plain and laminated configurations; as shown in Figure
2-6, steel shims can be used to reinforce the rubber layers, commonly called steel-reinforced
elastomeric bearing.

(a)

(b)
Figure 2-6 (a) plain elastomeric bearing (b) reinforced elastomeric bearing.

The performance of elastomeric bearings relies on the mechanical properties of the rubber
such as the shear modulus. Bearing sliding is an undesired reaction to the shear forces;

27

hence, the design provision recommends limiting shear strain to avoid the sliding of the
elastomeric bearing. Similarly, there is a limitation for the compressive load carried by
elastomeric bearings. To pass the requirements of maximum compressive stress, two factors
should be checked: 1) the compressive deformation and 2) the shear strain at the interface of
the steel shims and rubber pads in the case that they are bonded together. Rubber pads
exhibit negligible volume change under hydrostatic pressure; however, rubber pads with
higher hardness experience more crucial volume changes, and thus, it should be considered
as an effective factor on the performance of elastomeric bearings. Under compression loads,
the rubber pads of elastomeric bearings bulge from the free sides. This specific behavior can
highly affect its compressive deformation. To determine the tendency of the elastomeric
bearing to bulge, a dimensionless geometrical factor is introduced called the “shape factor”,
defined as the ratio of the loaded area to the bulging area, with a higher shape factor showing
the lower bulging capability under compression loads [62].
To evaluate the performance of both plain and reinforced elastomeric bearings, some
experimental studies were done. The shear and compressive behaviors of elastomeric
bearings with different configurations were experimentally studied [63]. The bearings were
tested in both plain and reinforced shape under static and cyclic loading for ambient and low
temperatures. In addition, the test samples were prepared using four different types of
rubbers including neoprene, natural rubber, urethane, and butyl. According to the obtained
results, the stress-strain relationship of the reinforced elastomeric bearing was more desired
than that of the plain form.
The shear modulus of the elastomeric layers was shown to be the main design parameter
rather than the hardness factor for the performance [64]. A wide range of elastomers like

28

neoprene, natural rubber, and others was studied. Both configurations of elastomeric bearings
including the plain and steel-reinforced were tested in a wide range of dimensions. In
addition, the hardness of the elastomeric pads and the number of layers of the reinforced
bearings were taken into account as test variables. The reduction in the shear modulus of
elastomeric bearings with higher hardness was more significant than those ones with lower
hardness under shear fatigue loading. Furthermore, the compressive deformation of
elastomeric bearings was analytically formulated based on an exponential function.
The feasibility of larger elastomeric bearings acting as the support for steel bridges was
evaluated [9]. A comprehensive material test series was performed to characterize the
properties of the elastomeric pads used in the bearing structure for higher displacement
demand, with the material tests’ samples being extracted from the real bearings with different
sizes. Thus, the evaluation of the elastomer’s material properties under large shear
deformation regimes was achieved. The variability of mechanical properties of the
elastomeric bearing was not significantly dependent on the bearing sizes which were used to
prepare the material tests’ samples. To provide the required input for finite element analyses
(FEA) of the elastomeric bearings, the material modeling of the elastomers was created based
on the different models of the hyperelastic theory. Full-scale tests of elastomeric bearings for
high demand applications were conducted [65] and accompanied by FEA on the shear
performance of the high-scale elastomeric bearing in terms of distinct variables such as
aspect ratio and the load direction.
Besides bridge structures, elastomeric bearings have been also employed in building
applications to act as the base isolation against earthquake forces. Regarding the fact that the
base of building acts as a link to transmit the earthquake forces to the building, it was

29

proposed that a specific structural design that can damp the seismic force at the base might
be a promising solution for buildings. Similar to bridge structures, the elastomeric bearing
has been also used as the common base isolation for buildings, as it can decouple the
deformation of the superstructure form the substructure of the building due to the high shear
deformability. However, the elastomeric bearings used as base isolation are more slender
than those employed in bridge structures. Hence, the effect of axial load on the dynamic
performance of isolation bearings, and also the risk of structural instability were considered
as the crucial challenges [66].
Regarding the material properties of rubber, previous studies characterized the non-linear
constitutive behaviour in terms of hyperelastic models such as Mooney-Rivlin [67], NeoHookean [68], Yeoh [69], and Ogden [70]. Curve fitting techniques were employed to
determine the material constants of natural rubbers, high damping rubbers, and elastomeric
bearings’ rubbers using experimental data [71, 72, 73, 74, 75].
2-7

Summary of literature review

Previous experimental research on CLT structural connections can be categorized into three
main groups, component-level scale, wall systems and full-scale building structures. The
results showed that the deformation and released energy are mainly absorbed by connection
systems of CLT walls while their deformation is almost negligible. While hold-down is
mainly resisting against uplifting, angle brackets can be activated for both CLT walls’ sliding
and rocking. Further experimental studies, theoretical methods also confirmed the significant
role of vertical joints on the kinematic of CLT wall systems. Vertical joints are also
considered to act as the main energy dissipation system for CLT-based shear walls.

30

It was found that traditional connections are not the proper choice for tall CLT buildings.
Novel connection systems were introduced for CLT structures to increase the load-carrying
capacity of shear walls. However, most HDs were all designed to behave as an energy
dissipative connection. Elastomeric bearings have been used in the bridge structures and
seismic isolators for tall buildings. Regarding the elastomeric bearings’ concept, the
hyperelastic rubber HD is considered to have high potential.
As discussed in Chapter 1, the main objectives of the research presented in this thesis are to
evaluate the structural performance of a hyperelastic HD solution for CLT shear walls. The
following chapters address the experimental work conducted to determine the: Rubber’s
material properties, specifically its hyperelastic coefficients (chapter 3); required connection
parameters to avoid brittle failure modes in the CLT (chapter 4); and the required geometry
of the elastomeric HD solution for given target performance (chapter 5).

31

3 Chapter 3: Rubber material tests
3-1

Materials and methods

Rubber Masticord pads, produced from a Random-Orientated-Fiber (ROF) elastomeric pads
[76], were chosen as the elastomeric member of the rubber HD system. Due to the specific
features like high compressive strength and cost-effective option, these pads are being used
as a viable structural bearing for short span bridges and precast parking decks. Some relevant
mechanical properties of Masticord are listed in Table 3-1 [76]. Masticord is made of a
polymeric rubber; it contains carbon black, zinc oxide, sulphur, rubber processing oils, and
other chemicals encapsulated in the rubber crumb [77].
Table 3-1 Mechanical properties of Masticord [76].
Shear
modulus
[MPa]

Tensile
strength
[MPa]

Compressive
strength [MPa]

Initial
cracking
strain [%]

1.6 ±0.21

6.9

55.2

40

To evaluate the material properties of the rubber pads, three different test series were
conducted: 1) compression tests to evaluate the compressive performance; 2) shear tests to
evaluate the shear performance, and 3) volumetric tests to help estimate the material
properties of rubber pads.
3-1-1 Compression tests
The objective of the compression tests was to determine the load-displacement behavior of
Masticord under different boundary conditions. The performance of bearing pads relates to
the load distribution and their movement dictated by the applied compressive and shear

32

forces. In general, the bearing pads should be strong in compression and soft in shear. For
compression loads, the rubber needs to provide the required displacement. The maximum
recommended strain for Masticord is 40%; as beyond this limit, the reinforcing fibers start to
de-bond and pull away from the outer surface which results in the rubber failure,
consequently [76].
The shape factor (SF) is the parameter that represents the geometry of the rubber pads, and
defined as the ratio of the loaded area to the free-bulge area, as shown in Figure 3-1.

Figure 3-1 Rubber geometry under compression load.

The shape factor (SF) for rubber pads with rectangular shape can be estimated as [78]:
SF =

d R wR
2( d R + wR ) t R

(3-1)

Where, dR, wR, and tR are the depth, width, and thickness of rubber pads, respectively.
According to the manufacturer design guideline [76], an empirical equation is provided to
predict the stress-strain relationship of the rubber under uniaxial compressive loading:
 c = 6900  (0.6 SF + 2) c1.8

(3-2)

Where,  c and  c are the uniaxial compressive stress and strain on the rubber pad,
respectively.

33

Four different rubber configurations were tested; their geometrical features are listed in Table
3-2. Besides the 40% strain limit, the applied loads calculated using equation (3-2), were
reduced by 50 kN to avoid any damage to the rubber. However, for R3 and R4, the capacity
could not be reached because of the 300 kN maximum capacity of the Universal Test
Machine (UTM).
Three boundary conditions were considered: 1) four sides free; 2) two sides free two sides
restrained along the width direction; and 3) all four sides restrained, as shown in Figure 3-2.
Each specimen was measured in length, width, and thickness on eight points of the surface
using a caliper. The samples were loaded; after unloading, the measurements were repeated.
The test was repeated five times for each specimen. According to EN 12512 [46], the
displacement rate of monotonic test of joints made with mechanical fasteners is
recommended to vary from 0.05 mm/s to 0.2 mm/s. Hence, to characterize rubber behavior
when it is used as the elastomeric member of rubber HD, the compression tests were done
under the displacement control loading with two different displacement rates: a) 2 mm/min;
and b) 10 mm/min for low and high displacement rate tests, respectively.
Table 3-2 Features of four different rubber specimens used in the compression test.
Rubber
specimen

wR

dR

tR

[mm]

[mm]

[mm]

R1

101

180

R2

102

R3
R4

SF

Capacity
[kN]

13

2.5

323

179

27

1.3

254

153

179

22

2.1

454

153

178

27

1.6

409

34

(a)

(b)

(c)
Figure 3-2 Compression test on the rubber specimens under different boundary conditions: (a) 4
sides free (b) 2 sides free - 2 sides restrained (c) 4 sides restrained.

3-1-2 Simple shear test
The double-lap shear test set-up included four steel plates and four rubber pads in a square
shape, as shown in Figure 3-3. Three specimens were prepared; two first specimens had same
dimensions, the average thickness of rubber pads was 13.5 mm while the larger specimen
was made of the average 20.5 mm-thick rubber pads. To meet the requirements of ASTM

35

standard D4014-3 [79], the width of steel plates and the rubber pads were the same, and it
was higher than four times of rubber pads’ thicknesses, 55 and 95 mm for small and large
samples respectively. The geometrical factors of three different specimens are listed in Table
3-3. To bond the steel plates and rubber pads, the glue Loctite produced by Henkel was used
which exhibits a maximum shear strength of 2 and 24MPa when rubber and steel are used as
the substrate, respectively [80]. In preliminary investigation, the glue Loctite produced by
Henkel was found as a recommended adhesive for the rubber-to-metal bonding [81].

(b)

(c)

(a)

Figure 3-3 Simple shear test: (a) specimen sketch (b) 13.5 mm-thick (c) 20.5 mm-thick rubber

Based on ASTM D4014-3 [79], the simple shear test on rubber specimens must be a cyclic
process, with six loading cycles for each test. The tests were conducted using a hydraulic
UTM, see Figure 3-4, and five tests were done with different displacement rates (Lr) per each
specimens. The details of the test series are also provided in Table 3-3.

36

(a)

(b)
Figure 3-4 Simple shear test: (a) set-up (b) loaded sample.
Table 3-3 Simple shear test series

Specimen

#1

#2

#3

tR

wR

dR

[mm]

[mm]

[mm]

13.6

13.4

20.5

55

55

95

Lr

Label

Dtar
[mm]

[mm/sec (in/sec)]

1

13.5

0.254 (0.01)

2

13.5

0.312 (0.0125)

3

13.5

0.38 (0.015)

4

13.5

0.51 (0.02)

5

13.5

1.02 (0.04)

1

13.5

0.254 (0.01)

2

13.5

0.312 (0.0125)

3

13.5

0.38 (0.015)

4

13.5

0.51 (0.02)

5

13.5

1.02 (0.04)

1

20.5

0.38 (0.015)

2

20.5

0.51 (0.02)

3

20.5

1.02 (0.04)

4

20.5

1.52 (0.06)

5

20.5

1.91 (0.075)

55

55

95

37

To achieve 50% as the target strain, the target displacement (Dtar) had to be at least equal to
the thickness of rubber pads, 13.5 mm and 20.5 mm for small and large specimens,
respectively. Based on the geometry of double-lap shear test and the number of shear planes,
the shear strain (γ), and shear stress (τ) were calculated using [79]:
 =

D
2 tR

(3-3)

=

F
2 AR

(3-4)

Where, D, tR, F, and AR are the displacement, the rubber pad thickness, the applied force, and
the area of rubber pads bonded to steel plates, respectively. The sixth cycle of each test must
be considered to calculate the shear strain and stress, according to ASTM D4014-3 [79].
3-1-3 Volumetric test
In the volumetric tests, the rubber samples must be compressed along their thickness
direction while those are only able to deform along the thickness direction; in other words,
the rubber samples must be constrained in two directions. A circular steel ring was
manufactured to act as a rigid container. The outer and inner diameter of the steel ring was
equal to 76 mm and 54.5 mm, respectively. A 54.5 mm diameter steel plug was produced;
the clearance of plug diameter and the inner diameter of the ring was less than 0.1 mm. Both
the steel ring and plug were machined to have the parallel and flat contact surfaces. The
components of the volumetric test set-up and their dimensions are shown in Figure 3-5. To
provide replications for the volumetric test series, six circular samples were prepared using
rubber sheets. While all samples had the same diameter, their thicknesses were varied: 13

38

mm, 22 mm, and 27 mm. To have 10% volume reduction, the target displacement (Dtar) was
considered as 10% of the rubber thicknesses.
To consider the influence of loading conditions, the circular rubber specimens were tested
under two different pre-load conditions as well as two different displacement rates. The two
different pre-loadings were 200N and 400N. The two different displacement rates (and target
displacements) were determined based on the specimen thickness. In this study, the lower
and higher displacement rates for volumetric test series are referred to “low-displacement
rate” and “high-displacement rate”. A hydraulic UTM was employed to do these test series;
low displacement rate was equal to 10% of the target displacements per minute and the high
displacement rate was twice of them.

