Master’s Thesis in Sustainable Structural

Engineering

Numerical analysis of a
timber-concrete composite floor
system

Author: Negar Ghasemi Khameneh

Supervisors: Wit Derkowski & Osama Abdeljaber
Examiner: Björn Johannesson
Semester: Spring 2024
Course Code: 5BY31E
Linnaeus University, Faculty of Technology
Department of building Technology
a
Acknowledgments
I would like to start by thanking my amazing supervisor, Dr. Wit Derkowski. Your guidance,
unwavering support, and insightful suggestions throughout my thesis have been invaluable.
Your feedback and encouragement have kept me on track and made this project possible. I am
very grateful for all the time and effort you have invested in my work and for proposing this
interesting and innovative topic.

I also want to extend my heartfelt thanks to my other supervisor, Dr. Osama Abdeljaber, for
your invaluable assistance with the FEM modeling and exceptional mentorship. Your patience
and expertise have been crucial in helping me understand and apply these techniques. You were
always there to help whenever I had a question, and your unwavering support has been vital
and indispensable throughout this research endeavor.

I am very thankful to the faculty of the Department of Building Technology at Linnaeus Uni-
versity in Växjö for providing a kind, positive, and professional environment. The resources
and support provided have been invaluable during this journey.

Lastly, I want to express my deepest gratitude to my family to my beloved husband Soheil,

and my dear son Hakan. Your love, support, patience, and belief in me have been my greatest
source of strength. Thank you for always being there for me.

Thank you all for your contributions and support.

b
c
Abstract
Reinforced concrete is one of the most widely used materials for flooring systems due to its
easy availability worldwide, low price, good strength and stiffness, high durability, and its
ability to be formed into any shape and size. However, to prevent the exhaustion of natural
resources and reduce the carbon footprint of cement production, it is becoming necessary to
reduce the use of concrete. To pursue sustainability and implement the findings of the Paris
Agreement, the construction industry is increasingly starting to use timber as an environmen-
tally friendly material. In particular, the use of cross-laminated timber (CLT) has increased
in flooring systems. However, the use of CLT as a system comes with structural challenges,
such as ensuring adequate vibration-related comfort, sound insulation, and fire resistance. To
address these performance shortcomings of CLT or reinforced concrete floors, timber-concrete
composite (TCC) has recently become an increasingly preferred alternative, offering a viable
solution for both the construction of new floor systems and the renovation of existing floors.
TCC floors typically feature a thin concrete slab connected to a timber component via shear
connectors. The timber component is designed to withstand tension and is located on the bot-
tom, while the concrete slab is designed to bear compression and is located at the top. The
connection system transmits the shear forces between the two components. However, design-
ing for disassembly, reducing resource consumption and waste, and decreasing dependency on
virgin resources is crucial. To contribute to this aim, this study focused on numerical inves-
tigations of a novel dismountable TCC floor system comprising prefabricated concrete and a
three-layer CLT connected with notches and screws.
To enable disassembly after the end of service life, wooden dowels were placed in the concrete
layer, and screws were drilled inside the wooden dowels to connect the concrete and CLT.
In this research, concrete and steel’s elastic and plastic material properties were modeled, along
with timber-only elastic material properties. Specimens were modeled in 3D in Abaqus/Standard.
The scope of the research was to analyze the proposed TCC model in the short-term SLS stage,
and subsequently check the bending performance, and composite action behavior of the slab
under a four-point bending test. To achieve this, the research results were first validated and
compared with the gamma method provided in EC5, Annex B. Then, under four times larger
displacement, the performance of the slab was tested, and a parameter study was conducted on
the fiber direction for wooden dowels, the material properties of the wooden dowels, and the
depth of the notch.
Overall, the analyses carried out in the study of the TCC flooring system with connectors to
ensure its future demountability tentatively demonstrated the feasibility of such a solution.
The calculations showed the correct composite action and therefore stiffness of the entire floor
system so that deflections are within an acceptable range. Importantly, it is a solution that fits
in well with the requirements of a circular economy, which could be a competitive advantage
in future years.

Key words: Timber Concrete Composite, Cross-Laminated Timber, concrete, Abaqus, gamma-
method, design for disassembly, effective bending stiffness

d
e
Contents
1 Introduction 1
1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.1 Climate change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 Concrete . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.3 Timber . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.1.4 Timber Concrete Composite Solution . . . . . . . . . . . . . . . . . . 3
1.1.5 Circular economy in TCC floor system . . . . . . . . . . . . . . . . . 3
1.2 Problem description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 Aim & objectives of the thesis . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Scope & limitation of study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

2 Literature information regarding the topic 7

2.1 Self-Tapping Screw Connections . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.2 Dovetail-Notched Connection . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.3 Enhancement of shear resistance at plate-ends in TCC floors using coach screw
connectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Investigation of demountable shear connection . . . . . . . . . . . . . . . . . 11
2.5 Deconstructable connector for TCC floors using self-tapping screws . . . . . . 12
2.6 Benefits of prefabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

3 Theory and Model implementation 14

3.1 Mechanical properties of timber . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.2 Comparison between hardwood and softwood . . . . . . . . . . . . . . . . . . 16
3.3 Shear Connector Systems in TCC Slabs . . . . . . . . . . . . . . . . . . . . . 17
3.4 Modeling with an analytical method based on the Gamma-method . . . . . . . 18
3.4.1 Necessary parameters in Gamma method . . . . . . . . . . . . . . . . 19
3.4.2 Determination of the effective bending stiffness . . . . . . . . . . . . . 19
3.4.3 Determination of the neutral axis of the TCC floor . . . . . . . . . . . 21
3.4.4 Design criterion for Serviceability Limit State (vibrations) . . . . . . . 21
3.4.5 Design criterion for Ultimate Limit State . . . . . . . . . . . . . . . . 22
3.4.6 Remaining Input Parameters in the Gamma Method . . . . . . . . . . . 24
3.5 Finite Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.5.1 Elements and shape function . . . . . . . . . . . . . . . . . . . . . . . 25
3.5.2 Finite element formulation in 3D . . . . . . . . . . . . . . . . . . . . . 26
3.5.3 Finite element formulation of beams . . . . . . . . . . . . . . . . . . . 27
3.5.4 Modeling with Finite Element method, using Abaqus software . . . . . 27
3.5.5 Modeling decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
3.5.6 Modeling procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

4 Analysis of obtained results 47

4.1 Analytical results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
4.2 Finite element results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.2.1 Results for SLS checks . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.2.2 Verification and discussion for uniformly distributed load . . . . . . . 51
4.2.3 Results of four-point bending test . . . . . . . . . . . . . . . . . . . . 53
4.2.4 Discussion for the TCC floor under four-point bending test . . . . . . 58
4.3 Parameter study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

f
 4.3.1 Fiber direction in wooden dowel . . . . . . . . . . . . . . . . . . . . . 59
4.3.2 Depth of the notch . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.3.3 Timber material properties . . . . . . . . . . . . . . . . . . . . . . . . 62

5 Conclusions 64

Future study 66

References 67

Appendix i

g
List of Figures
1.1 Engineered wood products based on sawn timber boards [16]. . . . . . . . . . 3
1.2 Schematic figure of the novel TCC floor slab. . . . . . . . . . . . . . . . . . . 5
2.3 Detailed specimen geometry (Unit: mm) [80]. . . . . . . . . . . . . . . . . . . 7
2.4 Configuration of pushout specimen (unit: mm) [65]. . . . . . . . . . . . . . . 8
2.5 Details of the coach screws (unit: mm)[14]. . . . . . . . . . . . . . . . . . . . 9
2.6 Details of push-out specimens (unit: mm) [14]. . . . . . . . . . . . . . . . . . 9
2.7 Details of bending specimens:(unit: mm) [14]. . . . . . . . . . . . . . . . . . 10
2.8 Illustration of demountable timber–concrete connections: (full-scale model) [31]. 11
2.9 Illustration of demountable timber–concrete connections: (down-scale model)
[31]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.10 The TCC floor model proposed by Derikvand and Fink [23]. . . . . . . . . . . 13
3.11 Normal (σ) and shear (τ ) stresses in different directions in timber [16]. . . . . . 14
3.12 Stress-strain relationship for clear wood under tension parallel to the fiber di-
rection [16]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.13 Stress-strain relationship for clear wood under compression parallel to the fiber
direction [16]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.14 Stress-strain relationship for clear wood under tension perpendicular to the fiber
direction [16]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.15 Stress-strain relationship for clear wood under compression perpendicular to
the fiber direction [16]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.16 Cross-sections and normal stresses distribution based on gamma-method [30]. . 18
3.17 Illustration of 1D element . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.18 Illustration of elements and shape functions in 2D. . . . . . . . . . . . . . . . 25
3.19 Illustration of elements and shape functions in 3D. . . . . . . . . . . . . . . . 26
3.20 Outline of the specimen in Abaqus. . . . . . . . . . . . . . . . . . . . . . . . . 29
3.21 Kent and Park model for confined and unconfined concrete [47]. . . . . . . . . 31
3.22 Response of concrete to a uniaxial loading condition under tension [56]. . . . . 32
3.23 Concrete compressive and tensile behaviour for class C30. . . . . . . . . . . . 33
3.24 Details of the SWC – Countersunk Screw. . . . . . . . . . . . . . . . . . . . . 39
3.25 Bilinear stress-strain curve of steel (Unit of stress: MPa). . . . . . . . . . . . . 40
3.26 Mesh sensitivity analysis for slab. . . . . . . . . . . . . . . . . . . . . . . . . 41
3.27 Finite element mesh of the model. . . . . . . . . . . . . . . . . . . . . . . . . 42
3.28 Illustration of distributed load and boundary conditions. . . . . . . . . . . . . . 44
3.29 Specimen under four point bending test. . . . . . . . . . . . . . . . . . . . . . 45
3.30 Illustration of concentrated load and boundary conditions. . . . . . . . . . . . . 45
4.31 Short term normal stress in concrete. . . . . . . . . . . . . . . . . . . . . . . 48
4.32 Short term normal stress in CLT. . . . . . . . . . . . . . . . . . . . . . . . . . 49
4.33 Short term normal stress in screws. . . . . . . . . . . . . . . . . . . . . . . . . 49
4.34 Short term normal stress in dowels. . . . . . . . . . . . . . . . . . . . . . . . . 50
4.35 Rolling shear stress check in short term. . . . . . . . . . . . . . . . . . . . . . 50
4.36 Tension perpendicular to the grain stress check in short term. . . . . . . . . . . 51
4.37 Normal stress in concrete. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.38 Concrete damage in tension. . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
4.39 Normal stress in CLT. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.40 Normal stress in wooden dowel. . . . . . . . . . . . . . . . . . . . . . . . . . 54
4.41 Normal stress in screws. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

h
4.42 Tension perpendicular to the grain stress check in timber parts. . . . . . . . . . 55
4.43 Rolling shear stress check in timber parts. . . . . . . . . . . . . . . . . . . . . 56
4.44 Linear combination effects of all stresses. . . . . . . . . . . . . . . . . . . . . 57
4.45 Vertical displacement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
4.46 Comparison of different fiber direction for wooden dowels. . . . . . . . . . . . 59
4.47 Linear combination effects of all stresses. . . . . . . . . . . . . . . . . . . . . 61
4.48 Ratio of tension perpendicular to the grain to strength in wooden dowel part. . . 62
4.49 Linear combination effects of all stresses. . . . . . . . . . . . . . . . . . . . . 63

i
1 Introduction
This Master’s thesis addresses the development of a new floor system in the construction in-
dustry that can fit in well with the requirements of circular construction, while at the same
time ensuring that the required load-bearing and serviceability conditions are met. Timber-
concrete composite floors with connectors to ensure their full demountability can be a realistic
alternative to the conventional floor systems currently in use, particularly those based on con-
crete, whose carbon footprint and consumption of natural resources contribute significantly to
environmental degradation.

1.1 Background
1.1.1 Climate change
Climate change, primarily driven by greenhouse gas emissions from both natural processes and
human activities, has led to approximately 1.0°C of global warming above pre-industrial levels,
with projections indicating a potential increase to 1.5°C by 2030 − 2052 if current emission
rates continue, meaning that 0.5°C, in addition, increase to the current situation [6]. In 2018,
the world faced 315 climate-related natural disasters, affecting around 68.5 million people and
causing economic losses of $131.7 billion, mainly from storms, floods, wildfires, and droughts.
Notably, the economic toll of wildfires in 2018 nearly matched the total losses from wildfires
over the previous decade. Sectors such as food, water, health, ecosystems, human settlements,
and infrastructures are particularly vulnerable. To tackle the problem, the Paris Agreement
established in 2015 aims to limit global temperature rise to 2.0°C by 2100, with efforts to
achieve a more ambitious target of 1.5°C [33].
On the other hand, the built environment, including buildings and infrastructures, one of the
fundamental components of economic and social development, involves large quantities of
material and energy consumption. The construction sector extensively relies on a substantial
volume of nonrenewable energy and releases a significant quantity of CO2 [42]. The building
industry contributes approximately 39% to the annual global CO2 emissions [43], and the con-
struction sector is documented to represent over one-third of CO2 emissions and total energy
consumption in both industrialized and developing nations [50] while producing 50% of global
waste [22].
To mitigate building industries’ environmental footprints use of the right material for each
specific situation is crucial. Furthermore, researchers in the environmental impact of building
design prove that up to 75% of the carbon emission of high-rise buildings is from floors [71, 34].
The most popular materials for floor systems are concrete, timber, and steel.

1.1.2 Concrete
Concrete is a vital material in the construction industry worldwide [72]. It is the most com-
monly utilized building material due to its considerably low production and maintenance cost,
durability, availability of constituent materials, high thermal and water resistance, and ability
to be formed into any shape and size [12]. The materials used for producing concrete are also
important in construction technology [51, 58]. As an example, cement is the most extensively
used binding material in plain concrete and reinforced concrete (RC) applications. It is esti-
mated that approximately 6 billion tons of conventional concrete are manufactured annually on
a global scale. However, it’s important to note that the manufacture of ordinary Portland ce-
ment (OPC), the primary binder in concrete, accounts for roughly 8% of global anthropogenic

1
carbon emissions and consumes around 3% of the world’s energy [70, 63, 20].
Considering the environmental implications associated with concrete utilization, the applica-
tion of this construction material should be restricted to essential circumstances. So one way to
prevent depletion of the natural aggregates and simultaneously consider environmental sustain-
ability is to eliminate the dependence on OPC by using aluminosilicate precursors such as Fly
ash (FA), Rice husk ash (RHA), granulated blast-furnace slag (GBFS), metakaolin (MK), in ce-
ment production [70] that are waste materials. Although the replacement can be partial or full,
the produced concrete is called Geopolymers and the binders are alkaline activating solutions
such as potassium hydroxide, sodium silicate, and sodium hydroxide. The use of geopolymer
concrete as an alternative to conventional Portland cement concrete has been found to result
in up to an 80% reduction in embodied carbon, depending on the precursor and activator used
[77], and it is a more significant improvement in sustainability by using alternative activators
and locally available precursors [11].
Regarding the floor system (RC) is one of the traditional choices as a floor for its ability to
provide high load-bearing capacity, minimal deflections, and favorable comfort attributes such
as vibration control, transmission, and acoustic properties all at a relatively low cost [7].