(a)

(b)

(c)

Figure 3-5 Volumetric tests: (a) sketch, (b) components of the fixture, and (c) rubber samples.

39

To use the volumetric data for evaluation of the material constants, the compression Pv vs.
the volumetric ratio Vr curve must be generated; these data can be evaluated as follows:
F
AR

(3-5)

t −D
Vr = R
tR

(3-6)

Pv =

Where, the terms of F, AR, D, tR are the applied force, the surface area of the rubber
specimen, the axial displacement, and the thickness of the rubber specimen, respectively.
3-2

Hyperelastic coefficients

Rubbers can carry high loads with no permanent deformation. Unlike for linear elastic
materials like steels where the modulus of elasticity can be evaluated as the linear stressstrain curve’s slope according to Hooke’s law, nonlinear stress expressions have been
formulated for elastomeric materials based on the hyperelastic models. The material
constants called hyperelastic coefficients must be found using the experimental data. To
describe the rubber behavior, the strain energy function based on the polynomial hyperelastic
model was developed to consider the state of strain at any point [82]:
N

N

1
( J − 1) 2k
d
k =1 k

W =  cij ( I1 − 3)i ( I 2 − 3) j + 
i + j =1

(3-7)

Where cij and dk are the hyperelastic coefficients and must be determined using the
experimental data. The common tests are the uniaxial, the biaxial, the planar, and the
volumetric test. The first three tests allow evaluating the material constants cij, and the
volumetric test results provide the required data to estimate the constants dk. The volume
ratio J is defined as the ratio of the deformed volume to the original volume. I1 and I 2 are

40

respectively representatives of the first and second strain invariants of the deviatoric CauchyGreen tensor [82].
The Yeoh model [83] was developed based on the reduced polynomial model and it is used
as a common model in FEM packages to simulate the behavior of rubber materials. There are
two benefits related to the Yeoh model; firstly, the rubber behavior under large deformation
would be tracked by the stress-strain expression results more consistent with the
experimental data compared to some other models like the Neo-Hookean and Mooney-Rivlin
models. Moreover, the Yeoh model makes it possible to evaluate the hyperelastic coefficients
using only one of the required test data mentioned before and expresses the stress-strain
relationship for different deformation modes. However, some recent studies showed that the
hyperelastic coefficients could also be evaluated using the simple shear test data [9]. The
strain energy based on the Yeoh model is defined in terms of the first invariant I1 of
deviatoric Cauchy-Green tensor [82]:
3

3

1
( J − 1) 2i
i =1 d i

W =  ci 0 ( I1 − 3)i + 
i =1

(3-8)

In general, the strain energy resulted from the hyperelastic models is a function of the strain
invariants, and the strain invariants are defined based on three principal stretches (λ1, λ2, and
λ3) [84]. The principal stretches show the body deformation in three principal directions; as
those can be defined as the ratio of the deformed element dimension after deformation to the
initial dimension. The relationship of the strain invariant and the principle stretches varies
according to the deformation mode; as the left Cauchy-Green deformation tensor has
different configurations per different deformation mode. According to the left Cauchy-Green
deformation tensor related to the simple shear deformation mode, the first strain invariant is

41

defined as I1 = 3 +  2 , and similarly, the first strain invariant for the bulk compression
deformation is defined as I1 = 2 + 2 [65]. Where, γ and λ are defined as the shear strain and
the stretch (the ratio of the final dimension to the initial dimension) in the axial compressive
direction, respectively. Thus, by differentiating of the strain energy function proposed by the
Yeoh model in terms of the shear strain and the axial compressive stretch the stress
expressions for shear stress (τ) and bulk compression stress (σ) are achieved [65]:
 = 2c10 + 4c20 3 + 6c30 5
 =

(3-9)

2
2
2
( − 1) + ( − 1)3 + ( − 1)5
d1
d2
d3

(3-10)

Where the constant coefficients c10, c20, c03 of the shear stress-strain relationship presented in
equation (3-9) must be estimated using the simple shear test data and likely, the constant
coefficients d1, d2, d3 of another relationship presented in equation (3-10) can be estimated
using the volumetric test data. The curve fitting technique was utilized to fit a degree-5
polynomial function to the shear stress-strain data recorded in the simple shear test. To use
the curve fitting technique for the volumetric test data, the engineering strain which is
defined as  E =  − 1 must be evaluated based on the displacement data recorded during the
volumetric test, so the most fitted degree-5 polynomial function to the bulk compression
stress-engineering strain curve of the volumetric test can determine the constant coefficients
in equation (3-10).
3-3

Finite element analyses

To simulate the rubber behavior based on the hyperelastic coefficients derived from the tests,
a numerical analysis was carried out using the FEM package ANSYS. The simple shear test

42

was modeled with similar dimension of specimens #1. Linear hexagonal 3D elements
(Solid185) were employed to mesh the parts, as shown in Figure 3-6a. The rubber pads were
simulated based on the Yeoh model as one of the hyperelastic models supported by ANSYS.
The steel plates were modeled linear elastic; the interface between steel plates and rubber
pads were chosen as a fully bonded without any relative displacement allowed between them.

(a)

(b)

Figure 3-6 FEM model: a) double-lap shear test; b) volumetric test.

Another FEM simulation was done to verify the hyperelastic coefficient resulted from
volumetric test data. A circular rubber sample was modeled; the thickness and diameter are
27 mm and 54 mm, respectively. The 3D parts were meshed using the 3D hexagonal
elements dominantly, as shown in Figure 3-6b. To avoid high nodal deformation on the
surface elements of the rubber pad under high pressure, the steel plug was also simulated in
contact by the rubber pad; no-separation contact type was chosen to allow the relative
displacement between the nodes of the rubber sample and the nodes of the steel plug. The
displacement of rubber was restrained in all directions except along its thickness to simulate
the effect of a rigid steel ring on the rubber pad.

43

3-4

Results

3-4-1 Compression tests
The load-displacement curve of the compression rubber specimen R4 in terms of three
different boundary conditions is illustrated in Figure 3-7. The individual load-displacement
curves of all rubber specimens are presented in Appendix A.1. The load-displacement curve
is based on the fifth test, as no difference was observed between the tests, see Appendix A.1.

Figure 3-7 Compressive load-displacement of specimen R4 under different boundary conditions.

To describe the effect of two other variables on the rubber compression, the loaddisplacement curves of four-sided free rubber specimen R4 in terms of the different number
of rubber layers (NR) and two various displacement rates are shown in Figure 3-8 and Figure
3-9, respectively.

44

Figure 3-8 Compressive load-displacement of specimen R4 for different rubber layers layout.

Figure 3-9 Compressive load-displacement of specimen R4 under different load speeding.

As shown in Figure 3-7 to Figure 3-9, the effect of the number of layers on the compressive
behavior of rubber pads is higher than that of the two other factors. The boundary condition
impacts the rubber deformation more than the displacement rate. The number of layers and
the boundary conditions which impact the rubber pad shape factor are highly effective on the
compressive behavior of the rubber pads.

45

The compression test results are listed in Table 3-4. The shape factor (SF) of the specimen
has considerably affected the maximum displacement under the compressive load: smaller
SF leads to higher deformations. The maximum displacement of the rubber specimens vs. the
shape factor is illustrated in Figure 3-10; the rubber specimens under two sides free boundary
conditions, and both one layer and two-layers configurations are considered here. The
maximum load capacity of rubber specimen R2 was 200 kN, because of this limitation the
maximum displacements of the other rubber specimens corresponded to the 200 kN
compressive load. A power function describes the relationship between the maximum
displacements versus their shape factor (SF):
Dmax = a  SF b with a = 14.7 ,

b = −1.1

(3-11)

Table 3-4 Compression test results for four rubber specimens.
Low Displacement Rate
Test

R1
R2
R3
R4

NR

load
(kN)

4 sides free

2 sides free

High Displacement Rate

4 sides
restrained

4 sides free

2 sides free

SF

Dmax
[mm]

SF

Dmax
[mm]

SF

Dmax
[mm]

SF

Dmax
[mm]

SF

Dmax
[mm]

1

300

2.50

4.5

3.90

3.9

N/A

2.4

2.50

4.1

3.90

3.4

2

300

1.24

8.7

1.94

7.1

N/A

-

1.24

-

1.94

6.9

1

200

1.20

8.8

1.90

6.8

N/A

2.5

1.20

8.3

1.90

6.7

2

200

0.60

18.7

0.94

14.8

N/A

-

-

-

0.94

14.1

1

300

1.87

7.1

3.48

5.0

N/A

2.7

1.87

6.4

3.48

4.6

2

300

0.93

14.1

-

-

N/A

-

-

-

1.73

10.3

1

300

1.52

8.5

2.83

6.2

N/A

3.1

1.52

7.5

2.83

5.6

2

300

0.76

18.0

1.42

14.0

N/A

-

-

-

1.42

12.9

46

Figure 3-10 Maximum displacements vs. shape factor.

3-4-2 Double-lap shear tests
The load-displacement curves obtained from the double-lap shear tests on three different
specimens are illustrated in Figure 3-11; considering the first displacement rate. The loaddisplacement corresponding to the other displacement rates are presented in Appendix A.2.
According to ASTM D4014 [79], each simple shear test has six cycles; and the sixth cycle
was considered to be used for evaluation of the shear characteristics. To compare the effect
of different displacement rates, the shear strain-stress curves are also illustrated in Figure
3-11 (b), (d), and (f) based on the sixth cycle of each test. The load-displacement curves and
their corresponding shear stress-strain curves obtained by the rubber shear tests were
nonlinear with a similar trajectory to the 3 rd or 5th-degree polynomial functions. Although
each test was done under six cycles, no degradation on the shear strength was found.
The obtained shear stress-strain curves corresponding to different displacement rates were
very similar to each other. In other words, the displacement rate has a negligible impact on
the simple shear tests of the rubber pads. Moreover, the results obtained from specimens #1
and #2 were quite similar, even though their result were slightly different than the results

47

achieved by specimen #3; the shear stress of specimen #3 at 0.5 mm/mm shear strain was
10% lower than that of the specimens #1 and #2.

Figure 3-11 Load-displacement (six cycles) of the simple shear test and the shear stress-strain
curves: (a) and (b) specimen #1, (c) and (d) specimen #2, and (e) and (f) specimen #3.

48

3-4-3 Volumetric tests
The volumetric test series were done using six different circular samples with different
thicknesses (tR) but the same diameter. The variable parameters were the pre-loading and the
displacement rate, see Table 3-5. The target displacement (Dtar) was almost not achieved for
any specimens, because of the used UTM had a 100 kN limit. The maximum displacement
(Dmax) achieved during the volumetric test and the corresponding maximum load (Fmax)
under different loading speed (Lr), also presented in Table 3-5.
Table 3-5 Volumetric test results.
Pre-load [N]
Specimen
label

tR
[mm]

Dtar
[mm]

200

400

Lr
[mm/min]

Dmax
[mm]

Fmax
[kN]

1

25.8

2.6

0.26

2.11

99

2

27.4

2.6

0.26

2.17

99

3

21.7

2.2

0.22

2.2

87

5

21.5

2.2

0.22

2.16

99

4

13.0

1.3

0.13

1.26

76

6

13.2

1.3

0.13

1.1

84

Fmax

Lr
[mm/min]

Dmax
[mm]

[kN]

0.26

2.00

98

0.50

2.03

98

0.26

2.07

98

0.50

1.83

83

0.22

2.03

88

0.45

2.00

98

0.22

2.13

98

0.45

1.94

98

0.13

1.23

99

0.25

1.20

98

0.13

1.16

98

0.25

0.90

74

The effect of the main variable parameters on the rubber pads' behavior under the volumetric
test series was evaluated as the compression-volumetric ratio curve, c.f. Figure 3-12. The
load-displacement curves and the compression-volumetric ratio curves of specimen #5 under

49

two different displacement rates are illustrated. Similarly, these curves under two different
pre-loading conditions are depicted in Figure 3-13. The load-displacement and the
compression-volumetric ratio curves related to the other rubber specimens are presented in
Appendix A.3. To compare the rubber specimens’ behavior against the volumetric test in
terms of various thicknesses, the load-displacement and the compression-volumetric ratio
curves of all six rubber specimens are illustrated in Figure 3-14.
The load-displacement curves obtained from the volumetric test were nonlinear with a
similar pattern to the exponential function; the reaction force was slightly increasing at the
beginning but it sharply increased after a certain displacement value. The displacement rate
was found effective on the reaction force but the pre-loading was not, see Figure 3-12 and
Figure 3-13; as the maximum displacement was decreased up to 10% when the loading speed
was doubled.
According to the compression-volumetric ratio curves shown in Figure 3-14 (b), different
volumetric ratios were achieved in the same compression by different samples; for instance,
the difference between the volumetric ratio corresponding to the specimens #2 and #5 was
almost 20%. In this regard, the average data was calculated based on all results obtained by
six different specimens. In Figure 3-14 (b), the compression vs. volumetric ratio curve based
on the average data is estimated and depicted. The final compression obtained for different
specimens was not the same due to the load capacity of UTM. Thus, to evaluate the average
curve, the target compression was considered equal to the least obtained compression among
all rubber specimens, 33 MPa; the average volume reduction was almost equal to 8.5%.
Eventually, the hyperelastic coefficients dk are estimated in the next section based on the
evaluated average curve shown in Figure 3-14 (b).