1.1.3 Timber
Timber is one of the environmentally friendly materials due to its carbon sequestration capa-
bility [67]. The term carbon sequestration refers to the process of capturing and storing carbon
that would otherwise remain in the atmosphere or be emitted. However, Timber throughout its
lifespan, sequesters carbon dioxide, which is subsequently released upon reaching the end of
its life cycle, such as during combustion for energy generation.
The lifespan of a structural timber item commences with the growth of the tree, during which it
absorbs carbon dioxide, and obtained logs from trees can be employed in diverse product man-
ufacturing while sequestering carbon, and residual wood that are not usable for the construction
industry or furniture are utilized for energy generation.
In the construction industry timber based on gluing together pieces of sawn timber can be
produced in the form of beams, glulam beams Fig. 1.1a, or in the form of thick panels, often
glued with the sawn timber boards oriented in different directions in different layers, referred to
as Cross-laminated timber (CLT) Fig. 1.1b called Engineered wood products ( EWPs). Notably,
CLT stands out for its low environmental impacts (particularly minimal carbon emissions, and
substantial carbon storage capabilities) and ability to distribute loads in two directions due to
its layered arrangement that alternates directions, making it ideal for use as a floor system
[68]. Its other benefits encompass a visually appealing wood aesthetic, exceptional mechanical
properties across orthogonal directions [31], remarkable thermal insulation quality, and a high
degree of prefabrication. These advantages position CLT as a favored material for construction
purposes [80].
Despite CLT’s numerous advantages, its lightweight characteristic can lead to floor vibration
concerns and, in specific instances, insufficient diaphragm behavior essential for lateral load
resistance [80]. Additionally, when assessing timber structural elements like slabs and beams,
certain drawbacks such as vibration, and limited sound insulation must be considered, par-
ticularly during the serviceability stage [31]. Overall, CLT floors may exhibit unsatisfactory
performance in terms of fire resistance, sound insulation, and vibration control [14].

2
 (a) Lay-up for a Glulam beam. (b) Cross-laminated timber panel with three layers.

Figure 1.1: Engineered wood products based on sawn timber boards [16].

1.1.4 Timber Concrete Composite Solution

In addressing the performance shortcomings of CLT or reinforced concrete floors, Timber Con-
crete Composite (TCC) has emerged as an increasingly favored alternative over recent decades,
offering a viable solution compared to traditional concrete or CLT flooring system options
[31]. TCC floors typically feature a thin concrete slab connected to a timber component via
shear connectors. This timber element can consist of a slab-type engineered wood product or
multiple individual timber beams in a joisted system.
The timber component is usually designed to withstand tension and is located on the bottom
while the concrete slab, is designed to bear compression and is located at the top while the
connection system transmits the shear forces between the two components. This feature of
concrete’s high compressive strength with timber’s tensile strength in parallel to fiber direc-
tion results in a floor system exhibiting better structural performance compared to timber or
concrete-only flooring solutions [23]. If the system is effectively developed, it can demonstrate
a load-bearing capacity three to five times greater than that of conventional timber or concrete
systems [31].

1.1.5 Circular economy in TCC floor system

A circular economy (CE) has the potential to reduce resource consumption and environmen-
tal impacts in the construction sector [54]. The concept of CE is related to changing patterns
of production and consumption with the aim of recirculating materials, such as recycling or
reusing, to reduce waste and decrease dependency on virgin resources [32]. Design for Disas-
sembly (DfD) is an approach that aims to promote the reuse of materials and change the use of
materials from linear to circular [59]. DfD encourages the design of structural elements with
reflection on their end-of-life scenario in addition to their load-bearing capacity and durabil-
ity [10]. DfD elements can be demounted, reused, or recycled with limited energy, time, and
resource consumption.
To develop new structural solutions, it is critical to consider two key factors. The first fac-
tor is the promotion of usage of fewer construction materials, while the second factor is the

3
incorporation of DfD [46].
To design a TCC floor system for circularity, connections in slabs are crucial and should be
designed in a way allowing for dismantling at the end of the service life of the structure to
adhere to circularity principles. The role of connectors is to connect all layers of the slab to
act as one structural element and transfer shear between layers. Connections directly influence
composite behavior, stress distribution, deformations, and slab design [37].

1.2 Problem description

Connections in TCC floors are categorized into two types: wet-dry and dry-dry systems. In
wet-dry systems, the concrete layer is directly cast against the timber part on-site. This method
requires additional propping and formwork for pouring the wet concrete onto the timber, sig-
nificantly slowing down the construction process [57]. Typically, wet-dry systems involve the
use of permanent connectors, making the timber components and concrete slab inseparable,
which limits recycling or reusing materials at the end of their service life.
Conversely, dry-dry systems employ prefabricated components to increase construction speed
and enable disassembly. In this system, the concrete part is precast with the shear connection
embedded inside. It is then fastened to the timber component using screws or bolts, enhancing
the slab’s ability to be dismantled and promoting sustainability.
Several studies have aimed to develop disassemblable connectors for TCC floors [48, 49, 36].
This area of research primarily involves conducting experimental analyses of TCC, utilizing
various connection types frequently within wet-dry systems. However, investigations in dry-
dry systems are focused on enabling disassembly, utilizing methods such as notch systems [31],
pre-casting holes in concrete, applying steel connections, and incorporating fillers like epoxy
glues to fill gaps between the fastener and concrete [57].
Sustainability concerns linked to connectors necessitate additional research on designing TCC
floors for disassembly to promote eco-friendly construction solutions [23]. Connectors de-
signed for deconstruction facilitate the dismantling and recycling of composite floors once they
reach the end of their service life [49]. Because of these reasons, this study on a TCC floor
system will be conducted within a dry-dry system, focusing on the types of connectors that
increase the dismantling ability of the slab.

4
1.3 Aim & objectives of the thesis
Developing innovative structural solutions requires considering two essential factors: advanc-
ing the utilization of sustainable construction materials and designing structural elements for
easy deconstruction at their end-of-life.
In pursuit of this goal, the study aims to enhance understanding regarding the design of an
innovative TCC floor system. This system comprises a three-layer CLT and a prefabricated
concrete slab, connected with both notches and disassembly screw fasteners from a structural
engineer’s perspective.
To achieve a highly prefabricated TCC floor system with disassembly capabilities, there are
two notches close to the ends of the slab and two cut-outs in the CLT panel, into which the tabs
of the precast concrete unit are geometrically perfectly aligned. This joint can be classified as
a notched joint in the transfer of shear forces between the components, and at the same time
allows correct positioning of the precast concrete element on the CLT panel during assembly.
Because of the uncertainty of the joint’s performance, such as potential manufacturing inac-
curacies of the precast components, the longitudinal shear resistance is additionally ensured
by mechanical screw connectors. Full demountability is ensured by the use of screws that are
driven into wooden dowels inserted into the precast concrete units and then into the CLT panel.
The wooden dowels, which are one of the innovations of this solution, also ensure that there is
no gap between the precast concrete element and the demountable screw. A general figure of
this model is presented in Fig. 1.2.
To contribute to this understanding, the thesis investigates various design solutions across the
following aspects:

• Different timber type for wooden dowel and its fiber direction.

• Change in the depth of the notch.

Objective of this study is the analysis of this novel TCC floor system in the short-term Service-
ability Limit State (SLS). Furthermore, validation of composite action by using the analytical
method (gamma method), is provided in Annex B of Eurocode 5 [4].

Figure 1.2: Schematic figure of the novel TCC floor slab.

5
1.4 Scope & limitation of study
This section outlines the boundaries of the research and is summarized as follows.

• Structures primarily utilized as floor slabs are included in the consideration, excluding
roof slabs and similar structures with an inclination and varying load distribution over
the span.

• The focus is solely on one type of cross-section.

• The study is limited to a single-span, one-way, and simply-supported configuration.

• Long-term behavior, fire design, and accidental loads are not within the scope of this
study.

• The investigation is centered only on the serviceability limit state of the slab in the short
term.

• This study is conducted independently of any physical experimentation.

• Local Phenomena around supports and concentrated loads are not taken into account
within the analysis.

The limitations of this study include modeling the timber material in a scenario of elastic de-
formations, which makes it difficult to determine the failure of the TCC floor system. This
limitation was addressed by modeling the elastic and plastic material properties of concrete
and steel, followed by the elastic material properties of timber, considering that timber exhibits
brittle behavior in tension parallel and perpendicular to the fiber direction. Since the CLT part
is in the bottom layer and tension dominates, the passing of stress can lead to strength failure,
which can be observed in this model design.

6
2 Literature information regarding the topic
A literature review was conducted to explore the regulations, various options, and recommen-
dations concerning the TCC floor system.
Most studies on floor system design involve comparing models using either an analytical
method or a finite element model with a physical experiment. However, many recent inves-
tigations have primarily focused on the shear and bending aspects of TCC floors, neglecting
the critical aspect of disassembly. Some of these investigations, summarized in subsections
2.1 to 2.3, overlook the importance of addressing disassembly, which is crucial for enhancing
sustainability and maximizing the end-of-life environmental benefits of a floor system. Only
a limited number of studies have explored the disassembly characteristics of TCC floors, and
some of these are outlined in subsections 2.4 to 2.6.

2.1 Self-Tapping Screw Connections

The mechanical performance of self-tapping screw connections in composite floors made of
CLT and concrete is examined in Wang et al.’s [80] study. They conducted laboratory tests
on twenty-four specimens, subjected to direct shear under monotonic loading. Each speci-
men, measuring 600 mm2 , comprised a 60 mm thick reinforced concrete layer atop a 105 mm
thick CLT panel (see Fig. 2.3), connected with two self-tapping screws. Various aspects such
as screw configuration, geometry, and thread type were varied for analysis. Failure modes at
the CLT-concrete interfaces were examined using endoscopy. Screws subjected to shear ten-
sion were withdrawn from the CLT panels, while those experiencing shear compression led to
yielding in the CLT panels due to flexure. Additionally, failures such as timber embedding and
concrete crushing occurred around screws loaded in shear compression. Analysis revealed that
slip stiffness was primarily influenced by screw configuration, inclination angle, and thread
type, while maximum capacity was affected by screw configuration, inclination angle, and
penetration length into the CLT panel.

Figure 2.3: Detailed specimen geometry (Unit: mm) [80].

2.2 Dovetail-Notched Connection

Ouch et al. [65] introduced a specific notched connection called the dovetail-notched connec-
tion for TCC slabs. This notched shape, along with V-shape reinforcement inside the notched
connection as depicted in Fig. 2.4, offers mechanical locking to restrict uplift without the need
for metallic screws. The study comprised two parts: an experimental pushout phase followed
by a numerical investigation into the impact of various parameters related to geometry and

7
material properties on the performance of the notched connection. The experimental pushout
phase revealed significant stiffness and resistance. However, it also indicated a lack of ductility
due to CLT failure by rolling shear. Two distinct failure modes emerged: concrete shear at the
notch and rolling shear of the cross-layer of the CLT panel. Avoiding the rolling shear failure
of the CLT panel is crucial for enhancing the post-peak performance of the notched connec-
tor. Reducing notch length and increasing heel length can induce concrete shear failure at the
notch, thereby enhancing connector ductility if adequate V-shaped rebars are placed inside the
notch. The stiffness of the connector can be adjusted by altering parameters such as concrete
resistance, concrete panel thickness, amount of V-shaped rebars, notch length, and heel length.

(a) Detail of specimen for push-out test. (b) Notched connector.

(c) V-shape rebar cage.

Figure 2.4: Configuration of pushout specimen (unit: mm) [65].

2.3 Enhancement of shear resistance at plate-ends in TCC floors using coach screw con-
nectors
Bao et al. [14] evaluated the shear and bending performance of coach screw connectors in
a TCC floor. To design specimens, reinforced bars (diameter = 8 mm) were used to form a
steel mesh framework in all specimens to avoid unpredictable brittle failure of concrete slabs
during the loading process. Since specimens were made in a wet-dry process, waterproof wax
was applied between the CLT panel and the concrete slab to prevent moisture transfer into the
timber. For all specimens, three-layer CLT panels were used (with each layer being 35 mm
thick), and the thickness of concrete slabs was designed as 70 mm. Four coach screws Fig 2.5
were installed in each push-out specimen and the only variable parameter for the two groups
of specimens was the angle (45° and 90°) of the screw embedded into the CLT panel. Fig 2.6
illustrates the details of the specimens for the push-out test.

8
 Figure 2.5: Details of the coach screws (unit: mm)[14].

(a) (b)

Figure 2.6: Details of push-out specimens (unit: mm) [14].

The results of the push-out test showed that vertical coach screws exhibited a typical double
plastic hinge bending failure. In contrast, inclined coach screws under tension-shear load did
not exhibit obvious bending deformation but were instead pulled out from the CLT panel, re-
sulting in brittle localized shear failure of the CLT panel. A comparison between the two groups
of specimens revealed that the CLT panel in (b) showed a more ductile failure mode. Regard-
ing the concrete panels, group (a) showed a small longitudinal crack, but the concrete slab in
(b) showed a full-length transverse crack. Therefore, vertical screws were considered more
suitable for the pure shear zone at the plate-end of the TCC floor for the following reasons:

• improving the structural ductility of the floor

• avoiding exacerbating the cracking of the concrete slab in the pure bending zone of the
TCC floor.

Results from the bending test showed that all bending specimens exhibited similar brittle failure
modes. However, a comparison between (a), (b), and (c) revealed that the plate-ends of (a)
and (b) exhibited an obvious uplift effect. This phenomenon demonstrated that increasing the
number of shear connectors at the plate end can effectively prevent the occurrence of the uplift

9
effect. It was reported that the first and second layers at the bottom of the CLT panel in the
pure bending section experienced tensile failure and rolling shear failure, respectively, while
the third layer showed almost no obvious failure. Specimens for bending tests are illustrated in
Fig. 2.7.

(a)

(b)

(c)

Figure 2.7: Details of bending specimens:(unit: mm) [14].

10
2.4 Investigation of demountable shear connection
In a paper, Eslami et al. [31] experimentally and numerically investigated the load-bearing
behavior of a down-scaled model of a demountable prefabricated TCC floor system. They
explained that this system was designed using a wet-dry process, with concrete cast onto the
timber part. After casting, the TCC module can be transferred to the building site and installed
using a dry-dry process. The TCC floor system comprises a concrete layer on a timber layer
connected via a demountable shear connection. The timber part can take the form of joists or a
timber slab made from engineered wood products. The shear connection included a triangular
notch to transfer shear force between the two layers without forming a permanent bond, with
vertical bolts used in notched connections to connect the layers together and prevent concrete
uplift. The bolts were designed to operate within the elastic zone to avoid permanent defor-
mation, facilitating deconstruction and reuse. Before concrete casting, notches and bolt holes
in the timber would be cut, and formwork for the bolts in the concrete would be placed. After
concrete hardening, the bolt formwork would be removed, bolts installed, and the TCC module
completed. These modules were connected to primary beams with simple supported supports.

Figure 2.8: Illustration of demountable timber–concrete connections: (full-scale model) [31].

However, they modeled the down-scaled concrete layer with a thickness of 10 mm, which is
insufficient for conventional aggregate sizes. Therefore, a special self-compacting mixture with
fine aggregates and high tensile strength was developed. Additionally, the placement of steel
reinforcement in the concrete layer was impossible. A mixture of water, high-performance
binder, and superplasticizer was utilized to develop the special concrete mixture for the down-
scaled element. One shear connection was placed on each side of the slab, with the timber
board sourced from the C24 strength class, as presented in Fig. 2.9.

11
Figure 2.9: Illustration of demountable timber–concrete connections: (down-scale model) [31].

2.5 Deconstructable connector for TCC floors using self-tapping screws

Derikvand and Fink [23] proposed an innovative deconstructable screw connector suitable for
both wet-dry and dry-dry construction systems. The connector comprises a self-tapping screw
with an upper section featuring a thin protective layer and a reusable rubber lid. To prevent
direct interlocking with concrete, the screw’s main body was covered. Additionally, it was nec-
essary to protect the screw head from being covered by fresh concrete during casting so that
the screw head could be accessed for disassembly purposes. In wet-dry systems, the connector
was driven into the timber before concrete casting, after which the concrete was cast on top,
similar to permanent connectors. In dry-dry systems, the connector was inserted into the con-
crete formwork, removed after the concrete had dried, and then the concrete slab was placed
on top of the timber member, with connectors driven into the timber through a hole inside the
concrete section.
For the disassembly process, the screw could be accessed and removed through the open hole
of the connector in the concrete section, which was covered by the rubber lid during casting.
The protective layer remained inside the concrete after twisting and removing the screws. Both
wet-dry and dry-dry systems are depicted in Fig. 2.10.
Experimental investigations examined single and cross-paired screw arrangements with varied
insertion angles and screw types, including failure analysis and comparison with permanent
connectors. However, this study did not cover the stiffness test of reassembled deconstructed
screw connections, as it can lead to reduced stiffness due to the looseness of the screw in the
timber section after reassembly.

12
(a) Application steps and disassembly of the deconstructable con- (b) prefabricated concrete and glulam
nector in the wet-dry system. members and assembling the elements.

Figure 2.10: The TCC floor model proposed by Derikvand and Fink [23].

2.6 Benefits of prefabrication

Dry construction systems are ideally suited to the realization of floor slabs with DfD character-
istics, meaning they can be easily dismantled in the future. Such solutions are based on the use
of precast concrete components as part of a composite element, which improves environmental
performance by better adapting the material and geometric properties of the precast component
to the requirements of the floor and enhancing the energy efficiency of the production process.
The implementation of prefabricated solutions can improve social impacts, such as increasing
the health and safety of workers, mitigating the impact of construction processes on local res-
idents by reducing construction inconvenience time (noise and dust reduction), and lowering
global costs while shortening the construction time [9, 24]. Derkowski’s study highlights [25]
that floors made from prefabricated concrete elements offer a wide selection of available tech-
nologies. These technologies provide significant advantages in terms of implementation costs,
versatility, fire resistance, and insulation and soundproofing properties.
A study by Jaillon et al. [44] suggests that precast concrete construction projects can achieve
a significant reduction in construction waste of 52%, including reducing the use of timber
formwork by more than 70%. As highlighted [8], the external benefits of precast structures go
beyond environmental issues and include social and economic perspectives. Further research
into the use of prefabricated components in a TCC floor system, with connections that have
disassembly properties allowing TCC floors to be separated into individual parts, is justified
and is the aim of this study.

13
3 Theory and Model implementation
This section covers basic knowledge about the mechanical properties of timber, connections,
and analytical and numerical models.
The analytical method that has been used for the scope of this research is the Gamma method for
the composite action evaluation, given in the Eurocode 5 Annex B [4]. The analytical method
was implemented in Mathcad Prime 7.0.0.0 as a template for hand calculation and presented in
detail in Appendix A.
Finite element modeling was employed in this thesis, utilizing the commercial software Abaqus/CAE
2022.

3.1 Mechanical properties of timber

From a structural standpoint, timber is characterized as an anisotropic material, meaning its
mechanical properties vary across all three spatial dimensions [60]. Fig. 3.11 illustrates stresses
in timber in different directions [16].

Figure 3.11: Normal (σ) and shear (τ ) stresses in different directions in timber [16].

Regarding directions in timber the longitudinal direction is parallel to the grain and height of
trees, while the radial direction extends from the annual rings to the pith (center of the tree), and
the tangential direction follows the annual ring tangent. Stress in parallel direction is denoted
by σ0 or σ// , and perpendicular to the fiber direction σ90 or σ⊥ while the difference between
the radial and tangential directions is often overlooked.
Fig. 3.12 illustrates the stress-strain relationship for clear wood (timber that does not have
knots, the slope of grain, and imperfections) loaded in tension parallel to the fiber direction.
The strength of timber in tension parallel to the fiber direction (ft0 ) is very high. However,
failure in this direction tends to be highly brittle.
Compressive strength of timber in parallel to the fiber direction (fc0 ) Fig. 3.13 has a ductile
behavior. Since in compression parallel to the fiber direction, the main stress aligns with the
fibers’ axial direction. Tubes loaded axially are stable and can withstand high loads. However,
if the load is too high, some fibers will buckle and be driven into others, reducing the load-
bearing capacity and leading to plastic behavior [16].
Fig. 3.14 illustrates the behavior of clear timber under pure tension perpendicular to the fiber
direction (ft90 ). Strength is very low with brittle failure since forces to pull apart the fibers or
break the fibers are too low [16].

14
By observing Fig. 3.15 behavior of timber under compression perpendicular to the fiber direc-
tion (fc90 ) can be defined. Timber has ductile behavior with low stiffness and strength. Under
loading tube-shaped wood cells will be crushed from the side with low force. However, after
crushing all cells of timber stress level can increase again [16].

Figure 3.12: Stress-strain relationship for clear wood under tension parallel to the fiber direction
[16].

Figure 3.13: Stress-strain relationship for clear wood under compression parallel to the fiber
direction [16].

15
Figure 3.14: Stress-strain relationship for clear wood under tension perpendicular to the fiber
direction [16].

Figure 3.15: Stress-strain relationship for clear wood under compression perpendicular to the
fiber direction [16].

The shear strength of timber is higher in planes parallel to the fiber direction (τRL , τT L ), ranging
from 5 to 8 MPa. However, for structural engineers, it is not possible to distinguish between
these two directions, so the lower of the two values is used. The minimum shear strength
is perpendicular to the fiber direction (rolling shear, τRT ), typically between 3 and 4 MPa,
since the failure line goes through the weaker early wood layer. In planes parallel to the fiber
direction, the failure line goes through both early and late wood [16]. Early wood, formed
during the early part of the season when growth is more active, is lighter in color and less
dense, while latewood, produced later in the season, is darker, denser, and stronger [41].

3.2 Comparison between hardwood and softwood

Tree species are grouped into two categories: softwoods and hardwoods. Softwoods include
spruces, pines, and firs, which keep their needle-like leaves all year. Hardwoods include oaks,
birches, and maples, which lose their broad leaves in winter. Both types are used for structural
timber, but softwoods are more common in Europe due to their greater availability and lower
cost [41]. The names softwoods and hardwoods can be confusing. Some softwoods, like

16
longleaf pine and Douglas-fir, are harder than some hardwoods, like basswood and aspen [69].
The mean density values for softwoods and hardwoods range from about 400 to 650 kg/m3
and 500 to 1200 kg/m3 , respectively. Wood of high density tends to shrink and swell more
with changes in moisture content than wood of low density [41] but most of the mechanical
properties of wood correlate positively with density [15].

3.3 Shear Connector Systems in TCC Slabs

Effectively transferring shear between timber and concrete components relies on a well-designed
shear connector system, a pivotal element in the performance of TCC slabs. This connection
directly influences composite behavior, stress distribution, deformations, and slab design [31].
Over the past two decades, diverse methods have emerged, including nails [62, 73] and screws
which are in dowel-type fasteners group [14], adhesives in glued connections type [62, 81],
notched connectors [81] facilitating the bond between concrete and timber [31]. These con-
nectors typically enable medium to high levels of composite action, depending on their type
[23]. Different properties of each type of connector have been described in the following bullet
points.

• Dowel type fasteners:

Dowel type fasteners are commonly used for TCC beams due to their affordability and
ease of installation. However, they tend to be more flexible. For timber slabs to provide
enough composite action requiring a large number of dowel-type fasteners may not be
needed so it can not be an ideal choice [19].

• Glued connectors:
Glued connections offer several advantages, including the potential to achieve perfect
composite action and provide a simple and fast solution for prefabrication without the
need for drilling or insertion of mechanical connectors [19]. It has also been reported
that it has good behavior in fatigue and cyclic loading [66]. However, there are con-
cerns regarding the brittle behavior of adhesives, sensitivity to hydrothermal changes,
and performance under elevated temperatures (such as fire conditions) [55].

• Notched connectors:
The notched connector system involves grooving the timber element to distribute the
shear force across a larger surface area, thereby enhancing the stiffness and strength of
the shear connector [19]. To mitigate the risk of brittle failure, cracking, and uplift in the
concrete, a dowel-type fastener is typically incorporated at the notch. The depth, spacing,
and edge distance of the notches significantly influence the behavior of the connection
[19]. Experimental research has revealed that cracks are most commonly observed near
notched connections located at the ends of composite slabs, particularly those with rela-
tively shallow notches. Such cracking may expedite the separation between the concrete
topping and the CLT panel [44].

While each connector type showcases unique properties, spanning from exceptional shear
strength to brittle failure, a shared characteristic among them is the lack of considering dis-
assembly capability in the design method [52].

17
3.4 Modeling with an analytical method based on the Gamma-method
The Gamma method, as defined in Eurocode 5, annex B [30], is utilized for the design of
mechanically jointed beams with flexible elastic connections, as illustrated in Fig. 3.16. This
method originally introduced by Möhler in 1956 [61] under the following conditions:

Figure 3.16: Cross-sections and normal stresses distribution based on gamma-method [30].

• Number of layers
This method is limited to three joint layers with non-rigid connections between them.
Consequently, the Gamma method is valid for TCC floors with 3-layer sections at most.
In cases where there are more than three load-bearing layers, it is expected that at least
one of these layers will contain multiple transverse layers positioned between itself and
the layer housing the neutral axis of the TCC section. This configuration poses a chal-
lenge for the conventional Gamma method, rendering it impractical. Consequently, an
adaptation is essential to ensure accurate calculation of the γ factor [26].
• Load case
The Gamma method is restricted to the load acting in the z-direction giving a moment
M = M (X) that varies sinusoidally or parabolically, i.e. a uniformly distributed load.
Furthermore, the Gamma method is restricted to single-span structures and is conse-
quently not valid for continuous floors [26].
• Shear connectors

The Gamma method requires the spacing, ’s’, between the fasteners with a constant
value or varies uniformly according to the shear force, between Smin and SM ax , with
SM ax ≤ 4Smin which implies an even distribution of shear connectors along the span
length. Thus, the individual parts are connected by mechanical fasteners with a slip
modulus, Ks [26, 4].
• Inelastic strains
Inelastic strains, such as those induced by temperature fluctuations, lead to material con-
traction or elongation, thereby causing deformation and generating bending moments in

18
 structural applications. These strains are particularly significant in structures made of
in-situ concrete due to shrinkage during the hardening process. In contrast, prefabricated
concrete undergoes most of its shrinkage before arriving at the construction site. There-
fore, TCC floors constructed with prefabricated concrete are less affected by inelastic
strains compared to in-situ concrete structures [82]. The conventional Gamma method,
used for determining the effective bending stiffness of composite structures, disregards
inelastic strains. However, by slightly modifying the Gamma method, the impact of in-
elastic strains can be addressed by introducing an external and eccentric effective load to
approximate its structural influence [26, 35].

• Cross-sectional characteristics
The Gamma method necessitates a consistent cross-sectional geometry along the span
length. Nevertheless, each layer can exhibit distinct cross-sectional properties. The
bearing capacity of the transverse layers is disregarded and set to zero due to their low
Young’s modulus, typically around E/30, where E represents the longitudinal Young’s
modulus of the load-bearing layers [26, 35, 82]. To fully satisfy the prerequisites of the
Gamma method concerning continuous stiffness and geometric properties along the span
length, the timber boards of the transverse layers should be continuously glued along
their edge interface. The Gamma method is grounded in the Bernoulli-Euler beam the-
ory, which primarily considers bending deformations while disregarding shear strains
[26].

3.4.1 Necessary parameters in Gamma method

The examination of a TCC floor in both Serviceability and Ultimate Limit States focuses pri-
marily on determining the effective flexure stiffness, represented as (EI)ef f , of the composite
cross-section, which is crucial for the design process in both states. This parameter enables the
calculation of deflection ω, first natural frequency f1 , and normal and shear stresses in the TCC
floor.

3.4.2 Determination of the effective bending stiffness

The effective bending stiffness of the TCC section (EI)ef f is calculated using the Gamma
method provided in Annex B of Euro code 5 [4]. However, only with consideration of two
longitudinal layers, and the concrete layer flexibly connected by the shear connectors. The
concrete layer is seen as the top longitudinal layer and the entire 3-layer CLT section is re-
garded as the bottom longitudinal layer of the fictive 2-layer CLT. The bearing capacity of
the transverse non-load bearing layers is neglected since Young’s modulus is equated to zero.
To calculate the effective bending stiffness, the cross-sectional area of the transverse layer is
ignored, considering only the longitudinal load-bearing layers [26].
The effective bending stiffness of the TCC floor is calculated as follows:

(EI)ef f = (EI)c + γc Ec Ac ac2 + (EI)t + γt Et At at2 [kNm2 ] (1)

where:
bi h3i
(EI)c /(EI)t Effective bending stiffness of the concrete/timber layer [kNm2 ], and Ii = 12
γc /γt Reduction factor for the concrete/timber layer

19
Ec /Et Young’s modulus of concrete/timber [MPa]
Ac Cross-sectional area of the concrete layer [m2 ]
At Cross-sectional area of the load-bearing layers in the CLT [m2 ]
ac /at Distance between the center of the concrete/timber layer and the cross-sectional neutral axis [m]

In the CLT part the reduction factor, denoted as γt , highlights the composite behavior between
two load-bearing layers, for each layer an individual factor is considered. The determination
of this factor varies somewhat depending on the specific layer of interest. It is designated as 1
for the layer housing the CLT neutral axis. For the next layer a modified version of the gamma
method, as detailed in the handbook by Gagnon and Pirvu [37], is utilized. This modification
involves computing the γ parameter based on the stiffness of the layers oriented perpendicular
to the load-bearing direction, rather than incorporating the stiffness of the connection. The
factor for this layer is computed as follows [78]:
1
γt = π 2 Et At ht
(2)
1+ GR L2 b

where:

Et Young’s modulus for each layer with grains in the longitudinal direction [MPa]
At Cross-sectional area of the i-layer of CLT [m2 ]
ht Height of the i-layer of CLT [m]
GR Shear modulus for layers with grains in the transversal direction [MPa]
L Span length of TCC floor [m]
b Width of layers [m]

The γ factor for the concrete layer corresponds to the stiffness and spacing, of the shear con-
nectors as well as the material properties of concrete.The γ factor for the concrete layer is
computed as follows [4]:
1
γc = π 2 Ec Ac Sef f
(3)
1+ KL2

where:

Ec Young’s modulus of concrete layer [MPa]