50

Figure 3-12 (a) Load-displacement curve and (b) compression-volumetric ratio curve of rubber
specimen #5 under 400N pre-load and different displacement rates.

Figure 3-13 (a) Load-displacement curve and (b) compression-volumetric ratio curve of rubber
specimen #5 under high displacement rate.

Figure 3-14 (a) Load-displacement curves and (b) compression-volumetric ratio curves of all six
rubber specimens together; under the low displacement rate and 200N pre-load.

51

3-5

Evaluation of hyperelastic coefficients

To evaluate the material constants of the rubber pads based on the Yeoh model, the
coefficients must be estimated using the experimental data. For this purpose, the assessment
of the first series (ci0) and second series (dk) of hyperelastic coefficients was performed based
on the experimental data. FEM commercial packages like ANSYS [85] provide curve-fitting
techniques to estimate the rubber-like material constants. However, these FEM packages are
not helpful if the simple shear test data is available, because the deformation of rubber is not
compatible with the direction of principal stretches under simple shear deformation.
Concerning this issue, the curve fitting toolbox of MATLAB [86] was utilized to evaluate the
coefficients of ci0 using the simple shear test data; the average data of three different
specimens were considered. The curve fitting of the simple shear test data based on a 5thdegree polynomial is illustrated in Figure 3-15, and also the shear stress-strain curves of
three different specimens including specimens #1, #2, and #3 and their average curve.
According to equation (3-9), the experimental result obtained from the rubber shear tests
must be fit a 5th-degree polynomial to extract the hyperelastic coefficients. Then, the
hyperelastic coefficients were implemented into the engineering materials section of ANSYS
to create the rubber material model for numerical analysis.
However, the volumetric test data is recognized as the required data to evaluate the
coefficients of volume ratio (dk) of hyperelastic materials in commercial FEM packages.
Thus, the compression vs. volumetric ratio data based on the average curve, c.f. Figure 3-14
(b), was imported to ANSYS and the hyperelastic coefficients of the rubber pads were
evaluated based on the Yeoh-3rd order model, as listed in Table 3-6.

52

Figure 3-15 Curve fitting for the average shear stress-strain data.
Table 3-6 Hyperelastic coefficients of the rubber.
d1

d2

d3

Model

c10
[MPa]

c20
[MPa]

c30
[MPa]

-1

[MPa ]

-1

[MPa ]

[MPa-1]

Yeoh-3rd order

1.15

-1.62

3.46

0.224

2.62e-4

1.2e-5

In Figure 3-16, the shear strain and stress distribution on the rubber pad resulted from 3D
FEM analysis is compared to the experimental shear stress-strain curve. There was a good
agreement between the FEA and the experimental data for the specimen #1. Similarly, the
compression vs. volumetric ratio resulted from 3D FEA is compared to the experimental
results for samples #1 and #2 under 200 N pre-load and low-displacement rate conditions,
see Figure 3-17. The thicknesses of sample #1 and #2 were almost the same; hence, the FEA
results are compared to experiments of both samples #1 and #2, showing a good agreement
between FEA and experimental results for the volumetric test data as well.

53

Figure 3-16 Shear stress-strain based on FEA and experiments of the specimen #1.

Figure 3-17 Experimental compression-volumetric ratio curve and vs. FEM results.

3-6

Summary of rubber tests

To characterize the rubber material behavior, three test series were conducted; the
compression test, the simple shear test, and the volumetric test.
1) The compression tests provided information regarding the compressive performance of
the rubber pads under different boundary conditions. The obtained results by the

54

compression tests made guidance in the design of the rubber dimensions employed for the
rubber HD system. Moreover, the compression test results showed that the shape factor of
rubber pads acted as a significant factor in the compressive behavior of rubber pads, with an
exponential relationship being achieved for variation of the maximum displacement of
rubber pads vs. their shape factor.
2) The load-displacement curve obtained from the rubber shear tests was nonlinear as
expected, with no degradation in strength and stiffness being observed after six cycles. The
displacement rate had no impact on the rubber shear experiments. The load-displacement and
shear stress-strain curves were quite similar for two different samples used in the rubber
shear experiments.
3) The load-displacement curve of rubber samples under the volumetric test was found
nonlinear almost similar to the exponential curve. The volumetric test results were found
slightly sensitive to the loading speed; doubling the displacement rate was able to reduce the
maximum displacement up to 10%.
4) The average data was estimated based on all results achieved by six different specimens
because the maximum volumetric ratios were different for various specimens; up to 20%.
5) The hyperelastic coefficients were estimated based on the Yeoh model and using the
experimental data achieved by the shear test and volumetric test. The numerical results
obtained by FEM had a good agreement with the experimental results.

55

4 Chapter 4: CLT tensile strength tests
4-1

Specimen description

The tensile strength of CLT panels with openings for internal hold-downs is one design
parameter of the rubber HDs. The objective of the tests presented in this chapter was to
evaluate the failure modes and the ultimate tensile strength of CLT panels with openings.
While the size of openings in all CLT panel specimens was the same, 76.2 mm in diameter,
the “loaded end distance” (al) between the center of the hole to the loaded edge of CLT
panels was the main variable of the test series. Based on preliminary estimates, the CLT
panels were prepared with two different widths, 190 mm and 290 mm, to consider the effect
of the “unloaded edge distance”, see Figure 4-1.
a c : edge distance.
a l : loaded end distance
d h : diameter of the centric
opening
w: width of the CLT panel
h: height of the CLT panel
t: thickness of the CLT panel
t l : thickness of the longitudinal
layer.
t t : thickness of the transverse
layer
w la m : width of the lamellas
x: distance of the center of the
opening and the lamellas.

(b)

(c)
(a)
Figure 4-1 Schematic of CLT specimens; (a) side view, (b) top view, and (c) definition of the
parameter x.

56

Some specimens were prepared from previously tested specimens by cutting off the broken
ends and drilling a new set of holes. According to the parameter x, the CLT specimens with
the same location of the centric hole were categorized in the same groups; as there were 11
groups totally. There were three replicates for each group except the groups 6 and 7
(specimens with the centric hole at the 300 mm-loaded end distance and the edge distance
equal to 57 mm) for which the number of replicates was two. A total of 30 specimens were
prepared and subsequently tested. Table 4-2 provides an overview of all tests.
4-2

Materials

In this research, 139 mm thick 5-ply CLT panels were employed with different thicknesses
for longitudinal and transverse layers: 35 mm and 17 mm for longitudinal and transverse
layers, respectively. According to ANSI/APA PRG 320 (2018) [87], the panels were rated as
stress grade E1; the longitudinal layers were made of machine stress-rated 1.8E Spruce-PineFir (S-P-F) lumber, while the transverse layers were made of No.3 S-P-F grade lumber. The
lamella width was equal to wlam = 130 mm. The young modulus E, the tensile strength ft, and
the rolling shear strength fs for the longitudinal and transverse layers are listed in Table 4-1
based on the manufacturer’s technical design guide [88].
Table 4-1 Strengths properties and modulus of elasticity of CLT panels [MPa] [88].
CrossLam
Grade
CLT
139 E

Longitudinal layers

Transverse layers

E

E1M4 12400

ft

fs

E

ft

fs

17.7

0.5

9000

3.2

0.5

57

Table 4-2 Test series overview
Group

1

2

3

4

5
6
7

8

9

10

11

Test
No.

al
[mm]

ac
[mm]

h
[mm]

Mass
[kg]

Density
[kg/m3]

MC
[%]

1

150

57

1160

14.7

479

11.5

2

150

57

775

9.8

479

11.5

5

150

57

591

7.3

465

11.9

4

150

57

1160

14.0

456

13.1

6

150

57

965

11.5

451

13.6

10

150

57

775

9.1

445

12.7

3

150

57

1160

15.6

508

13.4

7

150

57

950

12.6

503

14.1

9

150

57

760

10.0

497

12.3

12

300

57

1460

19.2

497

12.6

14

300

57

1120

14.7

496

12.4

15

300

57

781

10.1

488

12.1

13

300

57

1458

17.7

461

13.2

19

300

57

781

9.2

446

13.0

16

300

57

1460

19.7

510

12.8

20

300

57

909

12.0

498

11.6

17

300

57

1160

14.1

461

13.0

18

300

57

816

9.7

450

13.0

21

150

107

1160

23.6

504

12.6

26

150

107

975

19.6

498

12.6

32

150

107

783

15.6

495

13.1

22

150

107

1163

22.7

484

13.6

25

150

107

975

18.9

482

12.6

27

150

107

786

15.1

477

12.8

24

300

107

1463

29.2

495

11.5

29

300

107

1119

22.0

489

12.0

31

300

107

779

15.2

483

11.8

23

300

107

1462

29.6

501

13.0

28

300

107

1123

22.6

500

12.7

30

300

107

782

15.7

499

12.9

58

x
[mm]
61

30

5

16

28
56
8

11

39

5

37

4-3

Methods

Testing was performed in a hydraulic UTM at a constant rate of displacement of 3 mm/min.
Two steel brackets and two 76.2 mm diameter solid circular steel sections (SCSS) acted as
the fixture of the specimens, see Figure 4-2. As the main purpose of these tests was to
measure the strength, no extra displacement recording devices were employed; therefore, the
displacements reported herein are the actuator head movements. The maximum force Fmax
applied before CLT failure occurred is considered for ultimate strength computation. The
displacement recorded at the maximum force is labeled dFmax.

(a)

(b)

Figure 4-2 (a) schematic of the test set-up (b) the real test set-up.

4-4

Results

The tensile test results of the CLT panels are reported in Table 4-3. The shear failure mode
and the average value of the maximum force F̄max related to each group of CLT panels are
also listed. The load-displacement curve of each CLT specimens and a picture of one of the
failed specimens in each group are shown in Figure 4-3 to Figure 4-12.

59

Table 4-3 Results summary of CLT tension test.
Group

1

2

3

4

5
6
7

8

9

10

11

Test
No.

ac
al
[mm] [mm]

Fmax
[kN]

dFmax
[mm]

1

150

55

188.2

4.3

2

150

55

192.7

4.4

5

150

55

172.6

3.6

4

150

55

119.7

4.0

6

150

55

109.3

3.6

10

150

55

114.4

3.4

3

150

55

209.7

4.4

7

150

55

189.0

3.7

9

150

55

181.7

3.6

12

300

55

278.8

6.0

14

300

55

280.1

5.8

15

300

55

274.8

5.5

13

300

55

211.1

5.6

19

300

55

212.6

4.8

16

300

55

266.4

6.6

20

300

55

271.8

5.2

17

300

55

268.9

6.2

18

300

55

275.2

6.2

21

150

110

177.6

4.4

26

150

110

168.2

3.7

32

150

110

178.8

3.8

22

150

110

123.4

3.8

25

150

110

119.3

3.3

27

150

110

119.8

4.0

24

300

110

285.1

6.1

29

300

110

312.9

7.3

31

300

110

285.2

5.7

23

300

110

249.6

7.6

28

300

110

250.6

5.1

30

300

110

249.8

5.2

60

F̄max
[kN]

x
[mm]

Failure

182.7

61

Row shear

114.9

30

Row shear

193.5

5

Row shear

277.9

16

Row shear

211.8

28

Row shear

269.0

56

Row shear

272.0

8

Row shear

174.8

11

Row shear

120.8

38

Row shear

294.4

5

Row shear
+ net
tension

250.0

37

Row shear

(a)
Figure 4-3 (a) Row shear failure and (b) load-displacement curve of specimens in group 1.

(a)
Figure 4-4 (a) Row shear failure and (b) load-displacement curve of specimens in group 2.

(a)
Figure 4-5 (a) Row shear failure and (b) load-displacement curve of specimens in group 3.

61

(a)
Figure 4-6 (a) Row shear failure and (b) load-displacement curve of specimens in group 4.

(a)
Figure 4-7 (a) Row shear failure and (b) load-displacement curve of specimens in group 5.

(a)
Figure 4-8 (a) Row shear failure and (b) load-displacement curve of specimens in group 6.

62

(a)
Figure 4-9 (a) Row shear failure and (b) load-displacement curve of specimens in group 7.

(a)

Figure 4-10 (a) Row shear failure and (b) load-displacement curve of specimens in group 8.

(a)

Figure 4-11 (a) Row shear failure and (b) load-displacement curve of specimens in group 9.

63

(a)
Figure 4-12 Row shear failure and (b) load-displacement curve of specimens in group 10.

(a)
Figure 4-13 Row shear failure and (b) load-displacement curve of specimens in group 11.

4-5

Discussion

The load-displacement curves for all tested specimens were almost linear before the brittle
failure occurred at low deformation capacity. Not only the strength of the CLT specimens
from the same group were very similar but also the slopes of the force-displacement curves
(as a proxy for stiffness). Thus, the ultimate tensile strength of the CLT specimens with big
openings only depends on the loaded end distance (al), and the edge distance (ac).

64

Although the experimental results showed that the loaded end distance (al) could
considerably affect the strength of the CLT panels, the failure mode of the CLT specimens
was not affected by this parameter. In other words, all specimens were failed under row shear
failure mode regardless of the size of the loaded end distance (al) except the CLT specimens
of group 11 which were failed in the combination of row shear and net tension failure modes.
The existed gaps between the lamellas on the face layers in particular those were adjacent to
the opening acted as the shear plane for CLT panels under tensile loading, see typical failed
samples shown in Figure 4-3 to Figure 4-13. Based on the visual inspection, three different
failure mechanisms were recognized, and the location of opening to the lamellas was the
main effective factor on it. As shown in Figure 4-14 (a) and (b), when either there was no
intersection between the opening and the lamellas’ edge or the lamellas’ edge almost located
in the middle of the opening (symmetrical intersection) the cracks propagation was started
from both sides of the opening. However, when there was an unsymmetrical intersection, see
Figure 4-14 (c), a shear plane already existed on one side of the opening; so one single crack
was enough to cause failure on the CLT panels.

(a)

(b)

(c)

Figure 4-14 Three different failure mechanisms; (a) no intersection, (b) symmetrical intersection, and
(c) unsymmetrical intersection between the opening and the lamellas.

65

The influence of the loaded end distance (al) on the average strength of different CLT
specimens is depicted in Figure 4-15; the experimental results of the groups with almost
similar opening locations are compared to each other to provide reasonable comparisons.

Figure 4-15 Average strength of CLT specimens vs. the loaded end distance (aL).

As shown in Figure 4-15, the effect of the loaded end distance (al) on the strength also relied
on the opening location to the lamella, parameter x. For instance, the tensile strength of the
CLT specimens was almost doubled or more by increasing the loaded end distance (al) from
150 mm to 300 mm when the distance between the center of the opening and the lamellas
was almost equal to the radius of the opening (groups 2 and 5 or groups 9 and 11).
To study the effect of the unloaded edge distance (ac) on the strength of the CLT specimens,
a comparison between the different groups of the CLT specimens with a similar situation
regarding the opening location to the lamellas (x) is presented in Figure 4-16. The parameter
x of the CLT panels and the loaded end distance (al) are considered as the main factors.
Compared to the loaded end distance (al), the strength of CLT panels was less affected by the
unloaded edge distance (ac); as the maximum variation belonged to the average strength of
group 5 and 11 CLT panels (similar x and different ac), and it was almost 20%.

66

Figure 4-16 Average strength of CLT specimens vs. the unloaded edge distance (ac).

The shorter distance between the center of the opening and the lamella (x) made the CLT
Panels more resistant against the tensile load. Besides, the tensile strength of the CLT panels
was increased if there was no intersection between the lamellas and the centric opening: the
parameter x was higher than the radius of the centric opening (x > 38 mm). In general, the
more symmetrical location of the opening to the lamellas is recommended. To exemplify, the
tensile strength of the CLT specimens in group 3 (x = 5 mm) was far higher than those in
group 2 (x = 30 mm). Similarly, this situation was also governed for the CLT specimens with
the opening at the 300 mm loaded end distance. Moreover, the location of the centric
opening to the lamellas was effective on the strength of the wider CLT panels; the longer
distance of x reduced the strength of the wider CLT panel considerably.
4-6

Summary of CLT material tests

To equip the CLT walls with the innovative internal rubber HD system, a large opening
inside the CLT panel is required. Hence, it is essential to evaluate the strength of CLT panels
with large openings. For this purpose, this experimental study was conducted to figure out

67

the CLT specimens’ behavior with symmetrical big centric openings against the tensile
loading. According to the experimental tests, some highly interesting results are achieved:
1. The load-displacement behavior of CLT panels with large centric openings was almost
linear up to brittle failure while low deformation capacity.
2. The CLT panels predominantly failed under the row shear failure mode. A few CLT
panels with longer loaded end distance (al) failed under a combination of row shear and net
tension failure modes.
3. The loaded end distance (al) impacted the ultimate tensile strength of the CLT specimens,
markedly. For wider panels, doubling the loaded end distance (al) increased their strength
around twice while it made the narrower panels stronger around 50%, averagely.
4. The edge distance (ac) had no big impact on the tensile strength of CLT panels; the
maximum increase of strength was almost 18% related to the wider CLT panels with 4d in
diameter openings. However, this factor must be considered in the connection design based
on the size of the openings.
5. The distance of the openings’ center to the lamella, parameter x, highly affects the tensile
strength of the CLT panels; to increase the strength of the CLT panels, two different
approaches are suggested: i) no intersection between the opening and the lamellas edge, or ii)
closer distance between the openings’ center and the lamellas.
6. The distance of the openings’ center to the lamella, parameter x, had a higher impact on
the strength of the CLT panel with shorter loaded end distances than that of the CLT
specimens with longer loaded end distances; the strength was increased around 70% when
the opening was made in the proposed ranges.

68

5 Chapter 5: Hold-down tests
5-1

Specimen description

The elastomeric pads plus steel plates – acting as the main components –were installed inside
the CLT panels and fixed by two steel rods. For this purpose, a rectangular opening was
prepared in the middle of the CLT panel with rounded corners (r = 25 mm) to reduce stress
concentrations. A 50 mm diameter hole to insert the steel rod was drilled into the center of
the panels’ longitudinal cross-section. The CLT panel prepared for the HD test is illustrated
in Figure 5-1.

(b)

(a)

ac: edge distance
al: loaded end distance
h: height of CLT panel
hs: height of the slot
r: rounded corner radius
w: width of the CLT panel
ws: width of the slot
t: thickness of the CLT panel
tl: thickness of longitudinal layers
tt: thickness of transverse layers
wlam: width of lamellas

(c)

Figure 5-1 CLT panel for HD test; (a) sketch of the face view, (b) sketch of the top view, and (c)
photo.

69

The parameters varied were the rubber pad geometry and the number of rubber layers. Both
steel plates and rubber pads were prepared in a rectangular shape, and a 35 mm-diameter
hole was drilled in the middle of them. The parameters to label the dimensions of the steel
plates and rubber pads used in the rubber HD prototype are shown in Figure 5-2.

(b)
(a)
wR (wst): width of rubber pads (steel plates)
dR (dst): depth of rubber pads (steel plates)
tR (tst): thickness of rubber pads (steel plates)
φ: diameter of the centric hole

(c)

Figure 5-2 (a) Sketch of rubber pads and steel plates, (b) steel plate, and (c) rubber pad.

High-strength steel rods were employed to fix the hold-downs inside the CLT panels, see
Figure 5-3. The outside end of the steel rod had a diameter 44 mm to allow attaching the
fixture to the Universal Test Machine (UTM); the inside end of the steel rod had a diameter
of 32 mm to fit through the rubber and steel plates. The steel rods were designed with
sufficient overstrength to transfer the load to the rubber hold down while remain in the
elastic zone.

70

(a)

(b)

Figure 5-3 (a) Technical sketch of the steel rods and (b) real steel rod

Two different CLT panels were used for the hold-down tests; labeled herein group A and
group B. In group A, 175 mm thick panels with equal layer thicknesses for the longitudinal
and transverse layers of 35 mm were used. The CLT panels used in group B were also 5-ply
but with unequal layer thicknesses for the longitudinal and transverse layers (35 and 19 mm,
respectively) for a panel thickness of 139 mm. The slots had a constant height of 300 mm
with a loaded end distance al = 500 mm to avoid brittle CLT failure. Different widths were
prepared, 100 mm, 120 mm, and 140 mm, which resulted in unloaded edge distances ac of
200 mm, and 175 mm and 155 mm for CLT panels used in groups A and B, respectively.
The dimensions of CLT panels used for groups A and B are listed in Table 5-1, in detail.
Table 5-1 CLT prototypes dimensions for groups A and B [mm].
Group

5-2

CLT panel dimension

slot dimension

h

w

t

tl

tt

hs

ws

A

1000

400

175 35

35

300

100

B

1000

450

139 35

17

300

120
140

al

ac

dc

wlam

500

150

45

130

45

130

500

175
155

Materials

Both types of CLT panels of groups A and B were rated as stress grade E1 according to
ANSI/APA PRG 320 (2018) [87]. The longitudinal layers were machine stress-rated (MRS)
2100 1.8E Spruce-Pine-Fir (S-P-F) lumber, while the transverse layers were made of No. 2

71

and No. 3 S-P-F grade lumber. The relevant strength and stiffness properties (tension
strength ft, modulus of elasticity E, compression strength parallel to the grain fc, compression
strength perpendicular to the grain fcp, rolling shear strength fs) of the CLT panels are listed
in Table 4-1 based on the manufacturer’s technical design guide [88]. The Masticord rubber
properties are reported in Table 5-3. The steel plates and steel rods were made of steel grade
AR400F and AISI 4140, respectively. The yield and ultimate tensile strengths for the steel
plates and rods, respectively, were 105 MPa and 125 MPa, and 415 MPa and 655 MPa.
Table 5-2 Strengths properties and modulus of elasticity of CLT panels [MPa] [88].
Longitudinal layers

Transverse layers

CLT

Grade

fc

fcp

fs

E

ft

fc

fcp

fs

175 E

E1M5 12400 17.7 19.9

6.5

0.5

9500

5.5

11.5

5.3

0.5

139 E

E1M4 12400 17.7 19.9

6.5

0.5

9000

3.2

9

5.3

0.5

E

ft

Table 5-3 Mechanical properties of Masticord [76].

5-3

Shear
modulus
[MPa]

Tensile
strength
[MPa]

Compressive
strength [MPa]

Initial
cracking
strain [%]

1.6 ±0.21

6.9

55.2

40

Test series overview

The CLT panel layout and layer thicknesses for the specimen from groups A and B are
shown in Figure 5-4 (a) and (b), respectively. In total, nine CLT panels were prepared and
subsequently tested, two and seven samples for groups A and B, respectively.

72

(b)

(a)

Figure 5-4 CLTs’ layers thicknesses (a) group A and (b) group B

In group A, the rubber pads were sandwiched between two steel plates, c.f. Figure 5-5 (a)
and (b). To reduce the number of required parts, in group B, the elastomeric bearing was
placed directly onto the CLT panel and only one steel plate was inserted, c.f. Figure 5-5 (c)
and (d). The rubber pads were also rounded fit the rounded corners of the slot. Regarding the
number of rubber layers (NR), two different cases were tested: one-layered and two-layered
for both groups A and B.

(a)

(b)

(c)

(d)

Figure 5-5 Rubber layers; (a) one-layered and (b) two-layered rubber pads for group A, and (c) onelayered and (d) two-layered rubber pads for group B.

73

For group A, two dimensions of rubber pads were considered: 100 x 175 mm and 100 x 100
mm, while one single thickness of rubber pad was considered as 27 mm. Four steel plates
were prepared with the dimensions of the bigger rubber pads: 100 x 175 mm, and 25 mm
thick. Two hold-down configurations were considered; (S) single configuration with rubber
pads placed on one side of the slot only; and (D) double configuration with rubber pads
placed of rubber pads on both sides of the slot, effectively testing two hold-downs in one
specimen. In addition, for each geometry (100 x 100 mm and 100 x 175 mm), the number of
rubber layers was varied as one and two layers resulting in two total rubber thicknesses tR,t:
27 mm and 54 mm.
For group B, two different thicknesses of rubber pads (tR): 22 mm and 27 mm; and two
different dimensions; 140x140 mm and 140x120 mm were considered. The number of rubber
layers was varied as one and two layers resulting in four levels of tR,t: 22 mm, 27 mm, 44
mm, and 54 mm. Two steel plates were prepared per each size of rubber pads: 140 x 140 mm
and 140 x 120 mm, with a thickness of 25 mm. It could be mentioned that the nominal
thicknesses of rubber layers provided by manufacturers were 3/4" (19 mm) and 1" (25.4
mm). All rubber pads were t = 139 mm deep while two widths (120 mm and 140 mm) were
tested. All specimens in group B exhibited hold-downs at both ends of the slot (double
configuration).
In total, 53 test specimens were prepared and subsequently tested. The parameter
combinations of the rubber HD system are listed in Table 5-4.
According to the shape factor definition mentioned in chapter 3, and assuming a hole at the
center of the rubber, see Figure 5-6, the shape factor of rubber layers can be calculated by:

74

SF =

d R wR − 

2

(5-1)

4

2(d R + wR ) + t R

Where, dR, wR, tR, and φ are the depth, width, and thickness of rubber pads, and the diameter
of the centric hole, respectively. The total rubber thicknesses tR,t is considered to calculate the
shape factor, listed in Table 5-4, for the two-layered arrangements.

Figure 5-6 Geometry of Rubber layers with a centric hole under compression.
Table 5-4 Overview of tested hold-down arrangements
Group

A

B

Label

dR
[mm]

wR
[mm]

tR
[mm]

NR

tR,t
[mm]

SF

Config.

100-1-S

100

100

27

1

27

1.08

single

100-1-D

100

100

27

1

27

1.08

double

100-2-S

100

100

27

2

54

0.54

single

100-2-D

100

100

27

2

54

0.54

double

175-1-S

175

100

27

1

27

1.98

single

175-1-D

175

100

27

1

27

1.98

double

175-2-S

175

100

27

2

54

0.99

single

175-2-D

175

100

27

2

54

0.99

double

120-1-0.75

140

120

22

1

22

2.06

double

120-1-1

140

120

27

1

27

1.68

double

120-2-0.75

140

120

22

2

44

1.03

double

120-2-1

140

120

27

2

54

0.84

double

140-1-0.75

140

140

22

1

22

2.17

double

140-1-1

140

140

27

1

27

1.77

double

140-2-0.75

140

140

22

2

44

1.09

double

140-2-1

140

140

27

2

54

0.89

double

75

5-4

Test set-up and loading protocols

To separately record the deformations of the two hold-downs in the double configuration,
four Linear Variable Differential Transformers (LVDT) were mounted, one on each face of
the hold-down at the top and bottom of the test specimen, see Figure 5-7. In the single
configuration, only two LVDT were mounted. The applied load was recorded through a
calibrated load cell.
Quasi-static monotonic and cyclic tests were conducted. According to ISO 6891 1983 [89],
the monotonic loading protocol included a pre-loading, and then full loading toward the
target load, see Figure 5-8 (a). All monotonic tests were done under the displacement control
with a rate of loading equal to 5 mm/min. Different monotonic target load (Ftar,m) was
applied to the different cases of the rubber HDs: 100, 150, and 120 for 100 mm depth, 175
mm depth, and 120 mm and 140 mm width rubber HDs. Additional monotonic tests were
conducted for group B under lower target loads (80 kN or 100 kN).