Ac Cross-sectional area of the concrete layer [m2 ]
Sef f Effective spacing between shear connectors [m]
K Stiffness of the shear connector [N/m]
L Span length of the TCC floor [m]

20
3.4.3 Determination of the neutral axis of the TCC floor
To define the spacing between the center of each longitudinal layer and the cross-sectional
neutral axis of the TCC floor, it is essential to determine the position of the neutral axis. The
gamma method mandates that the neutral axis of the TCC floor be located within the CLT
section, which also contributes to enhancing the effective bending stiffness [4]. By establishing
a reference point at the center of the first layer of the CLT, the distance between the neutral axis
and this layer, denoted as d− N L1 , can be calculated as follows:

γc Ec Ac ( h2c + h2t ) − γt Et At ( h2t + ht

2
+ ht )
d− N L 1 = [mm] (4)
γc Ec Ac + γt Et At

where:
Ec /Et Young’s modulus of concrete/timber layer [MPa]
Ac Cross-sectional area of the concrete layer [m2 ]
At Cross-sectional area of each load-bearing layers of CLT [m2 ]
hc Height of the concrete layer [m]
ht Height of each CLT’s layer [m]

Once the neutral axis is determined, by using Eq. (5) and Eq. (6) distance from the center of
the concrete and third layer of the CLT to the cross-sectional neutral axis are calculated.
ht hc
ac = + − NA [mm] (5)
2 2

ht ht
at = + ht + + NA [mm] (6)
2 2

3.4.4 Design criterion for Serviceability Limit State (vibrations)

Analysis of TCC floors in the Serviceability Limit State (SLS) is done in the short term. Short-
term analysis is computationally more convenient than long-term analysis since the material
time dependency does not need to be considered. Analysis primarily involves defining the
maximum (mid-span) deflection and examining the dynamic response to vibrations. To define
the dynamic response of the TCC slab the first natural frequency should be calculated. Par-
ticularly for rectangular floors with overall dimensions l x b, supported along all four edges,
and timber beams spanning l. The fundamental frequency f1 can be approximately calculated
using the formula outlined in Eurocode 5, section 7.3.3 [4]. The mean values of the material
elasticity moduli should be applied in the SLS calculations.
r
π (EI)L
f1 = 2
[Hz] (7)
2L m

where:
L Span length of the TCC floor [m]
(EI)L Equivalent plate bending stiffness of the floor about an axis perpendicular to the beam direction [kNm2 ]
m Self-weight per meter of the TCC floor [kg/m]

21
The initial natural frequency, denoted as f1 needs to exceed 8 Hz to prevent discomfort. For
residential floors exhibiting a fundamental frequency below 8 Hz, (f1 ≤ 8 Hz), it is necessary
to conduct a dedicated examination, and for residential floors with a fundamental frequency
greater than 8 Hz (f1 > 8 Hz) the following requirement according to Eurocode 5 [4] should
be satisfied:
w
≤a [mm/kN] (8)
F
where:
F L3
w Peak immediate vertical displacement, calculated as w = 48EI

F Vertical concentrated static force applied at any point on the floor, taking account of load
distribution.

a Factor which can be determined by Fig 7.2 in EN 1995-1-1, for the Swedish standard is set to
1.5 mm/kN by Boverket [74], lower value for a in Fig 7.2 results in higher performance.

3.4.5 Design criterion for Ultimate Limit State

The analysis of TCC floors in the Ultimate Limit State (ULS) for the short-term verification
concentrates on evaluating internal stresses, encompassing both normal and shear stress, and
validating them through comparison with material strength properties. As for the SLS calcu-
lations, the elasticity moduli for timber and concrete should be taken as their respective mean
values, rather than the characteristic values [18]. It is been outlined that due to the absence
of characteristic values for the slip modulus of the connection, mean values are advocated for
all stiffness properties. The rationale behind this approach lies in capturing the most realistic
behavior, thus resulting in the stress distribution deemed most probable [40, 35]. Normal and
shear stress in the Ultimate Limit State is explained as follows:

• Normal Stress
A TCC slab subjected to a bending moment, experiences normal stress, influenced by
both the individual normal stress σi accounting for coupling effects, and the normal stress
distribution σmi due to bending moment over the layer height. The normal stress distri-
bution within the concrete layer and longitudinal CLT layers is determined by computing
the stresses individually at the top and bottom of each layer [35]. These stresses are
computed following the methodology outlined in Eurocode 5 Annex B [4].

γi Ei ai
σi = M [MPa] (9)
(EI)ef f

0.5Ei hi
σmi = M [MPa] (10)
(EI)ef f

where:

22
 γi Gamma factor of the i-layer of slab

Ei Young’s modulus of the i-layer [MPa]

ai Distance between the geometric center of the i-layer and the neutral axis [m]

(EI)ef f Effective bending stiffness of the TCC floor [MNm2 ]

hi Height of the i-layer [m]

M The maximum design bending moment of the simply supported TCC floor subjected
to a uniformly distributed load across the span length [kNm]

The determination of normal stress at the upper and lower boundaries of each longitudinal
layer involves summing defined stresses from Eq. (10) and Eq. (9). The sign of the σi
stress during summation hinges on whether the respective layer experiences compression
or tension. Typically, layers positioned above the TCC cross-sectional neutral axis are
under compression, while those below it are under tension. The sign of the σmi stresses
during summation varies based on the layer’s location, with a negative sign for stresses
at the upper boundary and a positive one for stresses at the lower boundary of each
layer [35]. The stress state is validated by assessing the total normal stresses against
the material strength characteristics of each longitudinal layer. The criterion is met only
when the total normal stresses remain below the material strength properties across the
entire cross-sectional area.
• Shear stress
The largest shear stress in the CLT section occurs where the normal stress is zero, i.e. at
the location of the cross-sectional neutral axis. It is calculated by Eq. (11)
2
γi Ei Ai ai + 0.5γi Ei bZend
τmax = V [MPa] (11)
(EI)ef f b
where:
V Maximum design shear force of a simply supported TCC floor subjected to a uni-
formly distributed load across the span length L [kN]

Zend The distance from the bottom edge of the layer containing the neutral axis to the
neutral axis, calculated as: Zend = h2t + N.A. [m]

Ei Young’s modulus of the i-layer [MPa]

Ai Cross-section of the i-layer [m2 ]

ai Distance between the geometric center of the longitudinal load-bearing layer and the
cross-sectional neutral axis [m]

(EI)eff Effective bending stiffness of the TCC floor [MNm2 ]

ht Height of the longitudinal load-bearing i-layer of CLT part [m]

b Width of the TCC floor [m]

Rolling shear stress at the joints is defined by Eq. (12) [4]:

γi Ei Ai ai
τij = V [MPa] (12)
(EI)ef f b

23
 And the load on a fastener should be taken as:

Fi = τ12 bsi [MPa] (13)

where:
τ12 Rolling shear stress at the joint between the concrete layer and the first layer of CLT
[M P a]

b Width of the composite floor [m]

si Spacing of the fasteners [m]

3.4.6 Remaining Input Parameters in the Gamma Method

The rest of the input parameters for calculation based on gamma method is presented as follows:

Ei−0mean Young’s modulus for all layers with grains in the longitudinal direction [kN/m2 ]

Ei−90mean Young’s modulus for CLT layers with grains in the transversal direction [kN/m2 ]

Gi−0mean Shear modulus for CLT layers with grains in the longitudinal direction [MPa]

Gi−rmean Rolling shear modulus for CLT layers with grains in the transversal direction [MPa]

ρCLT −mean Density of CLT [kg/m3 ]

ρC Density of concrete [kg/m3 ]

ft0k Characteristic tensile strength along the grain for CLT [MPa]

ft90k Characteristic tensile strength perpendicular to grain for CLT [MPa]

fmk Characteristic bending strength for CLT [MPa]

kmod Modification factor

kdef Deformation factor

qsurf −k Uniformly distributed characteristic load as a surface load [kN/m2 ]

qsurf −F Uniformly distributed frequent load as a surface load [kN/m2 ]

qsurf −Q Uniformly distributed quasi-permanent load [kN/m2 ]

qsurf −U Uniformly distributed ultimate load as a surface load [kN/m2 ]

24
3.5 Finite Element Method
The Finite Element Method (FEM) is a widely used approach for addressing complex structural
and mechanical engineering challenges. It takes advantage of the computational power of com-
puters to efficiently solve numerous equations within a brief time frame. Today, there exists a
plethora of commercial software options satisfying various needs. While some are utilized for
specific scenarios, others offer broader applicability across different domains.

In the finite element method, the process of dividing a continuous geometric model or domain
into a smaller and finite number of elements is known as discretization. Nodes represent points
in space with defined degrees of freedom (DOFs), indicating potential movements under load-
ing. DOFs also illustrate the transfer of forces and moments between elements.

3.5.1 Elements and shape function

In engineering applications of FEM, approximate continuous models using data obtained from
a finite set of discrete points are estimated. The shape function is the function that interpolates
between nodes of the primary unknowns for example temperature or displacement. Various
options for element and their related element shape functions are available for selection and
presented as follows:

• Line element(1D)

Figure 3.17: Illustration of 1D element .

• surface element(2D)

Figure 3.18: Illustration of elements and shape functions in 2D.

25
 • Plain element(3D)

Figure 3.19: Illustration of elements and shape functions in 3D.

3.5.2 Finite element formulation in 3D

The general process through which FEM operates involves solving the mechanical differential
equations for individual elements using approximate numerical methods. Each structural com-
ponent, within the mesh of finite elements for mechanical or structural design, undergoes this
procedure. The quantity of elements composing the mesh significantly influences the compu-
tation time since each element essentially represents a set of differential equations that must
be resolved. However, as the number of elements increases, resulting in a finer mesh, the
calculation outcome tends to converge more closely to the exact solution.
Every variety of elements is addressed by solving a matrix system, comprising an element
stiffness matrix (K), a displacement vector (a), and a force vector (f ), as depicted in the Eq.
(14).
Ka = f (14)
Where (f ) is the force vector containing a boundary load vector (fb ), a load vector (fl ), and an
initial force vector due to initial strains (f0 ). Force vector for a 3D element formulated in the
Eq. (15).
Z Z
T
fb = N hdS + N T tdS
S Sg
Z h
fl = N T bdV (15)
V
Z
f0 = B T Dϵ0 dV
V
The stiffness matrix K for 3D element formulated in the Eq. (16).
Z
K= B T DBdV (16)
V

26
3.5.3 Finite element formulation of beams
In the context of three-dimensional structures, the dominance of a beam’s extension in the ax-
ial direction enables certain assumptions about structural deformation to be made (i.e., about
kinematic relations: Plane cross sections normal to the beam axis remain plain and perpendic-
ular to the beam axis during deformation), thereby significantly simplifying the problem. The
finite element formulation of a single beam element is presented in Eq.(17). It is important to
note that boundary conditions for an element are unknown except when an element boundary
coincides with the boundary of the beam.
Z
e
K = B eT EIB e dx
Lα
dNeT
 
e eT
 
fb = N V Lα − M (17)
dx Lα
Z
e
f1 = N eT qdx
Lα

The parameters in Eq. (15),(16) and (10) are described as follows:

N/N e Global/Element shape function matrix

Sh /Sg Natural/Essential boundary condition region

t Thickness

h Known quantity on natural boundary condition region

b Body forces per unit axial length, i.e. [N/m]

D Constitutive matrix including the material’s stiffness properties

ϵ0 Strains for zero stresses

K/K e Element stiffness matrix

Be Derivative of element shape function N e , to the length

EI Stiffness properties

M/V Bending moment/vertical shear force

Lα Region of the element

q Force per unit length

3.5.4 Modeling with Finite Element method, using Abaqus software

Complete Abaqus environment (Abaqus CAE), is a software suite for finite element analysis
and computer-aided engineering, originally released in 1978. It is a software application used
for both the modeling and analysis of mechanical components and assemblies (pre-processing)
and for visualizing the finite element analysis results [5]. Abaqus is a frequently utilized soft-
ware in research, owing to its extensive capabilities for adjusting input parameters and smooth
usability. It provides flexibility in performing analyses in both two and three dimensions, facil-
itating the utilization of one-dimensional (1D), two-dimensional (2D), and three-dimensional

27
(3D) elements.
In this thesis, three-dimensional modeling in Abaqus software is employed. The process of
implementing Abaqus model is described in the next subsections including the decisions for
the modeling phase, and the modeling procedures.

3.5.5 Modeling decisions

This section outlines a numerical framework to analyze the structural behavior of the studied
TCC slab under static loading conditions, as observed in a four-point bending test and SLS
loading.
By utilizing the Abaqus software package, a three-dimensional finite element (FE) model was
developed and validated against analytical results discussed earlier.
To investigate the proposed floor with reduction in the computational time, a down-scaled
model of the TCC slab was designed. The samples have dimensions of 2300mm x 500mm
x 180mm (length × width × height) and feature a composite structure comprising concrete,
timber, and steel, exhibiting orthotropic and isotropic material characteristics.
To streamline computational resources, a symmetry boundary condition was employed, reduc-
ing the model’s size and computational time. This boundary condition assumes zero displace-
ments normal to the plane of symmetry and rotations about the axes within that plane. By
implementing symmetry in the y-z plane, the model’s size is effectively halved.
As shown in Fig. 3.20, the specimen primarily consists of four parts: precast concrete slab,
CLT slab, wooden dowels, and screws.

28
 (a) Concrete part.

(b) Three-layer CLT part.

(c) Wooden dowel and screw.

Figure 3.20: Outline of the specimen in Abaqus.

29
3.5.6 Modeling procedure
In the following sections modeling procedures, encompassing the selections and input param-
eters utilized in the Abaqus model are presented.

3.5.6.1 Concrete constitutive law

The elastic characteristics of concrete are defined through the calculation of the secant modulus
of elasticity, as suggested in Eurocode 2 [3], using the Eq. (18)
Ecm = 22[(fcm )/10]0.3 (18)

In this study, the mean value of the cylindrical compressive strength at 28 days, fcm was used.
Using Eq. (18), the modulus of elasticity was calculated to be Ecm = 32.84 GP a. Additionally,
the Poisson’s ratio was set to 0.2.
To examine displacement in SLS in the short term the mean value of secant modulus of elas-
ticity is considered, while for long-term effects according to EN 1992 [3] time dependency
material property needs to be considered including creep and shrinkage.
To examine displacement in SLS in the long-term in prefabricated concrete models, Since
part of the creep and shrinkage (autogenous shrinkage) occurs before erection, the effects of
shrinkage would be ignored. To do this in concrete members, for loads with a duration causing
creep, the total deformation including creep can be calculated by using an effective modulus of
elasticity according to Eq. (19).
Ecm
Ec,eff = (19)
1 + φ(∞, t0 )
where:
φ(∞, t0 ) Creep coefficient relevant for the load and time interval which in this calculation is
assumed to be equal to 2.0

Ecm Secant modulus of elasticity of concrete

32.84
Ec,eff Effective modulus of elasticity of concrete, Ec,eff = 3 = 10.947 GPa

To evaluate the crushing of the concrete in compression and the cracking of the concrete in
tension two common failure modes of concrete a simplified concrete damage plasticity (SCDP)
model was used [39]. Concrete damage plasticity (CDP) is considered as a precise and practi-
cal constitutive model to simulate concrete behavior. CDP, introduced by various researchers,
involves the formulation of stress-related plasticity in the effective stress space. Initially de-
veloped by Lubliner et al. [56] and subsequently modified by Lee [53]. The versatility of the
CDP model allows for its application in both static and dynamic scenarios and it is accessible
in Abaqus [39].
In this numerical simulation, all SCDP parameters were derived from the work of Hafezolgho-
rani et al. [39]. They specifically devised an SCDP model for concrete with a characteristic
strength of fck = 30 MPa. The requisite parameters provided by them have been adopted in
the current study for modeling concrete’s non-linear material properties. These parameters are
presented in Tab. 3.1 and explained in the following.
In the SCDP model, the Kent and Park Model [47] was utilized to capture the response of
concrete under a uniaxial loading condition in compression. Its graph is presented in Fig. 3.21.