(a)

(b)

(c)

Figure 5-7 (a) Sketch of test set-up; the experimental test set-up for (b) group A and (c) group B.

76

The cyclic tests were conducted under displacement control. The maximum displacements
from the monotonic tests were used to determine the target displacement of the cyclic tests
based on a modified version of “CUREE” method C according to ASTM E2126 [90] to
increase the magnitude of the initial cycles. The primary cycle of each phase had a certain
target amplitude while the three trailing cycles’ amplitudes were 75% of the primary ones.
The amplitudes of primary cycles were estimated based on the specific percentage of target
displacement; see Figure 5-8 (b). For group B, two cyclic tests were done for each holddown configuration with the corresponding maximum amplitude increased from 100% to
125% of the target displacement from the first to the second replicate. The displacement rate
varied between 1 mm/sec and 5 mm/sec) from lowest to highest displacement amplitudes.

Figure 5-8: Component level loading protocols: a) monotonic and b) cyclic loading.

5-5

Hold-down tests to determine the strength of the assembly

The rubber pads as a hyperelastic material are capable of withstanding high loads without
any permanent deformation. While the designer should aim to avoid the catastrophic failure
of the rubber HD, it was still of interest to determine the strength and failure modes of the
selected arrangements under uniaxial tensile loading. As discussed previously, the steel
components of the rubber HD (rods and plates) were designed overstrength to remain elastic.

77

Hence, the brittle failure on CLT panels was expected as the first failure mechanism of
rubber HD assembly, in the case of keep loading beyond the target loads. However, the steel
members can be re-designed to prevent any brittle failure mode on the CLT panels even
when the target load was achieved.
In total, seven CLT panels from group B, three 120 mm width slots, and four 140 mm width
slots, and one CLT panel from group A were tested to failure. It should be noted that six of
the CLT panels from group B were used multiple times in monotonic and cyclic tests and
exhibited partial failure. For 120 mm width rubber HDs, the first specimen (S1-120-2-0.75)
was the only one with no partial failure caused by monotonic and cyclic tests. The second
CLT specimen with the 120 mm width slot was already partially failed, and its rubber layers
arrangement was one layer of 22 mm thick rubber layers (S2-120-1-0.75). The failure test on
the third CLT specimen with the 120 mm width slot was performed while the rubber HD was
equipped with one layer of 27 mm thick rubber layers (S3-120-1-1); the third CLT specimen
was also partially failed. All CLT specimens with the 140 mm width slot were already
partially failed. The arrangement of the rubber HD was the same for all of them; there was
one layer of 27 mm thick rubber layers at both ends of the slot (140-1-1). The only group A
CLT panels tested for the failure was intact. For group A, the steel plate on the top of two
175 mm depth rubber layers (175-2-S) was thinner (12.7 mm) than the steel plates used in
the monotonic and cyclic test series (25.4 mm) but the other set-up characteristics were
similar to the monotonic loading.
The quasi-static monotonic tests were performed under displacement control loading at a rate
of 5 mm/min. Besides, some tests were done with no pre-loading step. It could be mentioned
that the displacement data used in the load-displacement curve come from the recorded data

78

by UTM because the load-carrying capacity of rubber HDs was the main objective, and the
failure on the face layers made that impossible to record data by LVDT continuously.
5-6

Results of monotonic stiffness tests

The results for all HD under quasi-static monotonic loading are reported in Table 5-5. The
deformations at the target loads at the bottom (dbot) and top (dtop) ends are presented; the
average of two LVDTs measurements at the front and back ends of the test specimens
observed when the target load (Ftar,mon) was applied. The load-displacement curves of the
rubber HD system under quasi-static monotonic loading are shown in Figure 5-9.

79

Table 5-5 Monotonic test results for all rubber HDs
Group

A

Label

SF

Ftar,mon
[kN]

Deformation [mm]
dbot

dtop

100-1-S

1.08

100

6.0

-

100-1-D

1.08

100

6.0

5.7

100-2-S

0.54

100

12.5

-

100-2-D

0.54

100

13.7

13.2

175-1-S

1.98

150

3.8

-

175-1-D

1.98

150

3.6

3.7

175-2-S

0.99

150

9.1

-

175-2-D

0.99

150

9.6

9.0

100

2.7

2.4

100

3.2

2.7

120

3.7

3.6

100

2.9

2.8

120

4.3

4.2

100

7.6

7.0

120

8.0

7.8

120

7.6

7.3

100

8.5

6.3

100

8.6

8.8

100

2.7

2.5

100

2.5

2.4

120

3.0

2.7

100

4.5

4.3

100

4.2

4.0

120

4.5

4.3

100

6.4

6.3

120

6.9

6.7

109

9.8

9.0

*

10.3

9.2

120

10.4

9.9

120-1-0.75

120-1-1

120-2-0.75

120-2-1

2.06

1.68

1.03

0.84

B
140-1-0.75

140-1-1

140-2-0.75

140-2-1

2.17

1.77

1.09

0.89

118

*

Partial failure occurred.

80

Figure 5-9 Load-displacement curves from monotonic tests: (a) 100 mm-depth, (b) 175 mm-depth, (c)
120 mm width , and (d) 140 mm width rubber HDs, and (e) one-layered and (f) two-layered rubber
HD in terms of the shape factor.

81

The double configuration (D) delivered two load-displacement curves per test. The
displacement data are based on the average of two LVDTs at the front and back ends of the
test specimens. The other load-displacement curves obtained from the replications of the 120
mm and 140 mm width slot using different CLT panels are presented in Appendix A.4.
According to Figure 5-9, the rubber HDs’ force-displacement followed a non-linear track
from the beginning to the target load, as expected. The target loads for different rubber HD
cases were achieved with no inelastic deformation and damage in the CLT panels even after
repetitions. However, there was an exceptional case for the 140-2-1 rubber HD under the
second replicate, with the brittle failure being occurred; as shown in Figure 5-10.

Figure 5-10 140-2-1 rubber HD with partial failure.

There are two load-displacement curves for each HD with the double configuration, and one
load-displacement curve per HD with the single configurations from group A, see Figure 5-9
(a) and (b). The double configurations’ individual displacements and their stiffnesses were
very similar and also similar to those of the single configuration with the same rubber HD.

82

The largest difference among all arrangements was obtained for the 100-2-D and 100-2-S
configurations; however, this difference was still less than 10%; see Table 5-5.
The maximum displacements achieved by two-layered rubber HDs were more than twice the
one-layered rubber HDs regardless of any other factors. Moreover, using thicker rubber
layers made the 140 mm width HDs more capable of having larger maximum displacements
compared to the 120 mm width ones, see Figure 5-9 (c) and (d).
As shown in Figure 5-9 (e), 100-1-D (lowest shape factor) had the highest displacement
among all one-layered HDs, and 175-1-D (highest shape factor) had the lowest displacement
at the same applied load. Similar behavior was observed for the two-layered HDs; see Figure
5-9 (f). The displacements of the 120 mm and 140 mm width rubber HD were almost the
same, because their shape factors were close to each other.
The deformation mechanisms of the different rubber HD arrangements are shown in Figure
5-11. The free area of rubber layers bulged towards the outside of the specimens. The bulged
shape was almost the same for all arrangements, even though their maximum deflections
varied as a function of the number of rubber layers and the slot’s width as expected
according to the shape factor definition. For two-layered configurations, the bulging
happened in one single shape with the centre line of the two layers experiencing the largest
bulging (similar to the one-layered). A separation at the interface of two rubber layers
happened under higher loads, see Figure 5-11 (h). For specimens from group B, the bulging
area was not as uniform as that of specimens from group A due to the lack of bottom steel
plates, especially for the two layered arrangements, see Figure 5-11 (d) and (h).

83

(a)

(b)

(e)

(f)

(c)

(d)

(g)

(h)

Figure 5-11 Hold-down deformation mechanism under different arrangements of rubber pads: a)
120-1-1, b) 140-1-1, c) 120-2-1, d) 140-2-1, (e) 100-1-D, (f) 175-1-D, (g) 100-2-D, and (h) 175-2-D.

5-7

Cyclic test results

The cyclic test results, specifically the cyclic target displacement (dcyc) determined based on
the monotonic loading displacement (dmon), and the maximum load (Fcyc,max) are listed in Table
5-6. The load-displacement curves of the rubber HD system under both monotonic and cyclic
loading are shown in Figure 5-12 to Figure 5-15; for group B, the second cyclic test series
are considered, and the results obtained by the first cyclic test series are available in
Appendix A.4.

84

Table 5-6 Cyclic test results of all rubber HDs.
Group

A

Ftar,mon

dmon

[kN]

1.08

100-1-D

[mm]

dcyc
[mm]

Fcyc,max
[kN]

Notes

100

8.5

7.5

162.1

-

1.08

100

14.0

15.0

189.5

-

100-2-S

0.54

100

15.2

15.0

143.3

-

100-2-D

0.54

100

29.0

26.0

122.8

-

175-1-S

1.98

150

7.4

6.0

153.0

-

175-1-D

1.98

150

10.7

11.0

242.8

Partial failure

175-2-S

0.99

150

12.7

12.5

198.3

-

175-2-D

0.99

150

22.2

21.0

224.9

-

120-1-0.75

2.06

100

8.3

7.0

114.1

-

8.4

159.9

-

120-1-1

1.68

100

8.7

8.0

104.0

-

9.6

142.2

-

120-2-0.75

1.03

100

17.6

14.0

109.9

-

16.8

155.9

-

120-2-1

0.84

100

18.0

16.0

94.5

-

19.2

138.2

-

140-1-0.75

2.17

100

8.6

8.0

158.6

-

9.6

155.5

Done by partially failed CLT

140-1-1

1.77

100

12.0

10.0

145.2

-

12.0

198.5

Developed cracks

140-2-0.75

1.09

120

17.3

16.0

141.5

Developed cracks

19.2

199.1

Partial failure

20.0

111.9

Partial failure

24.0

153.0

Developed cracks – partial
failure

Label

SF

100-1-S

B

140-2-1

0.89

120

24.9

85

Figure 5-12 Load-displacement curves of 100 mm-depth rubber HD system: (a) 100-1-S, (b) 100-1D, (c) 100-2-S, and (d) 100-2-D under monotonic and cyclic loading.

86

Figure 5-13 Load-displacement curves of 175mm-depth rubber HD system: (a) 175-1-S, (b) 175-1-D,
(c) 175-2-S, and (d) 175-2-D under monotonic and cyclic loading.

87

Figure 5-14 Load-displacement curve of 120 mm width rubber HD system: (a) 120-1-0.75, (b) 120-11, (c) 120-2-0.75, and (d) 120-2-1 under monotonic and cyclic loading.

88

Figure 5-15 Load-displacement curve of 140 mm width rubber HD system: (a) 140-1-0.75, (b) 140-20.75, (c) 140-1-1, and (d) 140-2-1 under monotonic and cyclic loading.

The load-displacement curves obtained for the rubber HD system were non-linear and
different for the loading and unloading cycles. For the double configuration, two separate
hysteresis curves were recorded c.f. Figure 5-12 (b) while one single hysteresis curve is
available for the single configurations c.f. Figure 5-12 (a). Each hysteresis curve included
several cycles with distinct target displacements and stiffness. Each cycle started from zero
to its target displacement and always returned to exact zero displacement at the end of

89

unloading. Thus, no residual displacement was observed in the rubber HD and it remained
completely elastic with no strength degradation, even after multiple loading cycles.
The area between the two non-linear curves of loading and unloading steps depended on the
geometry of rubber layers and their numbers (NR); this area was small for 175-1-S, c.f. Figure
5-13 (a), compared to that of any HD arrangement with two layers, and the double
configuration c.f. Figure 5-12 (d) and Figure 5-13 (d). For the 140 mm width HDs, the area
under the cyclic curves was more affected by thicker layers compared to the 120 mm width
rubber HDs; see Figure 5-15 (a) and (b).
The rubber HD had different behaviour as a function of target displacement and loading
speed. The primary cycles was similar to that observed during the monotonic tests. This
difference in load-displacement trend between cyclic and monotonic tests was less
pronounced in the two layered arrangements, c.f. Figure 5-12 (d), Figure 5-13 (d), Figure
5-14 (d), and Figure 5-15 (d).
For some HD arrangements like 140-2-0.75 c.f. Figure 5-15 (c), the load-displacement
curves of the last cycle are smooth, but experienced a sudden reduction in load. This was
caused by a local or partial brittle failure in the CLT panel, as was reported in Table 5-6;
however, the HD system was still capable of carrying high loads after the partial brittle
failure and returned to the initial position. This partial brittle failure occurred as debonding of
the lamellas on the CLT face layer, as shown in Figure 5-16 (a). Besides, some cracks that
propagated at the rounded corner of the slot, c.f. Figure 5-16 (b), were observed during cyclic
loading. It can also be seen in Table 5-6, the maximum load related to the second cyclic test
series of the 140-1-0.75 rubber HD was lower than that of the first cyclic test series,
exceptionally. To justify, the second cyclic test series of this arrangement was done using the

90

partially failed CLT panels; one of its lamella on the CLT face layer was delaminated from
the adhesive bond line, as shown in Figure 5-16 (c). Due to the lack of complete support, see
Figure 5-16 (d), the loaded area of the rubber layers was decreased. Thus, it caused a
reduction in the stiffness of rubber HD, consequently.

(a)
(b)

(d)

(c)

Figure 5-16 (a) Partial failure on the CLTs’ face layers (b) Cracks at the rounded corners of the slot;
(c) Delaminated lamella, and (d) lack of complete support for the rubber layers.

5-8

Hold-down strength test results

During the cyclic tests, the rubber HD system was capable of carrying high loads even
though –in some cases– the CLT panels developed some cracks or partially failed.
91

Subsequent to the cyclic tests, eight CLT specimens were subjected to quasi-static monotonic
load until failure to determine the maximum load-carrying capacity of the HD system. The
resulting load-displacement curves are illustrated in Figure 5-17. For each specimen, the
curve followed the same path as during the non-destructive tests, specimen 170-2-S reached
the highest load level and exhibited large decrease in stiffness. Failure in all specimens was
characterized by a sharp decrease in load; each drop in the curves was associated with the
failure of one lamella adjacent to the ends of the slot. However, the rubber HD was still
carrying high loads and no complete collapse of the system was observed even under large
deformations.

Figure 5-17 Load-displacement of HD under failure tests (a) group A and group B 120 mm width
slot; (b) group B 140 mm width slot.

The results from the failure tests of both group A and B CLT specimens are listed in Table
5-7. Two loads Fcrack and Fmax are reported; Fcrack identifies the load at which the first brittle
failure occurred, and also Fmax identifies the load-carrying capacity of the CLT panels with
big openings. The displacement at maximum load dF,max of the rubber HDs and the average
capacity F̄max for the CLT panels with two different slots’ widths are also presented in Table

92

5-7. The specimen from Group A failed at 246 kN. The arrangement of rubber HD including
the number of rubber layers and their thicknesses had no impact on Fmax, even though thicker
rubbers were capable of withstanding larger deformations under high loads even after all
brittle failure occurred.
Table 5-7 Results summary of rubber HD failure test using.
Label

widthratio

x

Fcrack

Fmax

dF,max

F̄max

[mm]

[kN]

[kN]

[mm]

[kN]

170-2-S

0.77

N/A

246

246.1

37.7

-

S1-120-2-0.75

0.92

2

178

209.4

-

S2-120-1-0.75

0.92

30

155

191.7

29.0

S3-120-1-1

0.92

30

-

173.8

27.1

S1-140-1-1

1.08

3

153

191.9

24.9

S2-140-1-1

1.08

60

112

140.5

24.9

S3-140-1-1

1.08

3

199

207.5

23.0

S4-140-1-1

1.08

60

118

135.4

34.2

192

169

All failed specimens after the failure test can be found in Figure 5-18. For the group A
specimen, it was observed that the steel plate bent, which caused plastic deformation in the
rubber HD system. The broken parts of the CLT specimen were almost as wide as the slot.
Similar to the CLT strength test presented in chapter 4, it was found that the cracks
developed at the rounded corners of the slot, and then the adjacent lamellas delaminated from
their adhesive bond line. The failure mechanisms were similar to those presented in Figure
4-14. There was the intersection between the lamellas and the slot for all CLT samples. For
the CLT samples with the symmetrical intersection, two cracks were developed before the
failure occurred at the lamellas c.f. Figure 5-18 (a) and (d). For CLT panels with an
unsymmetrical intersection between the slot and the lamellas, one single crack was
propagated at each adjacent lamella and made them separated from the face layers c.f. Figure
93

5-18 (b) and (e). There were some local crushing and bearing on the longitudinal and
transverse layers, see Figure 5-18 (h) and (i). Moreover, the shear deformation of rubber pads
acted as another deformation mechanism; as the broken and crushed parts of CLT panels
started to squeeze the rubber along their thickness, see Figure 5-18 (j).

(a)

(f)

(b)

(c)

(d)

(e)

(h)

(i)

(j)

(g)

Figure 5-18 Failed specimens: (a) S1-120-2-0.75, (b) S2-120-1-0.75, (c) S3-120-1-1; (d) S1-140-1-1,
(e) S2-140-1-1, (f) S3-140-1-1, (g) S4-140-1-1. Local bearing on (h) transverse and (i) longitudinal
layers of CLT panels, and (j) squeezed rubber.

94

5-9

Discussion of monotonic tests

The load-displacement curves of the rubber HD systems under monotonic loading were
nonlinear elastic. The rubber HDs were capable of deforming elastically for a given target
load, with no residual deformation and damage in the assembly. The stiffness of the rubber
HD system was not constant; it was continuously changing, even though the rate of change
decreased for higher loads. As presented in Figure 5-9 and Table 5-5, the contribution of the
bottom and top rubber layers in the double configuration was almost the same. Further, it
was found that the response of the bottom rubber layers was the same in both single and
double configurations in particular for the rubber HDs with higher shape factors.
The load-displacement relied on the geometrical factors of rubber layers and their
arrangements. The monotonic test results showed that the number of rubber layers had a
significant impact on the stiffness of the rubber HD, c.f. Figure 5-9. The number of rubber
layers (NR), the loaded area (AL = wR x dR), and the rubber thickness (tR) affect the shape factor
SF under compression loads. In other words, any factor that changes the shape factor acted as
an effective parameter on the structural behavior of the rubber HD system. The HD
displacements at 50 kN and 100 kN as a function of SF are plotted in Figure 5-19 and a fitted
exponential function d = a · SFb. In general, a smaller SF led to an increase in the
deformation contribution of rubber layers. As a consequence, the rubber HDs with a lower
shape factor are generally able to have higher deformation (or lower stiffness) capacity
before brittle failure occurs.

95

Figure 5-19 Variation of rubber layers’ displacement vs. the shape factor; (a) maximum
displacement at F=50kN and (b) maximum displacement at F=100kN.

5-10 Discussion of cyclic tests
The stiffness of the lower cycles was found almost similar to the stiffness of the rubber HD
under monotonic loading, even though the higher cycles had higher stiffness. Using smaller
shape factor reduces the effect of loading speed on the stiffness of rubber HD under cyclic
loading; c.f. Figure 5-12 (d), the stiffness of rubber HD was almost the same under
monotonic and cyclic loading even for the last cycles. In general, reducing the shape factor
of rubber HD through either increasing the thickness or the number of rubber layers or
reducing the loaded area decreases the amplitude of the highest load regardless of the CLT’s
layer thicknesses.
The cyclic behavior of the rubber HD was significantly different from that of traditional HDs
for mass timber shear walls, exhibiting two different non-linear elastic load-displacement
trajectories for the loading and unloading parts of the cycles. Full capability of returning to
the initial position was achieved and no inelastic deformation remained. No strength
degradation occurred under higher amplitudes, as it is the normal behavior of common hold96

downs, and the rubber HD was also able to resist higher loads with increasing displacements
– that is until the whole assembly reached its brittle CLT capacity. Even when reaching the
brittle failure of the CLT panels, there was no inelastic deformation in the rubber HD system.
In other words, the rubber HD does not dissipate energy through plastic deformation unlike
common HDs.
As a result, the common method of hysteresis analysis, i.e. through creating an equivalent
energy elastic-plastic curve, is not applicable. However, there is no need for such analysis,
because the rubber HD is not supposed to be counted as the dissipation mechanism for CLT
shear walls. However, by adding more rubber layers (increasing the rubber thickness) to the
HD system, the area under the load-displacement curve can be increased. Thus, it can be
interpreted that the rubber HDs with smaller shape factors can absorb more energy than those
with higher shape factor.
5-11 Empirical description of the load-displacement behavior
Statistical analyses were conducted to describe the force-displacement trajectory of the
rubber HD system. As shown in equation (3-2), the maximum compressive stress on rubbers
can be expressed as a function of the shape factor (SF). On the other side, the monotonic test
results showed that the structural response of rubber HD relies on SF and the loaded area
(AL). Hence, an empirical formulation was developed, considering the monotonic loaddisplacement curves of all tested rubber HD systems under double configuration based on the
average of the bottom and top displacements:

97

F = f (d , SF , AL )

(5-2)

Using the curve fitting technique, the load-displacement relation of rubber HD can be
expressed by a power function where the width and the depth of rubber layers need to be
considered in mm:

7.48 AL


F = 1.18 exp(1.22 SF ) −
− 1.35  d 1.87
6
10



(5-3)

The proposed equation was used to plot force-displacement curves for all arrangements of
rubber HD systems and compared to the experimental data, see Figure 5-20.
The proposed formulation makes that possible to evaluate the maximum load-carrying
capacity of the rubber HD system in terms of the geometrical dimensions and also the
number of rubber layers. The evaluated target load should be lower than the maximum loadcarrying capacity of CLT panels to prevent the brittle failure of the rubber HD. Moreover,
the maximum displacement of rubber HD can also be achieved when the maximum target
load is evaluated, thus it is also feasible to design the rubber HD system with specific
geometrical factors so that satisfy the target displacement.

98

Figure 5-20 Estimated load-displacement curves for (a) one-layered, and (b) two-layered rubber
HDs.

5-12 Discussion of strength tests
The load-displacement curves of the rubber HD under the destructive quasi-static monotonic
tests revealed four potential brittle failure modes, with the adjacent lamellas to both ends of
the slot being delaminated sequentially. As a consequence, the delamination of adjacent
lamellas to the slot was the dominant failure mode; however, due to the high compressive

99

strength of rubber layers there was no collapse for the rubber HD even after all brittle failure
modes happened.
Based on the results, the maximum load-carrying capacity of the rubber HD is equal to the
tensile strength of CLT panels, and this should be accounted for in the rubber HD design to
prevent brittle failure. The strength of the CLT panel with a big opening was highly variable;
because different factors can impact on the tensile strength of CLT panels. Besides the
loaded end distance (al), the failure test results showed that the ratio of the slot’s width (ws)
to the lamellas’ width (wlam), called “width-ratio” in this study, is also important. The
average of the strength of CLT panels with the 120 mm width slot was around 15% higher
than that of those with the 140 mm width slot, see Table 5-7. As the CLT panels’ face layers
were made by 130 mm width lamellas, it seems that cutting slots with a smaller width than
the lamellas’ width is a better choice for designing the rubber HD. However, the desired
stiffness should be taken into account in designing the slot’s width, because the wider slot
makes employing the wider rubber pads feasible; as a higher loaded area makes the rubber
HDs stiffer, consequently.
Furthermore, the strength of CLT panels is susceptible to the location of the slot to the CLTs’
lamellas layout which was introduced as parameter x in chapter 4. As shown in Figure 5-21,
if the slot is cut at the middle of two adjacent lamellas (symmetrical intersection) the loadcarrying capacity of rubber HDs considerably increased, in particular for CLT panels with
the 140 mm width slot. If there is a demand to design a rubber HD with high stiffness and
high load-carrying capacity at the same time, it is recommended to employ the CLT panels
made by wider lamellas.

100

According to the results obtained by the group A CLT panel, the steel plate bending shows
that the rubber HD system has also the potential to be designed as a dissipative connection;
the steel plates or steel rods can deform plastically when a given target load is achieved if the
steel members are designed properly.

(a)

(d)

(c)

(b)

(e)

(f)

(g)

Figure 5-21 Slot location to the lamellas; (a) S1-120-2-0.75, (b) S2-120-1-0.75, (c) S3-120-1-1, (d)
S1-140-1-1, (e) S2-140-1-1, (f) S3-140-1-1, (f) S4-140-1-1.

5-13 Summary of hold-down tests
In this chapter, the results of the component level tests of the rubber HD system were
presented and discussed. The rubber HD tests were divided into two different groups, A and
B, as a function of their layer thicknesses. In the stiffness test series, specimens were
subjected to quasi-static monotonic and cyclic loading. Different arrangements of rubber HD
systems were considered; the number of rubber layers, the thickness, and the loaded area of
rubber layers. The results can be summarized as follows:

101

1)

The deformation of the rubber HD vs. the applied load followed a non-linear trajectory

from the beginning of loading (non-constant stiffness).
2)

The rubber HD behaved elastically without any inelastic deformation and no damage in

the CLT panels for the target loads.
3)

Smaller rubber shape factors led to smaller stiffness and higher deformation capacity

before the target load. Higher loaded area decreased the deformation capacity of rubber HD
(higher stiffness). A power function was developed to describe the load-displacement
behaviour.
4)

Under cyclic loading, due to the elastomeric behavior, the rubber HD showed a good

capability of returning to the initial position after unloading without any residual
deformation.
5)

The HD performance was affected by the loading speed: higher loading speed resulted

in higher stiffness. However, smaller shape factors decreased the difference between the
stiffness of monotonic and cyclic tests.
6)

The HD assemblies failed under four sequential brittle failure modes; there was a

deboning failure on the lamellas of the CLT panels’ face layers. No structural collapse was
observed in the rubber HD even after all brittle failures of CLT panels occurred.
7)

Using CLT panels with wider slots reduced the maximum load-carrying capacity of

rubber HDs. The slot’s location to the lamellas was found as an effective factor on the
maximum load-carrying capacity of rubber HD; in particular for the wider slot, a more
symmetrical location increased the load-carrying capacity of rubber HDs around 45%.