30
To simplify the entire constitutive model, a parabolic curve was considered. In this graph, they
defined a parabolic increase in compressive stress (A–B) during the hardening stage and a linear
decline (B–C) for both confined and unconfined concrete. The softening phase extended until
20% of the unconfined cylindrical compressive strength (Point C). After this point, the stress
was assumed to remain constant, exhibiting perfect plastic behavior (C–D).
Due to negligible differences among all constitutive models for concrete under tension because
of the concrete’s brittle nature the simplified model was used. In the SCDP model, for mod-
eling uniaxial tensile behavior, tensile strength was adopted to 10% of maximum compressive
strength σcu regardless of the realistic condition to prevent numerical instability [39].
As depicted in Fig. 3.22, as the hardening cracking strain εck,ht increases further, the tension
damage continues to rise. This trend can be described as follows:
σt
dt = 1 − (20)
σt0
where:
dt Scalar tension damage variable

σt Tensile stress in hardening stage [MPa]

σt0 Tensile strength at 28 days [MPa]

Figure 3.21: Kent and Park model for confined and unconfined concrete [47].

31
 Figure 3.22: Response of concrete to a uniaxial loading condition under tension [56].

Plasticity parameters for concrete damage plasticity model is described in the following:
• Dilation angle
The dilation angle, determined by the ratio of volume strain to shear strain, typically
ranges from 20° to 40° for concrete, influencing material ductility and exerting signif-
icant effects on the overall model. Higher dilation angles enhance system flexibility.
Practically, the internal dilation angle is contingent upon specific parameters like plastic
strain and confined pressure, where an increase in these factors reduces the internal dila-
tion angle. However, the material maintains a constant dilation angle across a wide range
of pressure stresses applied for its confinement [39].

• Eccentricity
The default flow potential eccentricity was set to ε = 0.1. Increasing this value height-
ened the curvature of the flow potential. However, if the default eccentricity was signifi-
cantly lower than its default value, convergence issues could arise, particularly when the
confining pressure is low [39].

• Compressive yield stress fb0 /fc0

The ratio of the initial equibiaxial compressive yield stress to the initial uniaxial com-
pressive yield stress has a default value of 1.16 [45].

• K ratio
’K’ is the ratio of the second stress invariant on the tensile meridian to the compressive
meridian at initial yield for any given value of the pressure invariant ’p’. This applies
when the maximum principal stress is negative σmax < 0 .’K’ should be 0.5 < K < 1, If
this field is left blank or a value of 0.0 is entered, the default of 2/3 is used [76].

• Viscosity parameter
The ABAQUS software employs a default viscosity parameter of zero to prevent vis-
coplastic regularization. This parameter improves the convergence rate of the model
during the softening process, yielding satisfactory results [39].

32
Graphs (a) and (b) in Fig. 3.23 independently illustrate the uniaxial stress-strain relationships
for the concrete in compression and tension, respectively. These relationships guide the deter-
mination of cracking and crushing trends, influenced by the values of hardening and softening
variables. These variables dictate the loss of elastic stiffness and the evolution of the yield
surface.

(a) Inelastic strain, compressive behavoir. (b) Cracking strain, tensile behavoir.

Figure 3.23: Concrete compressive and tensile behaviour for class C30.

33
 Material’s parameters C30 Plasticity parameters
Dilation angle 31
Concrete elasticity Eccentricity 0.1
E (GPa) 32.84 fb0/fc0 1.16
ν 0.2 k 0.67
Viscosity parameter 0
Concrete compressive behavior Concrete compression damage
Yield stress (MPa) Inelastic strain Damage parameter Inelastic strain
10.2 0 0 0
12.8 7.73585E-05 0 7.73585E-05
10.2 0 0 0
15 0.000173585 0 0.000173585
16.8 0.000288679 0 0.000288679
18.2 0.000422642 0 0.000422642
19.2 0.000575472 0 0.000575472
19.8 0.00074717 0 0.00074717
20 0.000937736 0 0.000937736
19.8 0.00114717 0.01 0.00114717
19.2 0.001375472 0.04 0.001375472
18.2 0.001622642 0.09 0.001622642
16.8 0.001888679 0.16 0.001888679
15 0.002173585 0.25 0.002173585
12.8 0.002477358 0.36 0.002477358
10.2 0.0028 0.49 0.0028
7.2 0.003141509 0.64 0.003141509
3.8 0.003501887 0.81 0.003501887
Concrete tensile behavior Concrete tension damage
Yield stress (MPa) Cracking strain Damage parameter (T) Cracking strain
2 0 0 0
0.02 0.000943396 0.99 0.000943396

Table 3.1: Material properties for concrete with SCDP model in class C 30 [39].

34
3.5.6.2 Timber constitutive law

The behavior of timber in the elastic stage is calculated with Hooke’s equation:

εi = Cij σj ∀i, j ∈ [ 1, 2, ..., 6] (21)

Where εi , σj , and Cij are components of engineering strain, stress vector, and entries of the
elastic compliance matrix (C) that has the following form for an orthotropic material respec-
tively [38]:

ε = {ε11 , ε22 , ε33 , 2ε23 , 2ε12 , 2ε13 }T (22)

σ = {σ11 , σ22 , σ33 , σ23 , σ12 , σ13 }T (23)
 
1
E1
− νE211 − νE311
 
 ν12
− E2 1
− νE322

E2

 
 ν13
− E3 − νE233 1 
E3

C=
  (24)
1 
E4
 
 

 1 

 E5 
1
E6

The vector E = {E11 , E22 , E33 , G23 , G12 , G13 }T contains the elasticity moduli for various ac-
tions, while νij represents Poisson’s ratios. These ratios are subject to conditions such as νij
are Poisson’s ratios that satisfy ν12 E1 = ν21 E2 , ν13 E1 = ν31 E3 and ν23 E2 = ν32 E3 ensuring
symmetry in the elastic compliance matrix. The elastic stiffness matrix(D) is then derived as
the inverse of the elastic compliance matrix D = C−1 .
To check SLS requirement according to EN 1995 [4], in serviceability limit states, if the struc-
ture comprises members or components with varying time-dependent characteristics, the final
mean values of modulus of elasticity (Emean.fin ), shear modulus (Gmean.tin ), and slip modulus
(Kser.fin ) which are used to calculate the final deformation should be taken from the following
equations:

Emean
Emean,fin = (25)
(1 + kdef )
Gmean
Gmean,fin = (26)
(1 + kdef )
Kser
Kser,fin = (27)
(1 + kdef )
where:

35
 Emean Creep coefficient relevant for the load and time interval which in this calculation is
supposed to be equal to 2.0.

Ecm Mean value of modulus of elasticity [MPa]

Gmean Mean value of shear modulus [MPa]

Kser,fin Slip modulus

kdef Factor for the evaluation of creep deformation taking into account the relevant service
class, for CLT with less than 7 layers, and service class 1, kdef = 0.85 [2]

The mechanical properties of the utilized timber in the elastic stage are specified in Tab. 3.2,
which Abaqus has an option for by the name of ”Elastic, Engineering Constants”. For SLS
long-term checks, one-third of the values of the Tab. 3.2 are considered.

Material E1 E2 E3 N u12 N u13 N u23 G12 G13 G23

Unit [MPa] [MPa] [MPa] [−] [−] [−] [MPa] [MPa] [MPa]
C24 11000 370 370 0.45 0.45 0.5 690 690 60
C40 14000 470 470 0.45 0.45 0.5 880 880 60
Oak, red 12500 1925 1025 0.35 0.448 0.56 792.1 720.1 180

Table 3.2: Mean elastic properties of utilized timbers.

Tab. 3.2 employs abbreviations derived from Abaqus, where ’E represents Young’s modulus,
’Nu’ stands for Poisson’s ratio, and ’ G’ denotes the shear modulus. The numerical labels 1,
2, and 3 correspond to the local material directions x, y, and z respectively, indicating parallel
alignment to grain (longitudinal) and perpendicular alignment to grain (radial and tangential).
Timber materials often exhibit complex responses due to directional dependency and material
non-linearity beyond their elastic limits. Compression in timber typically shows plastic behav-
ior with hardening potential, while tensile and shear responses generally demonstrate brittle
softening behavior. Therefore, by understanding timber’s strength, the brittle failure of tim-
ber can be defined even without modeling its plastic material properties, which require long
computational times.
To define each stress component, user-defined output variables named UVARM subroutine, a
program written in Fortran by Dorn [29], are used.
The characteristic strengths for application in the UVARM subroutine to be added in Abaqus
are presented in Tab. 3.3, and their descriptions are presented in Tab. 3.3.

36
 Material ρmean ft11 fc11 ft22 fc22 ft33 fc33 fs12 fs13 fs23
Unit kg/m3 [MPa] [MPa] [MPa] [MPa] [MPa] [MPa] [MPa] [MPa] [MPa]
C24 420 14.5 21 0.4 2.5 0.4 2.5 4 4 2.52
C40 480 26 27 0.4 2.8 0.4 2.8 4 4 2.52
Oak, red 740 50 20.7 5.2 4.2 5.2 4.2 8.1 8.1 4

Table 3.3: Density and strength value of utilized timbers.

Strength Value Description

ft11 Tensile strength in fiber direction
fc11 Compressive strength in fiber direction
ft22 Tensile strength in radial direction
fc22 Compressive strength in radial direction
ft33 Tensile strength in tangential direction
fc33 Compressive strength in tangential direction
fs12 Shear strength in plane 1 and 2
fs13 Shear strength in plane 1 and 3
fs23 Shear strength in plane 2 and 3

Table 3.4: Strength parameter to add in UVARM subroutine.

Once these values are defined by the user, 28 variables appear in the result file after the job’s
run is completed, by the title of UVARM 1 to 28. These variables help to analyze the results
and are described in more detail in the following.

• UVARM1-9
The first nine variables, referred to as UVARM1-9, describe the charecteristic strength
values entered by the user. The order of these is shown in Tab. 3.4. For instance,
UVARM1 and UVARM2 correspond to tensile and compressive strength in material di-
rection 1 (parallel to the fiber direction) respectively, and so on.

• UVARM10-18
The variables UVARM10 -18 represent the linearly independent utilization rate, irrespec-
tive of whether the interpolation point being calculated is under compression or tension.
UVARM10 denotes the correlation between stresses in direction 1(fiber direction) and the
tensile strength in the same direction which was introduced to the model by UVARM1,
Eq. (28).
σ11
UVARM10 = (28)
ft11

37
 UVARM11 signifies the relationship between stresses in direction 1 and the compressive
strength in that direction, as indicated in Eq. (29).
σ11
UVARM11 = (29)
fc11

Correspondingly, UVARM12 to 15 represent their counterparts in material directions 2

and 3.
UVARM16 describes the relationship between the shear stresses and the shear strength
in the 12-plane, Eq. (30).
σ12
UVARM16 = (30)
fs12
The variables UVARM17 to 18 describe the relationship between the shear stresses and
the shear strength, similar to Eq. (30), but in the 1-3 and 2-3plane.
• UVARM19-24
UVARM19-24 is the ratio between stress in each element of the specimen and the strength
value at that direction. In doing so UVARM19 is the ratio between stress in the normal
direction (S11) and strength value at direction 1 (fiber direction) which is the value of
UVARM10 or UVARM11 depending on whether S11 is in tension or compression, Eq.
(31).
(
UVARM10 if S11 > 0
UVARM19 = (31)
UVARM11 if S11 < 0

The variables UVARM20 to 21 are defined similarly to UVARM19, but for directions 2
and 3.
As the shear utilization rate is independent of whether the elements are in tension or
compression, the variables UVARM22-24 are the same as UVARM16-18 respectively.
• UVARM25-26
Although multi-axial stress conditions exist in almost all engineering applications, design
procedures often assume uniaxial stress conditions. However, this assumption is not
adequate for timber, which has orthotropic material properties [17]. There are a few
models developed specifically for wood to predict its multi-axial failure, such as those by
Norris [64] and van der Put [79].
In this study, two simple interactions of stresses were considered: a linear model and a
quadratic polynomial model by UVARM 26 and UVARM 28 respectively.
The variable UVARM25 creates a linear combination of all ratios of UVARM19-24 pre-
sented in Eq. (32). If the result is less than 1, it satisfies the criterion; otherwise, if it is
greater than 1, it does not meet the criterion. This outcome is represented in UVARM26
by the numbers 1 or 0, respectively Eq. (33).
24
X
UVARM25 = UVARMi (32)
i=19

σ11 σ22 σ33 σ12 σ13 σ23

+ + + + + ≤ 1, 0 (33)
f11 f22 f33 fs12 fs13 fs23

38
 • UVARM27-28
The variable UVARM27 creates a quadratic combination of all ratios of UVARM19-24
presented in Eq. (34). If the outcome meets the criterion, it is indicated in UVARM28 by
the number 1; otherwise, it’s represented by 0 in Eq. (35).
24
X
UVARM27 = (UVARMi)2 (34)
i=19

σ11 2 σ22 2 σ33 2 σ12 2 σ13 2 σ23 2

( ) +( ) +( ) +( ) +( ) +( ) ≤ 1, 0 (29) (35)
f11 f22 f33 fs12 fs13 fs23

3.5.6.3 Steel constitutive law

The fasteners utilized in this investigation are called Countersunk Screws sourced from a rel-
evant handbook [75], designed to withstand significant loads. The countersunk head gives a
flush fitting while allowing the CLT panels to close up firmly, also milling thread helps to re-
duce drive-in torque, and no pre-drilling is required, presented in Fig. 3.24. SWC screws are
provided in different lengths and diameters range and its overview is presented in Tab. 3.5, and
SW C10 × ℓ for this research is simulated.

Figure 3.24: Details of the SWC – Countersunk Screw.

Dimensions [mm]
Products
d l lg dh di tf ix
SW C6 × ℓ 6 200-300 70 11.8 3.9 130-230
SW C8 × ℓ 8 80-400 50-80 14.6 5.2 30-320
SW C10 × ℓ 10 100-400 50-80 17.8 6.2 50-320

Table 3.5: SWC–Range Overview [75].

The mechanical characteristics of screws are detailed in Tab. 3.6. Both elastic and plastic ma-
terial properties are employed to describe the performance of steel materials within numerical
simulations. Elastic characteristics are utilized, characterized by Young’s modulus of 200 GPa
and a Poisson’s ratio of 0.3. Additionally, plastic properties are incorporated through the defini-
tion of a bilinear stress-strain diagram, as depicted in Fig. 3.25, within the Abaqus framework.