102

6 Chapter 6: Conclusion and outlook
6-1

Conclusion

According to CSA-O86 [2], hold-downs should be designed as an elastic connection in CLT
shear wall systems. To satisfy this demand, HDs composed of a hyperelastic rubber layer
were studied. The rubber HD as an internal bearing system includes rubber pads and steel
plates which are fixed inside the CLT panel by steel rods. All steel members were designed
to have no inelastic deformation for a given target load. Therefore, the HD deformations
result from the rubber layers’ deformation.
The experiments were conducted at: 1) the material-level with rubber and CLT tensile
strength test; and 2) the component-level (HD assembly) tests. Three different test series
were conducted to determine the rubber material properties. The load-displacement behavior
of rubber was nonlinear elastic. The experiments showed that a dimensionless factor called
the shape factor determines the rubber layers’ compressive reactions. To characterize the
rubber hyperelastic material properties, the results obtained by simple shear test were used.
Moreover, the rubber layers compressibility was evaluated by conducting volumetric tests.
As a design parameter for the rubber HD, the CLT tensile strength was determined for panels
with a centric opening. CLT panels with big opening failed commonly under row shear
brittle failure mode; as the gap between the lamellas on the face layers acted as a shear plane
under uniaxial tensile load. In addition to the loaded end distance (al) and the edge distance
(ac), the location of the opening to the CLTs’ lamellas (x) affected the tensile strength.
For the component-level test, different arrangements of the rubber HD were tested under
quasi-static monotonic and cyclic loading. Different loaded areas and different rubber

103

thicknesses were considered (different shape factors). Two different types of CLT panels
were employed to take into account the effect of CLT panels’ layout. The results showed that
the performance of rubber HD relied on the compressive behavior of rubber layers; as the
HD load-displacement was similar to the compressive load-displacement of isolated rubber
layers. The rubber HDs exhibited a nonlinear elastic load-displacement trajectory. The
rubber HD assembly stayed in the elastic zone and no residual deformation was remained
even after replications. There was no damage in the CLT panels and higher target loads were
achieved under higher amplitudes during cyclic tests. Further, an empirical formula was
proposed to describe the HD load deformation response
According to the obtained results by this research study, the design procedure of the rubber
HD system is to: i) Define a target load and target displacement; ii) design the CLT panels
with big opening, steel plates, and steel rods with sufficient overstrength to avoid premature
failure; two recommendation for the opening are: 1) no opening wider than the lamellas’
width, and 2) making the opening more symmetrical to the lamellas; iii) define the required
rubber geometry using the proposed empirical formula.
6-2

Future research

To research on the rubber HD system can be extended as follows:
More experiments on the rubber itself can be conducted; more replications on the rubber
shear test and volumetric test using more specimens with different dimensions. Further, the
other test series including the uniaxial, biaxial and pure shear test can be performed to verify
the hyperelastic properties of rubbers. Moreover, numerical modeling can be expanded using
the hyperelastic coefficients achieved by this study.

104

More tensile tests on CLT panels with different thicknesses and layout are required. The
various size of the opening is another variable to investigate; more data makes that feasible
to find a relevant algorithm to predict the CLT tensile strength as a function of the opening
size. Moreover, it also provides an opportunity to investigate the effective factors such as the
width-ratio and parameter x, comprehensively.
More experiments can be performed on the rubber HD system; new test series for higher
target loads, expanded effective parameters, and under different loading conditions. CLT
panels with different layers thicknesses and various dimensions of opening are recommended
as potential future experiments. More variation on the rubber layers’ shape factor provides
larger data-base for statistical analysis. Although an empirical formula was developed based
on the monotonic experiments, more experiments are required to generalize the formula with
better adjustments to the experimental results. It also makes that possible to realize how to
consider the effect of loading speed on the structural performance of rubber HD when it is
subjected to cyclic loading.
New test series using the laminated rubber bearings rather the plain rubber are also
suggested. The laminated rubber bearing provides higher stiffness to the rubber HD. Hence,
there is no need to prepare larger slot inside of the CLT panels for higher loaded area. Two
main goals are predicted: i) the strength of CLT panels with smaller openings would be
higher (higher target load can be achieved) ii) the laminated bearings with lower loaded area
compared to the plain ones can provide higher stiffness for the rubber HD.
To avoid brittle failure on the rubber HD assembly, it is suggested to re-design the steel
members of the rubber HD system to make ductile failure as a dominant failure mode. The
new design can provide an elastic connection with high deformability for a certain target load

105

without brittle failure and then it can dissipate energy though plastic deformation caused by
steel members for beyond the target load. New test series can be conducted by the new
design of rubber HD’s components to evaluate its performance as a two-phased connection,
elastic and inelastic phases, for CLT shear walls.
The full-scale CLT shear walls test can be conducted, with the CLT shear walls being
equipped with the rubber HD to resist the uplifting. For a certain lateral load, the rubber HD
can be designed using the proposed design steps in this research study. Moreover, the shear
key can be used as a viable solution to minimize sliding. To provide dissipator fasteners, the
STS connection or other types of dissipating systems like inserted perforated steel plates can
be selected as the vertical joints. Moreover, the steel tube-type connectors can be other
choice as the energy-dissipation connection for CLT shear walls application.
As the rubber material properties were evaluated based on a hyperelastic model, numerical
analyses of rubber HD systems can be conducted. Then, the optimization analysis using the
numerical results makes that feasible to find the optimum design of the rubber HD assembly
for the CLT shear walls application. Moreover, numerical analysis at the structural level can
be conducted to predict the performance of tall buildings using the proposed rubber HD.
Unlike the common hold-down, the nonlinear elastic rubber HD delivers the non-constant
stiffness to the structural analysis of CLT shear walls, so it is not applicable to use the current
established formulation for CLT shear wall design. In this purpose, new research studies can
be conducted to develop a new analytical design procedure for the shear walls system
equipped with the nonlinear connections such as rubber HD.
Other considerations such as constructability, cost and particularly fire protection of the
rubber HD assembly are suggested as the other suggested research studies.
106

References

[1] M. Green, E. Karacabryli and E. Karch, "Chapter 1 - Introduction," in Technical Guide
for the Design and Construction of Tall Wood Buildings in Canada, Pointe-Claire, QC,
FPInnovations, 2014, pp. 1-12.
[2] Canadian CLT Handbook 2019 Edition, Pointe-Claire, QC, Canada: FPInnovation.
[3] S. Pei, D. Rammer, M. Popovski, T. Williamson, P. Line and J. W. van de Lindt, "An
Overview of CLT Research and Implementation in North America," in World
Conference on Timber Engineering, Vienna, Austria, 2019.
[4] "BC

gov

news,"

British

Columbia

Government,

[Online].

Available:

https://news.gov.bc.ca/releases/2019PREM0024-000383.
[5] National Building Code of Canada, Ottawa, Canada: National Research Council of
Canada, 2015.
[6] C. Loss, T. Tannert and S. Tesfamariam, "State-of-the-art review of displacement-based
seismic design of timber buildings," Construction and Building Materials, vol. 191, pp.
481-497, 2018.
[7] M. Izzi, D. Casagrande, S. Bezzi, D. Pasca, M. Follesa and R. Tomasi, "Seismic
behaviour of Cross-Laminated Timber structures: A state-of-the-art review,"
Engineering Structures, vol. 170, pp. 42-52, 2018.
[8] "CSA O86 (2016), Engineering Design in Wood," Canadian Standard Association.
[9] C. Sun, "Material Study of the Steel Reinforced Elastomeric Bridge Bearings," Ph.D.
Dissertation, the University of Texas at Austin, Texas, US., 2015.

107

[10] D. W. Green, "Wood: Strength and Stiffness," in Encyclopedia of Materials: Science
and Technology, Elsevier Ltd., 2001, pp. 9732-9736.
[11] C. Carll and A. C. Wiedenhoeft, "Chapter 4 - Moisture-related properties of wood and
the effects of moisture on," in Moisture control in buildings: the key factor in mold
preventions, West Conshohocken, PA, ASTM International, 2009, pp. 54-79.
[12] R. H. Falk, "Wood Handbook Wood as an Engineering Material," in Chapter 1 - Wood
as a Sustainable Building Material, Madison, WI: the U.S., General Technical Report
FPL-GTR-190., 2010, pp. 1-6.
[13] APA - The Engineered Wood Association. (2010). A Guide to Engineered Wood
Products. Form.
[14] B. S. Smith and A. Coull, "Chapter 1 - Tall Buildings," in Tall Building Structures, John
Wiley & Sons, Inc., 1991, pp. 1-8.
[15] Wood Design Manual 2017, Ottawa, Canada: Canadian Wood Council.
[16] M. Flaig and H. J. Blaß, "Shear strength and shear stiffness of CLT-beams," in In: 46th
CIB-W18 Meeting, Vancouver, 2013.
[17] R. Brandner, P. Dietsch, J. Döscher, M. Schulte-Wrede, H. Kreuzinger, M. Sieder, G.
Schickhofer and S. Winter, "Shear properties of cross laminated timber (CLT) under inplane load: Test configuration and experimental study," in In: 2nd INTER-Meeting,
Sibenik, 2015.
[18] G. Fink, J. Kohler and R. Brandner, "Application of European design principles to
cross-laminated timber," Engineering Structures, vol. 171, p. 934–943, 2018.
[19] R. Brandner, G. Flatscher, A. Ringhofer, G. Schickhofer and A. Thiel, "Cross laminated
timber (CLT): overview and development," European Journal of Wood Productions,
vol. 74, pp. 331-351, 2016.

108

[20] H. Unterwieser and G. Schickhofer, "Characteristic values and test configurations of
CLT with focus on selected properties," in European conference on cross laminated
timber (CLT). , The University of Bath, Bath, 2013.
[21] H. Ido, H. Nagao, M. Harada, H. Kato, J. Ogiso and A. Miyatake, "Effects of the width
and lay-up of sugi cross-laminated timber (CLT) on its dynamic and static elastic
moduli, and tensile strength," Journal of Wood Science, vol. 62, pp. 101-108, 2016.
[22] H. Blass and T. Uibel, "Tragfähigkeit von stiftförmigen Verbindungsmitteln in
Brettsperrholz.," Karlsruher Berichte zum Ingenieurholzbau, Band 8 , Karlsruhe,
Germany, 2007.
[23] P. Zarnani and P. Quenneville, "New design approach for controlling brittle failure
modes of small-dowel-type connections in Cross-laminated Timber (CLT),"
Construction and Building Materials , vol. 100, p. 172–182, 2015.
[24] B. S. Smith and A. Coull, "Chapter 2- Design Criteria," in Tall Building Structures,
John Wiley & Sons, Inc., 1991, pp. 9-17.
[25] S. M. Hassanzadehshirazi, "Propagation of Uncertainty in Light-Frame Wood
Buildings," PhD dissertation, Unversity of Ottawa, Ottawa, Canada, 2012.
[26] T. Sartori, M. Piazza, R. Tomasi and P. Grossi, "Characterization of the Mechanical
Behaviour of Light-Frame Timber Shear Walls Through Full-Scale Tests," in World
Conference on Timber Engineering, Auckland, New Zealand, 2012.
[27] J. W. Van de Lindt, S. Pei, S. E. Pryor, H. Shimizu and H. Isoda, "Experimental seismic
response of a full-scale six-story light-frame wood building".
[28] R. Tomasi, T. Sartori, D. Casagrande and M. Piazza, "Shaking table testing of a fullscale prefabricated three-story timber-frame building.," Journal of Earthquake
Engineering, vol. 19, no. 3, pp. 505-34, 2015.

109

[29] A. Filiatrault, I. P. Christovasilis, A. Wanitkorkul and J. W. van de Lindt, "Experimental
Seismic Response of a Full-Scale Light-Frame Wood Building," Journal of Structural
Engineering, vol. 136, no. 3, pp. 246-254, 2010.
[30] E. Poirier, M. Moudgil, A. Fallahi, S. Staub-French and T. Tannert, "Design and
Construction of a 53-Meter-Tall Timber Building at the University of British
Columbia," in World Conference on Timber Engineering, Vienna, Austria, 2016.
[31] R. B. Abrahamsen and K. A. Malo, "STRUCTURAL DESIGN AND ASSEMBLY OF
“TREET” - A 14-STOREY TIMBER RESIDENTIAL BUILDING IN NORWAY," in
World COnference on Timber Engineering, Quebec City, Canada, 2014.
[32] M. Popovski, J. Schneider and M. Schweinsteiger, "Lateral Load Resistance of CrossLaminated Wood Panels," in Proceedings of the 11th World Conference on Timber
Engineering , Riva del Garda, Italy, 2010.
[33] M. Popovski and E. Karacabeyli, "Seismic Behaviour Of Cross-Laminated Timber
Structures.," in Proceedings of the 12th World Conference on Timber Engineering,
Auckland, New Zealand, 2012.
[34] I. Gavric, M. Fragiacomo and A. Ceccotti, "Cyclic Behavior of CLT Wall Systems:
Experimental Tests and Analytical Prediction Models," Journal of Structural
Engineering, vol. 141, no. 11, p. 04015034, 2015.
[35] A. Ceccoti, M. Follesa, M. P. Lauriola and C. Sandhaas, "SOFIE PROJECT - Test
Results on the Lateral Resistance of Cross-Laminated Wooden Panels," in First
European Conference on Earthquake Engineering and Seismology, Geneva,
Switzerland, 2006.
[36] M. Shahnewaz, M. Popovski and T. Tannert, "Resistance of Cross-Laminated Timber
Shear Walls for Platform-Type Construction," Journal of Structural Engineering, vol.
145, no. 12, p. 04019149, 2019.

110

[37] D. Casagrande, G. Doudak, L. Mauro and A. Polastri, "Analytical Approach to
Establishing the Elastic Behavior of Multipanel CLT Shear Walls Subjected to Lateral
Loads," Journal of Structural Engineering, vol. 144, no. 2, p. 04017193, 2018.
[38] I. Lukacs, A. Björnfot and R. Tomasi, "Strength and stiffness of cross-laminated timber
(CLT) shear walls: State-of-the-art of analytical approaches," Engineering Structures,
vol. 178, pp. 136-147, 2019.
[39] I. Gavric and M. Popovski, "Design models for CLT shearwalls and assemblies based
on connection properties," International Network on Timber Engineering Research;
2014, INTER/47-15-4, pp. 267-80, 2014.
[40] I. Gavric, M. Fragiacomo and A. Ceccotti, "Cyclic behavior of typical screwed
connections for cross-laminated (CLT) structures," European Journal of Wood
Productions, vol. 73, pp. 179-191, 2015.
[41] A. Hossain, M. Popovski and T. Tannert, "Cross-laminated timber connections
assembled with a combination of screws in withdrawal and screws in shear,"
Engineering Structures, vol. 168, pp. 1-11, 2018.
[42] A. Hossain, M. Popovski and T. Tannert, "Group Effects for Shear Connections with
Self-Tapping Screws in CLT," Journal of Structural Engineering, vol. 145, no. 8, p.
04019068, 2019.
[43] R. Ganey, J. Berman, T. Akbas, S. Loftus, J. D. Dolan, R. Sause, J. Ricles, S. Pei, J. van
de Lindt and H.-E. Blomgren, "Experimental Investigation of Self-Centering CrossLaminated Timber Walls," Journal of Structural Engineering, vol. 143, no. 10, p.
04017135, 2017.
[44] Z. Chen, M. Popovski and A. Iqbal, "Structural Performance of Post-Tensioned CLT
Shear Walls with Energy Dissipators," Journal of Structural Engineering, vol. 146, no.
4, p. 04020035, 2020.

111

[45] "Seismic and Wind Anchor Tiedown System Guide," Simpson Strong-Tie Company
Inc, 2018.
[46] I. Gavric, M. Fragiacomo and A. Ceccotti, "Cyclic behaviour of typical metal
connectors for cross-laminated (CLT) structures," Materials and Structures, vol. 48, p.
1841–1857, 2015.
[47] L.-M. Ottenhaus, M. Li, T. Smith and P. Quenneville, "Mode Cross-Over and Ductility
of Dowelled LVL and CLT Connections under Monotonic and Cyclic Loading,"
Journal of Structural Engineering, vol. 144, no. 7, p. 04018074, 2018.
[48] L.-M. Ottenhaus, M. Li, T. Smith and P. Quenneville, "Overstrength of dowelled CLT
connections under monotonic and cyclic loading," Bulletin of Earthquake Engineering,
vol. 16, p. 753–773, 2018.
[49] W. Y. Loo, P. Quenneville and N. Chouw, "A new type of symmetric slip-friction
connector," Journal of Constructional Steel Research , vol. 94, pp. 11-22, 2014.
[50] W. Y. Loo, C. Kun, P. Quenneville and N. Chouw, "Experimental testing of a rocking
timber shear wall with slip-friction connectors," Earthquake Engineering & Structural
Dynamics, vol. 43, p. 1621–1639, 2014.
[51] P.

Zarnani

and

P.

Quenneville,

"A

resilient

slip

friction

joint".

Patent

WO2016185432A1, 2015.
[52] A. Hashemi, P. Zarnani, R. Masoudnia and P. Quenneville, "Experimental Testing of
Rocking Cross-Laminated Timber Walls with Resilient Slip Friction Joints," Journal of
Structural Engineering, vol. 144, no. 1, p. 04017180, 2018.
[53] A. Hashemi, P. Zarnani and P. Quenneville, "Development of Resilient Seismic
Solutions for Timber Structures in New Zealand Using Innovative Connections,"
Structural Engineering International, pp. 1-8, 2019.

112

[54] B. Murty, A. Asiz and I. Smith, "Tests and models for engineered wood product
connections using small steel tube fasteners," Engineering Structures, vol. 33, p. 63–68,
2011.
[55] J. Schneider, T. Tannert, S. Tesfamariam and S. F. Stiemer, "Experimental assessment
of a novel steel tube connector in cross-laminated timber," Engineering Structures , vol.
177, p. 283–290, 2018.
[56] H. M. Bita and T. Tannert, "Numerical optimisation of novel connection for crosslaminated timber buildings," Engineering Structures, vol. 175, p. 273–283, 2018.
[57] "Wood Innovation and Design Centre, A Technical Case Study," Wood Works, Canadia
Wood Council.
[58] L. Bathon, O. Bletz-Mühldorfer, J. Schmidt and F. Diehl, "Fatigue performance of
adhesive connections for wooden wind towers," Materials and Joints in Timber
Structures, p. 375–380, 2014.
[59] X. Zhang, M. Popovski and T. Tannert, "High-capacity hold-down for mass-timber
buildings," Construction and Building Materials , vol. 164, p. 688–703, 2018.
[60] A. Polastri, I. Giongo and M. Piazza, "An Innovative Connection System for CrossLaminated Timber Structures," Structural Engineering International, vol. 27, no. 4, pp.
502-511, 2017.
[61] "Standard Specifications for Highway Bridges, 16th edition," AASHTO, Washington,
DC, 1996.
[62] J. V. Muscarella and J. A. Yura, "An Experimental Study of Elastomeric Bridge
Bearings with Design Recommendations," Texas Department of Transportation, Austin,
Texas, 1995.
[63] G. T. Suter and R. A. Collins, "Static and Dynamic Elastomeric Bridge Bearing Tests at

113

Normal and Low," Ontario Joint Highway Research Program Report #24, Department
of Civil Engineering, University of Toronto, 1964.
[64] J. C. Minor and R. A. Egen, "Elastomeric Bearing Research, National Cooperative
Highway Research Program Report #109," Transportation Research Board, National
Research Council, Washington, D.C., 1970.
[65] K. V. Belivanis, "Expanding the Use of Elastomeric Bearings for Higher Demand," PhD
dissertation, the University of Texas at Austin, Austin, Texas, 2017.
[66] C. G. Koh and J. M. Kelly, "A Simple Mechanical Model for Elastomeric bearings Used
in Based Isolation," International Journal of Mechanical Sceinces, vol. 30, no. 12, pp.
933-943, 1988.
[67] R. S. Rivlin, Large elastic deformations of isotropic materials. iv: Further developments
of the general theory, Philosophical Transactions of the Royal Society of London, 1948.
[68] G. A. Holzapfel, Nonlinear Solid Mechanics, New York: John Wiley and Sons, 2000.
[69] O. Yeoh, " Characterization of elastic properties of carbon-black-filled rubber
vulcanizates," Rubber Chemistry and Technology, vol. 63, no. 5, pp. 792-805, 1990 .
[70] R. W. Ogden, Non-Linear Elastic Deformations, Mineola, NY, USA: Dover
Publications Inc., 1984.
[71] A. F. Amin, S. I. Wiraguna, A. R. Bhuiyan and Y. Okui, "Hyperelasticity Model for
Finite Element Analysis of Natural and High Damping Rubbers in Compression and
Shear," Journal of Engineering Mechanics, vol. 132, no. 1, 2006.
[72] M. Gajewski, R. Szczerba and S. Jemioło, "Modelling of elastomeric bearings with
application of Yeoh hyperelastic material model," Procedia Engineering 111 , vol. 111,
p. 220 – 227, 2015.

114

[73] J. Yoshida, M. Abe, Y. Fujino and H. Watanabe, "Three-Dimensional Finite-Element
Analysis of High Damping Rubber Bearings," Journal of Engineering Mechanics, vol.
130, no. 5, 2004.
[74] M. Takayama, H. Tada and R. Tanaka, "Finite-Element Analysis of Laminated Rubber
Bearing Used in Base-Isolation System," Rubber Chemistry and Technology , vol. 65,
no. 1, p. 46–62, 1992.
[75] M. Johlitz and S. Diebels, "Characterisation of a polymer using biaxial tension tests.
Part I: Hyperelasticity," Archive of Applied Mechanics, vol. 81, p. 1333–1349, 2011.
[76] "Masticord, Structural Bearing pad Design Guide," JVI, Inc., Illinois, USA., Third
Edition.
[77] "Safety Data Sheet - MASTICORD," JVI Inc., Lincolnwood, Illinois, US.
[78] F. Naeim and J. M. Kelly, Design of seismic isolated structures: from theory to practice,
John Wiley & Sons, Inc., 1999.
[79] "ASTM Standard D4014-3 (Reapproved 2018), Standard specification for plain and
steel-laminated elastomeric bearings for bridges," ASTM International, West
Conshohocken, PA, US.
[80] H.

Co.,

"LoctiteProducts,"

6

4

2018.

[Online].

Available:

https://www.loctiteproducts.com/en/products/fix/superglue/loctite_super_glueliquidprofessional.html.
[81] A. E. B. Martinelli, "Rubber Bearings For Precision Positioning Systems," Master
thesis, Massachusetts Institute of Technology, 2005.
[82] A. Dorfmann and A. Muhr, Constitutive Models for Rubber, CRC Press, 1999.
[83] O. H. Yeoh, "Some forms of the strain energy function for rubber," Rubber Chemistry

115

and Technology, vol. 66, no. 5, p. 754, 1993.
[84] R. S. Rivlin, "LARGE ELASTIC DEFORMATIONS," Springer, 1997, pp. 318-351.
[85] "Ansys Theory Reference, 001242, Eleventh Edition," SAS IP, Inc.
[86] "MATLAB and Curve Fitting Toolbox Release 2019a," The MathWorks, Inc., Natick,
Massachusetts, United States.
[87] "ANSI/APA (2018) PRG 320, Standards for performance-rated cross-laminated timber,"
American National Standard, APA, Tacoma, USA.
[88] "CrossLam CLT technical design guide v2.0," StructuraLam, Penticton, Canada.
[89] "Timber structures; Joints made with mechanical fasteners; General principles for the
determination of strength and deformation characteristics (ISO 6891 : 1983),"
International Organization for Standardization.
[90] "ASTM standard E2126−11 (Reapproved 2018), Standard Test Methods for Cyclic
(Reversed) Load Test for Shear Resistance of Vertical Elements of the Lateral Force
Resisting Systems for Buildings," ASTM International, West Conshohocken, PA, US.

116

Appendix A.1-Rubber compression tests

Figure A.1- 1 Load-displacement curves of (a) rubber specimen R1, (b) rubber specimen R2, (c)
rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression test; one
rubber layer, four sides free, and low displacement rate.

117

Figure A.1- 2 Load-displacement curves of (a) rubber specimen R1, (b) rubber specimen R2, (c)
rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression test; one
rubber layer, four sides free, and high displacement rate.

118

Figure A.1- 3 Load-displacement curves of (a) rubber specimen R1, (b) rubber specimen R2, (c)
rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression test; one
rubber layer, two sides free and two sides restrained, and low displacement rate.

119

Figure A.1- 4 Load -displacement curves of (a) rubber specimen R1, (b) rubber specimen R2, (c)
rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression test; one
rubber layer, two sides free and two sides restrained, and high displacement rate.

120

Figure A.1- 5 Load -displacement curves of (a) rubber specimen R1, (b) rubber specimen R2, (c)
rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression test; one
rubber layer, four sides restrained, and low displacement rate.

121

Figure A.1- 6 Load -displacement curves of (a) rubber specimen R1, (b) rubber specimen R2, (c)
rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression test; two
rubber layers, four sides free, and low displacement rate.

122

Figure A.1- 7 Load -displacement curves of (a) rubber specimen R1, (b) rubber specimen R2, (c)
rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression test; two
rubber layers, two sides free and two sides restrained, and low displacement rate.

123

Figure A.1- 8 Load -displacement curves of (a) rubber specimen R1, (b) rubber specimen R2,
(c) rubber specimen R3, and (d) rubber specimen R4 under five repetition of compression
test; two rubber layers, two sides free and two sides restrained, and high displacement rate.

124

Appendix A.2- Rubber simple shear tests

Figure A.2- 1 Total load-displacement curve (six cycles) of the simple shear test under 0.0125 in/sec
displacement rate: (a) specimen #1 and (b) specimen #2.

Figure A.2- 2 Total load-displacement curve (six cycles) of the simple shear test under 0.0125 in/sec
displacement rate: (a) specimen #1 and (b) specimen #2.

125

Figure A.2- 3 Total load-displacement curve (six cycles) of the simple shear test under 0.02 in/sec
displacement rate: (a) specimen #1 and (b) specimen #2.

Figure A.2- 4 Total load-displacement curve (six cycles) of the simple shear test under 0.04 in/sec
displacement rate: (a) specimen #1 and (b) specimen #2.

126

Figure A.2- 5 Total load-displacement curve (six cycles) of the simple shear test for specimen #4
under displacement rates: (a) 0.02, (b) 0.04, (c) 0.06, and (d) 0.075 in/sec.

127

Appendix A.3- Rubber volumetric tests

Figure A.3- 1 (a) Load-displacement curve and (b) compression-volumetric ratio of all six rubber
specimens under low displacement rate and 400N pre-load.

Figure A.3- 2 (a) Load-displacement curve and (b) compression-volumetric ratio of all six rubber
specimens under high displacement rate and 400N pre-load.

128

Appendix A.4- Monotonic and cyclic tests of rubber HD

Figure A.4- 1 Replications of monotonic tests (a) 120 mm width (b) 140 mm width rubber HDs.

Figure A.4- 2 First cyclic test series (a) 120 mm width (b) 140 mm width rubber HDs.

129