39
 SW C10.0 × ℓ
Fy (MPa) Fu (MPa)
240 400

Table 3.6: Yield and ultimate strength of SW C10.0 × ℓ.

Figure 3.25: Bilinear stress-strain curve of steel (Unit of stress: MPa).

3.5.6.4 Geometry model, element type and mesh

Specimens are simulated in 3 dimensional. To model the CLT part, the entire three-layer CLT
slab was initially modeled as a single unit. Then, it was subdivided into three layers through
partitioning. It was assumed that there was no separation of layers and the connection between
them was considered perfect. An appropriate local coordinate system was subsequently applied
to each layer to precisely account for variations in the primary direction of the layers. Thus the
fiber direction for the first and third layers was assigned along the span length, while for the
second layer, the fiber direction was perpendicular to the span.
The element type used for meshing components is a continuum, 3-dimensional, and 8-node
solid element employing the reduced integration method (C3D8R).
To reduce the sensitivity of the analysis results to the mesh size and to achieve an appropriate
size for accurate prediction of both local and global responses, various numerical models of
each specimen were prepared. These models had different maximum mesh sizes, ranging from
40 mm to 18 mm in the slab parts. The maximum vertical displacements of these models were
then compared with each other and the mesh size after obtaining constant vertical displacement
was utilized in the FEM modeling. The details are presented in the mesh size-displacement
diagram in Fig. 3.26.

40
 Figure 3.26: Mesh sensitivity analysis for slab.

Consequently, the mesh sizes for screws, dowels, concrete layer, and CLT layer were set at 6, 8,
20, and 20 mm, respectively. Details of the meshing of components are specified in Fig. 3.27.

41
 (a) Concrete.

(b) Three-layer CLT.

(c) Screw. (d) Wooden dowel.

Figure 3.27: Finite element mesh of the model.

42
3.5.6.5 Interactions

The interaction among surfaces, encompassing the contact between wooden dowels, as well
as concrete and CLT, was simulated utilizing the "General contact (Standard)" feature within
Abaqus. Employing the "hard contact" model inherent in the Abaqus software, normal contact
between surfaces, potentially inducing compressive or tensile stresses, was accurately captured
which means incorporating normal contact between the connector and the other components
(timber and concrete). Furthermore, frictional effects at all interfaces were accounted for based
on the Coulomb criterion. A hole was generated only in the concrete part by removing the
volume occupied by the wooden dowels. Regardless of the threading of the fasteners, no gap
existed between the holes and them. In contrast, a tangential contact model was employed
using friction coefficients as detailed in Tab. 3.7. Tangential behavior was characterized by
the Penalty method, with friction coefficients of 0.4 for timber-concrete interfaces, less than
0.57 from Dias et al.’s study [27], because in this model the prefabricated concrete was tried to
model, which has less friction than wet-dry surfaces and prefabrication process may introduce
gaps between layers, potentially diminishing friction between them.
In addition, the “embedded” constraint is used to embed the screws in the dowels and CLT slab,
which means the translational degrees of freedom of the embedded nodes of the screws were
restricted by the response of the host element which are wooden dowels and CLT [1].
Moreover, due to the extensive contact surfaces leading to extremely slow convergence during
analysis, an automatic stabilization feature with a damping factor of 0.0002 was set in the
Abaqus software.

Static coefficients of friction Dias et al. [28] Present study

Timber/Concrete 0.57 0.4

Table 3.7: Comparison of the static coefficients of friction values.

3.5.6.6 Loading and boundary condition

Two types of loading are applied to the specimens. The first type utilizes a distributed load
to assess the SLS displacement requirement, which is then validated against the analytical
method. The second type involves the composite action analysis of the TCC floor, specifically
the evaluation of the shear capacity of the joint and the identification of the failure model. This
includes recognizing the weakest elements of the joint. To do so, the floor section, comprising
the notched and bolted connections, was modeled as a beam in a 4-point bending test. Such
a model makes it possible to analyze the zone of the connectors subjected to a constant shear
force in the joint and thus answer the fundamental research question. Using a concentrated
load to simulate a four-point bending test on the TCC slab determines the TCC slab’s effective
bending stiffness.

Distributed load
In this model, a uniform load was applied to check the SLS requirement. The characteristic load
combination was used for the short-term SLS requirement to control displacement and check
for cracking at the bottom of the slab. It must be mentioned that to control the floor’s displace-
ment for this size of cross-section, the span should be approximately double what was utilized

43
in this model to simulate a real structure. However, FEM modeling of this novel TCC slab with
such connections takes a lot of time, and to reduce the time, this down-scaled model with a
real floor slab cross-section but a shorter span was studied. Therefore, the down-scaled model
cannot detect the real vertical displacement of a TCC floor slab, and the result of displacement
is only for comparison with the analytical method and validation.
The boundary conditions were transversal symmetry (U 1 = U R2 = U R3 = 0) at mid-span
and a line-hinged support (U 1 = U 2 = U 3 = 0) on the bottom left side of the slab.

Figure 3.28: Illustration of distributed load and boundary conditions.

Concentrated load
To simulate a four-point bending experiment with Abaqus all specimens were simply supported
at a distance of 100 mm from the end of the slab and spanned a distance of 2100 mm.
Load is a Quasi-static type which means that the material model is time-independent and rate-
dependent effects are not considered. The model for material does not cover dynamic loading
or viscoelastic effects. Point loads were placed at the 1/3 of the support span. The loading
was applied under displacement control conditions at a loading rate of 4 mm/s. In a four-point
bending test the specimen was modeled as a simply supported beam and two-concentrated loads
placed at an equal distance from the mid-span Fig. 3.29. In this study due to not modeling
timber plasticity, the model cannot detect failure and the results for timber parts are only in the
elastic stage.
Calculation of the bending modulus of elasticity for a four-point bending test where the point
loads are applied at a known distance from each support is done by Eq. (36) [13].

ap (3L2 − 4a2 )
EI = (36)
48δ

where:

44
 EI Bending modulus of elasticity [kN.m2 ]

p Applied point load on four-point bending test [N]

L Support span [mm]

a Distance between supports and point loads [mm]

δ Vertical mid-span displacement of specimen under test [mm]

Figure 3.29: Specimen under four point bending test.

The four-point bending test offers straightforward sample geometries, requiring minimal ma-
chining. It features a simple test setup and allows for the use of fabricated specimens [21].
To replicate the test conditions and because of utilizing symmetry, one of the two reference
points (PR1) was set up, and all degrees of freedom are restrained (U1 = U2 = U2 = U R1 =
U R2 = U R3 = 0) at this reference point. Subsequently, one-line nodal points at a distance
of 100 mm were selected and by coupling constraint available in Abaqus were coupled to
the reference point. This configuration simulates a simply supported beam. Point load with
introducing a reference point (PR2) was also applied at a distance of 1/3 of the support span,
700 mm from the support, and then again a line nodal point was selected and by coupling
constraint was coupled to this specific reference point.
To simulate a symmetry condition in the Y − Z plane half of the specimen was modeled then
the whole surface of the middle of the specimen was selected and symmetry available in Abaqus
by the name of ’XYSMM’ which has displacement in x and rotation around Y and Z directions
equal to zero was applied. A general picture of loading and boundary conditions is illustrated
in Fig. 3.30.

Figure 3.30: Illustration of concentrated load and boundary conditions.

45
3.5.6.7 Analysis procedure

The material and geometric nonlinearity challenges were addressed using a standard static so-
lution. Abaqus/Standard is a general-purpose finite-element solver designed to simulate true
static events and was employed for this purpose.
Displacement-control mode was utilized to apply force to the specimen, making a positive
bending moment on the slab. The reasons for not applying a force-controlled test are that non-
linear static analysis usually runs better in displacement-controlled and force-controlled non-
linear static analysis has a problem, when the resistance of the structure goes down during the
analysis and the solver cannot automatically reduce the applied force. Typically, the analysis
aborts in those cases.
In consideration of significant deformations, geometric nonlinearities were accommodated by
activating the NLgeom feature in Abaqus. Additionally, to mitigate potential slow convergence
caused by a large number of surfaces in contact, the automatic stabilization feature in Abaqus
was utilized with a damping factor set to 0.0002.

46
4 Analysis of obtained results
In this chapter, the obtained results from both analytical and numerical modeling are presented,
discussed, and after the validation of numerical results, a parameter study on some decisive
parameters, including material properties of wooden dowels, fiber direction of wooden dowels,
depth of the notch are undertaken.

4.1 Analytical results

In order to have composite action concrete part needs to have an appropriate connection with
the CLT part. So the reduction coefficient γ with the value between 0 and 1 was introduced
according to the types of connections.
By the fact that when γ = 0, the composite system is in the non-composite state, and when
γ = 1, the composite system is in the full-composite state.
Effective bending stiffness for the state connection of model, no and full composite action for
short term SLS are calculated, and presented in Tab. 4.8.

EIef f [kN.m2 ] No composite action The research model Full composite action
Short-term 1699 2345 3137

Table 4.8: Comparison of the effective bending stiffness.

The next parameter examined in the gamma method is the control of displacement in SLS. A
comparison of vertical displacement in the short, based on deflection limits, was outlined in
Tab. 4.9.

Deflection Short term SLS Limit

Quantity [mm] 0.423 7.667

Table 4.9: Comparison of vertical displacement by gamma method.

47
4.2 Finite element results
In the following subsections, results from two types of loading in the research model are pre-
sented. Results include the load versus mid-span displacement curve, concrete tensile damage
if applicable, ratio of rolling shear stress to strength (UVARM18), ratio of tension perpendicular
to the fiber stress to strength (UVARM26), and normal stress.
Since Abaqus does not have any defined built-in system of units, chosen units for this study are
presented in Tab. 4.10.

Quantity Unit
Length [mm]
Force [N]
Mass [ton]
Time [s]
Stress [MPa]
Density [ton/mm3 ]

Table 4.10: Table of SI units used in Abaqus.

4.2.1 Results for SLS checks

In this subsection Abaqus results for short-term SLS checks are presented.

Figure 4.31: Short term normal stress in concrete.

48
 Figure 4.32: Short term normal stress in CLT.

Figure 4.33: Short term normal stress in screws.

49
 Figure 4.34: Short term normal stress in dowels.

Figure 4.35: Rolling shear stress check in short term.

50
 Figure 4.36: Tension perpendicular to the grain stress check in short term.

In Tab. 4.11 a Comparison between maximum mid-span displacement from FEM modeling
and the analytical method for short term SLS is presented.

∆γ (mm) ∆F EM (mm)
0.42 0.32

Table 4.11: Maximum displacement in FEM modeling versus the γ-method.

4.2.2 Verification and discussion for uniformly distributed load

In this subsection, the validation of FEM modeling is conducted by comparing the vertical
mid-span deflection obtained from the gamma method and Abaqus modeling.
The results from both analyses exhibit similarity due to certain factors. In the short-term SLS,
the initial deflection was computed without considering time-dependent material properties
such as creep and shrinkage. This approach resulted in a maximum mid-span deflection of 0.42
mm and 0.32 mm from the gamma method and Abaqus modeling.
Fig. 4.31 provides a visual representation of the normal stress distribution in the concrete
section during the short term. It illustrates that the concrete experiences predominantly com-
pressive stress, with minimal tensile stress in the upper part of the end-plate end and the lower
part of the mid-span. The highest normal stress occurs at the mid-span in the top layers of
the concrete. Additionally, there is increased compressive stress around the wooden dowel,
particularly the one located near the mid-span.
The graphical depiction of the CLT part in Fig. 4.32 reveals an anticipated predominance of
tension. Areas of localized compression are observed near the support, along the left edge
of the notch, and on surfaces interacting with wooden dowels. The highest tensile stress is
concentrated within the notch.

51
To detect whether potential failures in wooden parts, such as tension failure perpendicular to
fiber directions or rolling shear failure, are occurring, the results of UVARM12 and UVARM18
Figs. (4.36, 4.35, respectively) are considered. VARM12, representing the ratio of tensile stress
to strength perpendicular to the fiber direction, is observed to be low. Similarly, UVARM18,
representing the ratio of rolling shear stress to strength perpendicular to the fiber direction, is
also low, with its maximum found in the second layer of the CLT.
As it is presented in Figs. 4.33 and 4.34 screws and wooden dowels are under compression
while stresses are in the elastic stage.
In Tab. 4.12, the results of stress from both gamma method and FEM are presented. It shows
that results from the two types of analysis are close to each other, and it validates the numerical
modeling. However, the gamma method has higher stress values than the Abaqus modeling.
This means the gamma method results are more conservative, and it considers lower stiffness
for TCC slab than the Abaqus model.

Gamma method Abaqus

Stress [MPa] σc.t σc.b σclt.t σclt.b σc.t σc.b σclt.t σclt.b
Short-term SLS -1.18 0.32 -0.26 0.7 -0.94 0.19 -0.21 0.61

Table 4.12: Normal stresses of γ-method versus Abaqus modeling.

52
4.2.3 Results of four-point bending test
After validating the TCC slab model under SLS, the results of the simulation for a four-point
bending test are presented for the model with a 30-mm notch depth, two wooden dowels with
fiber directions parallel to the span, and structural timber C24 used for both the CLT part and
the wooden dowels. In the following results, the plots were scaled by a factor of 50 to better
interpret and visualize the deformation of the TCC slab

Figure 4.37: Normal stress in concrete.

Figure 4.38: Concrete damage in tension.

53
 Figure 4.39: Normal stress in CLT.

Figure 4.40: Normal stress in wooden dowel.

54
 Figure 4.41: Normal stress in screws.

Figure 4.42: Tension perpendicular to the grain stress check in timber parts.

55
Figure 4.43: Rolling shear stress check in timber parts.

56
Figure 4.44: Linear combination effects of all stresses.

Figure 4.45: Vertical displacement.

57
4.2.4 Discussion for the TCC floor under four-point bending test
For the TCC floor slab under SLS loading, no damage was observed. Therefore, in the next
step, the TCC floor was tested under a four-point bending test with a displacement four times
greater to examine its mechanism, and identify probable damage zones, and improvements in
its design. The results are presented in the following.
• The comparison between the effective bending stiffness obtained from the FEM and the
gamma method shows that the gamma method predicts 25% higher bending stiffness than
the Abaqus modeling, as calculated in the appendix.

• Fig. 4.37 shows that at the end of the specimen on the top part, the uplifting effect causes
normal stress to be tensile. In the mid-span, the upper part experiences maximum com-
pressive stress, which transitions to tensile stress on the lowest surface. The compressive
and tensile stresses in the concrete are below the maximum strength, indicating no crush-
ing or cracking occurred. Fig. 4.38, which shows concrete tensile damage (DAMAGET),
confirms that no tensile cracks appeared under the loading conditions, verifying no ten-
sile failure.

• Fig. 4.39 illustrates that in the CLT part, the notch surface experiences maximum tension,
approximately 12 MPa. The first layer of CLT in the notch, with a depth of 30 mm,
experiences more stress concentration. Additionally, in the third layer from the top of
the CLT, tensile normal stresses are higher than the other layers and increase as they
approach the mid-span. Since material properties for timber parts are in the elastic stage,
this model cannot define the ultimate load-carrying capacity of the structure. However,
two common brittle failures of timber, rolling shear and tension perpendicular to the
grain, are so brittle that they do not exhibit plastic strain and can be easily checked by
this model.

• In Fig. 4.40, normal stresses in the dowels are mostly compressive. Fig. 4.41 indicates
that screws have compressive normal stress only in their upper part, and in the region
embedded in the CLT, they experience tensile normal stress. All stresses for screws are
below the yielding stress.

• Fig. 4.42 is utilized to check tension failure in timbers. The ratio of tensile stress perpen-
dicular to the fiber direction to strength is greater than one only at the support. Therefore,
there would not be any brittle tension failure for this model.

• To examine rolling shear failure in the CLT part, Fig. 4.43 is utilized. As expected,
the highest ratio of shear stress to strength in the radial-tangential plane is in the second
layer of the CLT. However, this ratio is lower than one, indicating no stress exceeding the
strength and no failure.

• However, Fig. 4.44, which represents the linear interaction of multi-axial stresses, shows
that the ratio in the notch and wooden dowel is greater than one. Thus, if the plastic
properties of timber had been modeled, and after reaching maximum stresses with stress
redistribution and elongation, these regions would be critical.

• And finally Fig. 4.45 illustrates the displacement of the specimen in the vertical direction,
with the maximum displacement occurring at the mid-span. Furthermore, an uplift effect
is observed at the end of the specimen.

58
4.3 Parameter study
By evaluating Figs. 4.39 and 4.44, it is evident that the wooden dowels, notch, and support parts
are critical regions in this model. However, the high stress in the support is caused by stress
concentration due to the line support used in the modeling, while the supports in the experi-
mental test have a larger cross-section. Furthermore, this stress is compression perpendicular
to the fiber direction, which exhibits completely ductile behavior. For these reasons, the high
stress in the support was ignored. In contrast, the maximum stress in the dowels and notch is
tension perpendicular and parallel to the fiber respectively, which all is associated with brittle
failure. Therefore, a parameter study is necessary to reduce the stresses in these regions.
According to the obtained results from the specimen, the parameter study includes three parts.
Firstly, the fiber direction for the wooden dowels is changed from parallel to the span to per-
pendicular direction. Secondly, two additional material properties are examined for the wooden
dowels. Lastly, the depth of the notch is decreased. The results of all these changes are com-
pared with the first specimen.

4.3.1 Fiber direction in wooden dowel

In the first model, the dowels had their fiber direction parallel to the span to resist stresses. In
this parameter study, the effect of having the fiber direction in the wooden dowels perpendicular
to the span was investigated. According to Fig. 4.46, the damage was larger in the dowels
with the fiber direction perpendicular to the span, confirming that the initial assumption was
correct. The reason for this could be that the stress in the dowels parallel to the span is primarily
compressive, while the stress in the perpendicular direction is tensile. By aligning the fiber
direction along the span, the dowels can resist more stress and reduce tensile stress in the
perpendicular direction, consequently decreasing the damage zone.

(a) Wooden dowel with fiber direction parallel to the (b) Wooden dowel with fiber di-
span. rection perpendicular to the span.

Figure 4.46: Comparison of different fiber direction for wooden dowels.

59
4.3.2 Depth of the notch
In this subsection, the effects of the depth of the notch, one of the critical areas in the first
model, was studied. The notch depth was decreased to 20 mm to increase the thickness of the
notch. A comparison of results between the first model and the model with the decreased depth
notch is presented in Fig. 4.47, illustrating the improvement in the model by decreasing the
notch depth. The figure shows that the critical elements are dramatically reduced.
However, it should be mentioned that the notch parallel to the span experiences high tensile
stress of around 12 MPa, which is in the fiber direction. In this direction, C24 has a character-
istic tensile strength of 14.5 MPa, meaning the fiber after reaching this high tensile stress has
very brittle failure. In the perpendicular direction, the notch experiences lower tensile stress
than its strength. Overall, notches are considered one of the critical regions in the model.
As presented in Tab. 4.13, compressive normal stresses in concrete increased slightly by 1.19%,
while tensile stresses in the CLT part, in both parallel and perpendicular directions, decreased
significantly by approximately 39% and 0.53%, respectively.

Notch depth [mm] CLT [MPa] Concrete [MPa] ∆M ax [mm]

30 σo 11.54 σ9o 0.3 11.54 Max Compression -15.97 4.34
20 σo 7.044 σ9o 0.14 7.044 Max Compression -16.16 4.325
Differences % -38.96 -0.53 1.19 -0.58

Table 4.13: Comparison of max stress & mid-span displacement, in two different notch depths.

60
 (a) 30mm-notch deep.

(b) 20mm-notch deep.

Figure 4.47: Linear combination effects of all stresses.

61
4.3.3 Timber material properties
In this subsection, the effects of the material properties of wooden dowels on the performance
of the studied slab are investigated. In the first model, structural timber C24 was utilized for
both the CLT and wooden dowel parts. However, since the thickness of the concrete at 40 mm
is smaller than the thickness of the CLT at 120 mm, the embedded length of the wooden dowels
in the concrete is shorter.
To achieve an efficient connection, it is beneficial to have similar resistance on both sides of
the shear plane. To address this, the material properties of the wooden dowels were changed.
First, structural timber C40 was used, and then red oak wood, which is a hardwood and much
denser, was used.
Wooden dowels made of C24 timber experience tension perpendicular to the fiber around 1.5
MPa, which exceeds the strength in this direction. By changing the dowel material to one with
greater strength, the stress in the dowels decreased. Fig. 4.48 illustrates the ratio of stress to
strength in the perpendicular direction for three different material properties. The dowel made
of oak wood shows no exceeding of the stress to strength. Fig. 4.49 clearly illustrates better
resistance of oak wooden dowels by the reduction in the critical elements in the dowels.

(a) C24-wooden dowels.

(b) C40-wooden dowels. (c) Oak-wooden dowels.

Figure 4.48: Ratio of tension perpendicular to the grain to strength in wooden dowel part.

62
(a) C24-wooden dowel. (b) C40-wooden dowel.

(c) Oak-wooden dowel.

Figure 4.49: Linear combination effects of all stresses.

63
5 Conclusions
This study focused on numerical investigations of a novel dismountable TCC floor system
comprising prefabricated concrete and a three-layer CLT connected with notches and screws.
To enable disassembly after the end of service life, wooden dowels were placed in the concrete
layer, and screws were drilled inside the wooden dowels to connect the concrete and CLT.
In this research, concrete and steel’s elastic and plastic material properties were modeled, along
with timber-only elastic properties. Specimens were modeled in 3D in Abaqus/standard.
The scope of the research was to analyze the proposed TCC model in short-term SLS stage.
Subsequently, the bending performance of the slab under a four-point bending test was exam-
ined. To achieve this, the research results were first validated and compared with the gamma
method provided in EC5, annex B. Then under 4 times bigger displacement, the performance
of the slab was tested and parameter study in fiber direction for wooden dowels, material prop-
erties of wooden dowel and depth of the notch were done.
The main conclusions of this investigation are summarized as follows:

• The deflection results in short-term SLS obtained from the gamma method and FEM
modeling were close to each other. However, the effective bending stiffness calculated
using the gamma method was slightly lower than that from FEM modeling, making the
gamma method results more conservative.

• In the SLS checks, the specimen exhibited compressive normal stress in the concrete
layer, with the maximum stress occurring at mid-span. Tensile normal stress was ob-
served at the end-plate of the concrete layer and the lower part of the mid-span. However,
no crushing due to compressive stress or tensile crack failure was observed.

• In CLT predominance normal stress was tension with its maximum concentrated within
the notch.

• No tension perpendicular to the fiber direction and rolling shear two common brittle types
of failure in timber members was found.

• In the four-point bending test tensile normal stress in the lower surface of the concrete
layer in mid-span increased significantly but no tensile crack was observed.

• In the four-point bending test in CLT part notch surface experienced maximum tension
parallel to the fiber direction, and since timber has high but brittle behavior in this di-
rection the notch was considered one of the critical regions in the model, and in the
parameter study by decreasing of the depth, tension more than 50% decreased.

• Screws have compressive normal stress only in their upper part, and in the region embed-
ded in the CLT, they experienced tensile normal stress. But all stresses are below yielding
stress.

• Dowels in the four-point bending test had compressive normal stress along the span while
the stress in the perpendicular direction for them was tensile. So the optimized fiber
direction for dowels is parallel to the span.

64
• Based on the parameter study, dowels with a fiber direction parallel to the span resist
stresses more effectively than those with a fiber direction perpendicular to the span.
Therefore, aligning the fiber direction along the span enhances the dowels’ stress re-
sistance and reduces the tensile stress in the perpendicular direction, thereby decreasing
the damage zone.

• The embedded length of the wooden dowels in the concrete is shorter due to the smaller
thickness of the concrete compared to the CLT. To achieve an efficient connection and
ensure similar resistance on both sides of the shear plane, using denser timber material
provided optimized results. Wooden dowels made of oak, being a hardwood, did not
exceed the tensile strength perpendicular to the grain direction and demonstrated better
tensile resistance perpendicular to the fiber direction.

• According to the parameter study decreasing the depth of the notch from 30 mm to 20
mm to have more thickness in the notch reduced tensile stress parallel and perpendicular
to the fiber in the CLT layer and reduced the probability of brittle failure in CLT.

• Maximum vertical displacement was larger in the mid-span with an uplifting effect at the
end of the specimen.

• These findings offer important guidance for designing and implementing TCC floors with
disassembly capability, aiding the development of environmentally friendly and reliable
structural systems within the construction industry.

65
Future study

This master thesis raises several questions that should be further investigated to enhance knowl-
edge and understanding of the structural behavior and design methods for the TCC floors.

• An experimental test on this designed TCC slab to see its performance in practice and
compare with obtained numerical results.

• Evaluating the TCC slab in the Ultimate Limit State.

• Evaluating the TCC slab under different configuration angles of this novel connector.

• Evaluating the performance of this novel connector by considering the shrinkage of

wooden dowels and comparing the results with this study.

• Defining optimized depth of concrete layer in this TCC slab.

• Evaluating the performance of the TCC slab with different notch shapes in both SLS and
ULS.

66
References
[1] Abaqus version 6.6 documentation,abaqus analysis user’s manual.

[2] Clt handbook: cross-laminated timber,clt structures – design and detailing,.

[3] En 1992-1-1 (2004), eurocode 5: Eurocode 2: Design of concrete structures - part 1-

1: General rules and rules for buildings [authority: The european union per regulation
305/2011, directive 98/34/ec, directive 2004/18/ec].

[4] En1995-1-1, eurocode 5: Design of timber structures - part 1-1: General - common rules
and rules for buildings, [authority: The european union per regulation 305/2011, directive
98/34/ec, directive 2004/18/ec] cen, 2004.

[5] https://www.3ds.com/products/simulia.

[6] ipcc, global warming of 1.5 ºc.

[7] Ayman Abd-Elhamed, Sayed Mahmoud, and Khalid Saqer Alotaibi. Nonlinear analysis of
reinforced concrete buildings with different heights and floor systems. Scientific Reports,
13(1):14949, 2023.

[8] Yong Han Ahn and Kyoon-Tai Kim. Sustainability in modular design and construction:
A case study of ‘the stack’. International Journal of Sustainable Building Technology and
Urban Development, 5(4):250–259, 2014.

[9] Faisal Alazzaz and Andrew Whyte. Uptake of off-site construction: benefit and future
application. International Journal of Civil and Environmental Engineering, 8(12):1219–
1223, 2014.

[10] Amr S Allam and Mazdak Nik-Bakht. From demolition to deconstruction of the built
environment: A synthesis of the literature. Journal of Building Engineering, 64:105679,
2023.

[11] Ahmad L Almutairi, Bassam A Tayeh, Adeyemi Adesina, Haytham F Isleem, and Abdul-
lah M Zeyad. Potential applications of geopolymer concrete in construction: A review.
Case Studies in Construction Materials, 15:e00733, 2021.

[12] YH Mugahed Amran, Rayed Alyousef, Hisham Alabduljabbar, and Mohamed El-
Zeadani. Clean production and properties of geopolymer concrete; a review. Journal
of Cleaner Production, 251:119679, 2020.

[13] Standard ASTM. Standard test methods for flexural properties of unreinforced and re-
inforced plastics and electrical insulating materials. astm d790. Annual book of ASTM
Standards, 1997.

[14] Yingwei Bao, Weidong Lu, Kong Yue, Hao Zhou, Binhui Lu, and Zhentao Chen. Struc-
tural performance of cross-laminated timber-concrete composite floors with inclined self-
tapping screws bearing unidirectional tension-shear loads. Journal of Building Engineer-
ing, 55:104653, 2022.

[15] Hans Joachim Blaß and Carmen Sandhaas. Timber engineering-principles for design.
KIT Scientific Publishing, 2017.

67
[16] Eric Borgström and R Karlsson. Design of timber structures—structural aspects of timber
construction. Swedish Forest Industries Federation, Stockholm, 2016.

[17] José M Cabrero, Kifle G Gebremedhin, and Jorge Elorza. Evaluation of failure criteria in
wood members. In 2009 Reno, Nevada, June 21-June 24, 2009, page 1. American Society
of Agricultural and Biological Engineers, 2009.

[18] Ario Ceccotti. Composite concrete-timber structures. Progress in structural engineering

and materials, 4(3):264–275, 2002.

[19] S Cuerrier-Auclair. Design guide for timber-concrete composite floors in canada. FPIn-
novations, Montréal, 2020.

[20] Jesper Sand Damtoft, Jacques Lukasik, Duncan Herfort, Danielle Sorrentino, and El-
lis Martin Gartner. Sustainable development and climate change initiatives. Cement and
concrete research, 38(2):115–127, 2008.

[21] Joseph R Davis. Tensile testing. ASM international, 2004.

[22] Mieke De Schepper, Philip Van den Heede, Isabel Van Driessche, and Nele De Belie. Life
cycle assessment of completely recyclable concrete. Materials, 7(8):6010–6027, 2014.

[23] Mohammad Derikvand and Gerhard Fink. Deconstructable connector for tcc floors using
self-tapping screws. Journal of Building Engineering, 42:102495, 2021.

[24] Wit Derkowski. Large panels buildings–the possibilities of modern precast industry.
Cement-Wapno-Beton= Cement Lime Concrete, 22(5):414–425, 2017.

[25] Wit Derkowski. New solutions for prefabricated fl oor slabs. Cement-Wapno-Beton =
Cement Lime Concrete, 24(5):372–382, 2019.

[26] Alfredo Dias. Design of timber-concrete composite structures: A state-of-the-art report

by COST Action FP1402/WG 4. Shaker Verlag, 2018.

[27] AMPG Dias, Helena Cruz, S Lopes, and JWG Van de Kuilen. Experimental shear-friction
tests ondowel type fastener timber-concrete joints. In 8th World Conference on Timber
Engineering, pages 305–308. WCTE 2004 Secretariat, 2004.

[28] AMPG Dias, Jan Willem Van de Kuilen, S Lopes, and Helena Cruz. A non-linear 3d
fem model to simulate timber–concrete joints. Advances in Engineering Software, 38(8-
9):522–530, 2007.

[29] Linnaeus University Faculty of Technology Department of Building Technology Dorn,

Michael. Timber strength criteria via uvarm in abaqus. DiVA, 2019.

[30] CEN EN. 1-1. eurocode 5: Design of timber structures-part 1-1: General-common rules
and rules for buildings. CEN, Brussels, Belgium, 2004.

[31] Hooman Eslami, Laddu Bhagya Jayasinghe, and Daniele Waldmann. Experimental and
numerical investigation of a novel demountable timber–concrete composite floor. Build-
ings, 13(7):1763, 2023.

[32] Parlamento Europeu. Circular economy: definition, importance and benefits, 2022.

68
[33] Samer Fawzy, Ahmed I Osman, John Doran, and David W Rooney. Strategies for mit-
igation of climate change: a review. Environmental Chemistry Letters, 18:2069–2094,
2020.
[34] Paolo Foraboschi, Mattia Mercanzin, and Dario Trabucco. Sustainable structural design
of tall buildings based on embodied energy. Energy and Buildings, 68:254–269, 2014.
[35] Albin Forsberg and Filip Farbäck. Timber concrete composite floors with cross laminated
timber-structural behavior & design. TVBK-20/5279, 2020.
[36] M Fragiacomo and E Lukaszewska. Time-dependent behaviour of timber–concrete com-
posite floors with prefabricated concrete slabs. Engineering Structures, 52:687–696,
2013.
[37] S Gagnon and C Pirvu. Clt handbook, fpinnovations. Quebec, Canada: Special Publica-
tion SP-528E, 2011.
[38] M Gharib, A Hassanieh, H Valipour, and MA Bradford. Three-dimensional constitutive
modelling of arbitrarily orientated timber based on continuum damage mechanics. Finite
Elements in Analysis and Design, 135:79–90, 2017.
[39] Milad Hafezolghorani, Farzad Hejazi, Ramin Vaghei, Mohd Saleh Bin Jaafar, and Keyhan
Karimzade. Simplified damage plasticity model for concrete. Structural engineering
international, 27(1):68–78, 2017.
[40] MZ Hailu. Long-term performance of timber-concrete composite flooring systems. PhD
thesis, 2015.
[41] Annette Harte. Introduction to timber as an engineering material. ICE manual of con-
struction materials, 2:707–715, 2009.
[42] Lizhen Huang, Guri Krigsvoll, Fred Johansen, Yongping Liu, and Xiaoling Zhang. Car-
bon emission of global construction sector. Renewable and Sustainable Energy Reviews,
81:1906–1916, 2018.
[43] I Iea. World energy statistics and balances, iea. France, 2019.
[44] Shahana Janjua, Prabir Sarker, and Wahidul Biswas. A review of residential buildings’
sustainability performance using a life cycle assessment approach. Journal of Sustain-
ability Research, 1(1):1–29, 2019.
[45] Tomasz Jankowiak and Tomasz Lodygowski. Identification of parameters of concrete
damage plasticity constitutive model. Foundations of civil and environmental engineering,
6(1):53–69, 2005.
[46] Jouri Kanters. Design for deconstruction in the design process: State of the art. Buildings,
8(11):150, 2018.
[47] Dudley Charles Kent and Robert Park. Flexural members with confined concrete. Journal
of the structural division, 97(7):1969–1990, 1971.
[48] Nima Khorsandnia, Hamid Valipour, and Mark Bradford. Deconstructable timber-
concrete composite beams with panelised slabs: Finite element analysis. Construction
and Building Materials, 163:798–811, 2018.

69
[49] Nima Khorsandnia, Hamid Valipour, Jörg Schänzlin, and Keith Crews. Experimental
investigations of deconstructable timber–concrete composite beams. Journal of Structural
Engineering, 142(12):04016130, 2016.

[50] Mustafa MA Klufallah, Muhd Fadhil Nuruddin, Mohd Faris Khamidi, and Nazhatulzalkis
Jamaludin. Assessment of carbon emission reduction for buildings projects in malaysia-a
comparative analysis. In E3S web of conferences, volume 3, page 01016. EDP Sciences,
2014.

[51] R Lakshmi and S Nagan. Investigations on durability characteristics of e-plastic waste

incorporated concrete. 2011.

[52] Serge Lamothe, Luca Sorelli, Pierre Blanchet, and Philippe Galimard. Engineering ductile
notch connections for composite floors made of laminated timber and high or ultra-high
performance fiber reinforced concrete. Engineering Structures, 211:110415, 2020.

[53] Jeeho Lee and Gregory L Fenves. Plastic-damage model for cyclic loading of concrete
structures. Journal of engineering mechanics, 124(8):892–900, 1998.

[54] Jonna Ljunge and Helena Nerhed Silfverhjelm. Reuse of structural clt elements: Assess-
ing the impact of inter-element joint solutions on the reuse potential and environmental
impact of a load-bearing wall panel. 2022.

[55] Louisa Loulou. Durabilité d’un assemblage mixte bois-béton collé sous chargement hy-
drique. PhD thesis, Université Paris-Est, 2013.

[56] Jacob Lubliner, Javier Oliver, Sand Oller, and Eugenio Oñate. A plastic-damage model
for concrete. International Journal of solids and structures, 25(3):299–326, 1989.

[57] Elzbieta Lukaszewska, Massimo Fragiacomo, and Helena Johnsson. Laboratory tests
and numerical analyses of prefabricated timber-concrete composite floors. Journal of
structural engineering, 136(1):46–55, 2010.

[58] S Mahaboob Basha, C Bhupal Reddy, and K Vasugi. Strength behaviour of geopolymer
concrete replacing fine aggregates by m-sand and e-waste. Int. J. Eng. Trends Technol,
40:401–407, 2016.

[59] Leonora Charlotte Malabi Eberhardt, Anne van Stijn, Freja Nygaard Rasmussen, Morten
Birkved, and Harpa Birgisdottir. Development of a life cycle assessment allocation ap-
proach for circular economy in the built environment. Sustainability, 12(22):9579, 2020.

[60] Cristoffer Mård. Composite timber structures–ribbed plate design: Evaluation of existing
and development of new design methods, 2022.

[61] Karl Möhler. Über das Tragverhalten von Biegeträgern und Druckstäben mit zusam-
mengesetzten Querschnitten und nachgiebigen Verbindungsmitteln. PhD thesis, 1956.

[62] João Henrique Jorge de Oliveira Negrão, Catarina Alexandra Leitão de Oliveira, Fran-
cisco Miguel Maia de Oliveira, and Paulo Barreto Cachim. Glued composite timber-
concrete beams. i: Interlayer connection specimen tests. Journal of Structural Engineer-
ing, 136(10):1236–1245, 2010.

70
[63] Adam M Neville and Jeffrey John Brooks. Concrete technology, volume 438. Longman
Scientific & Technical England, 1987.

[64] CB Norris. Strength of orthotropic materials subjected to combined stress us forest prod-
ucts. Laboratory Report, (1816), 1962.

[65] Vanthet Ouch, Piseth Heng, Hugues Somja, and Thierry Soquet. An experimental and
numerical investigation on a dovetail notched connection for cross-laminated-timber-
concrete composite slabs. In 13th World Conference on Timber Engineering: Timber
for a Livable Future, WCTE 2023, volume 6, pages 3333–3341. World Conference on
Timber Engineering (WCTE), 2023.

[66] Hoai Son Pham. Optimisation et comportement en fatigue de la connexion bois-BFUP

pour de nouveaux ponts mixtes. PhD thesis, Marne-la-vallée, ENPC, 2007.

[67] Ingolf Profft, Martina Mund, Georg-Ernst Weber, Eberhard Weller, and Ernst-Detlef
Schulze. Forest management and carbon sequestration in wood products. European jour-
nal of forest research, 128:399–413, 2009.

[68] Shahryar Qazizadeh, SK Hashemi, and HR Valipour. Numerical investigation of steel-

timber composite floors with flush end-plate connection. In Structures, volume 56, page
104892. Elsevier, 2023.

[69] Robert J Ross. Wood handbook: wood as an engineering material. 2021.

[70] Ali Sadrmomtazi, Negar Ghasemi Khameneh, Reza Kohani Khoshkbijari, and Morteza
Amooie. A study on the durability of the slag-based geopolymer concretes containing bi-
nary solid mixtures in corrosive environments. Revista Romana de Materiale, 51(2):195–
206, 2021.

[71] Michael Sansom and Roger J Pope. A comparative embodied carbon assessment of com-
mercial buildings. The Structural Engineer, 90(10):38–49, 2012.

[72] Faiz Uddin Ahmed Shaikh. Mechanical and durability properties of fly ash geopolymer
concrete containing recycled coarse aggregates. International Journal of Sustainable Built
Environment, 5(2):277–287, 2016.

[73] Eric Steinberg, Ricky Selle, and Thorsten Faust. Connectors for timber–lightweight con-
crete composite structures. Journal of structural engineering, 129(11):1538–1545, 2003.

[74] strongtie.eu. Application of the european construction standards, eks 12, 2023. Available
online.

[75] strongtie.eu. Connectors and fasteners for cross-laminated timber structures, 2023. Ac-
cessed on 2024-04-25.

[76] Yusuf Sümer and Muharrem Aktaş. Defining parameters for concrete damage plasticity
model. Challenge Journal of Structural Mechanics, 1(3):149–155, 2015.

[77] Bassam A Tayeh, Abdullah M Zeyad, Ibrahim Saad Agwa, and Mohamed Amin. Effect
of elevated temperatures on mechanical properties of lightweight geopolymer concrete.
Case Studies in Construction Materials, 15:e00673, 2021.

71
[78] Joakim Thilén. Testing of clt-concrete composite decks. TVBK-5259, 2017.

[79] TACM Van der Put. The tensorpolynomial failure criterion for wood. Delft Wood Science
Foundation, Delft, 2005.

[80] Mingqian Wang, Qingfeng Xu, Kent A Harries, Lingzhu Chen, Zhenpeng Wang, and
Xi Chen. Experimental study on mechanical performance of shear connections in clt-
concrete composite floor. Engineering Structures, 269:114842, 2022.

[81] Qingfeng Xu, Mingqian Wang, Lingzhu Chen, Kent A Harries, Xiaobing Song, and Zhen-
peng Wang. Mechanical performance of notched shear connections in clt-concrete com-
posite floor. Journal of Building Engineering, 70:106364, 2023.

[82] David Eng Chuan Yeoh. Behaviour and design of timber-concrete composite floor system.
2010.

72
Appendix

This appendix contains an extract from analytical calculations according to EN1995, Annex B.
To accurately model the behavior of a composite structural system comprised of Cross-Laminated
Timber (CLT) and precast concrete component which is called TCC, it is necessary to account for the
connections between the concrete and the timber as well as the connections within the longitudinal
layers of the CLT element. Given the absence of experimental validation for this composite slab in order
to be in the safe side, Smeared slip modulus of notch only be considered.

L ≔ 1.15 ⋅ 2 m = 2.3 m

b ≔ 500 mm

hc ≔ 60 mm

ht ≔ 40 mm

Ac ≔ b ⋅ hc = 0.03 m 2

At ≔ ht ⋅ b = 0.02 m 2
kN
ρC ≔ 24 ――
m3
kN
ρCLT_mean ≔ 4.2 ――
m3

Load calculation:

Imposed load:
kN
qk ≔ 2.5 ――
m2
kN
mpartition ≔ 1 ――
m2
kN
wself_weight ≔ ρCLT_mean ⋅ 3 ⋅ ht + ρC ⋅ hc + mpartition = 2.944 ――
m2
kN
mself_weight ≔ wself_weight ⋅ b = 1.472 ――
m
According to A1.4 in Annex A in EN 1990, Table 8, Characteristic, Frequent and Quasi-permanent load
combinations are defined.

Characteristic load (for checking of cracking at bottom of slab):

qSurf_k ≔ wself_weight + qk = 0.005 MPa

Non-Commercial
i Use Only
According to ''Table Al.I , Annex Al of Application for Buildings EN 1990 value for ψ for category B
is extracted.

ψ1 ≔ 0.5

ψ0 ≔ 0.7

ψ2 ≔ 0.3

Service stage - frequent (for checking of reversible effects like deflection):

kN
qSurf_F ≔ wself_weight + ψ1 ⋅ qk = 4.194 ――
m2

Quasi-permanent load (for checking of long term effects)

qSurf_Q ≔ wself_weight + ψ2 ⋅ qk = 0.004 MPa

Ÿ EN 1990 ,Load combination 6.10b , Table 6, section''2.1.2,2'' for Ultimate limit state (ULS)
Ÿ partial factor from section ''2.1.2.2 Structural failure'' in ''EN 1990 Basis of structural design''
for RC2 is extracted.

γd ≔ 0.91
kN
qSurf_U ≔ γd ⋅ 0.89 ⋅ 1.35 ⋅ wself_weight + γd ⋅ 1.5 ⋅ qk = 6.631 ――
m2

Material Properties:

-Concrete slab – concrete grade C30/37:

Ecm ≔ 32840 MPa

αcc : the coefficient taking account of long term effects on the compressive strength and of unfavorable
effects resulting from the way the load is applied, EN 1992-1-1, section 2.4.2.4 (1)

αcc ≔ 1

Non-Commercial
ii Use Only
γc :Partial factors for materials, EN 1992-1-1, section 2.4.2.4 (1)

γc ≔ 1.5

EN 1992-1-1, section 3.1.3, Table 3.1.

fck ≔ 30 MPa

EN 1992-1-1, section 3.1.6 (1)

fck
fcd ≔ αcc ⋅ ―― = 20 MPa
γc

αct :a coefficient taking account of long term effects on the tensile strength and of unfavorable effects,
resulting from the way the load is applied, EN 1992-1-1, section 3.1.6
αct ≔ 1

EN 1992-1-1section 3.1.3, Table 3.1.

fctk.0.05 ≔ 2 MPa

fctd : EN 1992-1-1, section 3.1.6 (2)

fctk.0.05
fctd ≔ αct ⋅ ――― = 1.333 MPa
γc

-Timber C24:

Ei_0mean ≔ 11000 MPa

Gi_0mean ≔ 690 MPa

Ei_90mean ≔ 370 MPa

Gi_rmean ≔ 80 MPa

fmk ≔ 24 MPa

ft0k ≔ 14.5 MPa

Non-Commercial
iii Use Only
ft90k ≔ 0.4 MPa

fc0k ≔ 21 MPa

fc90k ≔ 2.5 MPa

fvk ≔ 4 MPa

Residential building interior,[SS-EN 1995-1-1, 2.3.1.3]

Service class=1

Load-duration class, Medium term action (Imposed floor load),[Design_of_timber_structure_Vo2.pdf,

Table2,1]

Modification factor, [The CLT handbook, Table 2.2] for CLT

Kmod_sh ≔ 0.9 Service class1, short term

Kmod_L ≔ 0.7 Service class1, long term

Deformation factor, [The CLT handbook, Table 2.3]

Kdef ≔ 0.8

The CLT Handbook section 3.1.5, table 3.2.

γM ≔ 1.25

Design value for strength in Ultimate Limit State(The CLT Handbook, section 3.1):
Kmod_L ⋅ fmk
fmd ≔ ――――= 13.44 MPa
γM

Kmod_L ⋅ ft0k
ft0d ≔ ――――= 8.12 MPa
γM

Kmod_L ⋅ fvk
fvd ≔ ―――― = 2.24 MPa
γM

Kmod_L ⋅ fc90k
fcd_t ≔ ―――― = 1.4 MPa
γM

Non-Commercial
iv Use Only
-Connection:
Type of fasteners: Notch and wooden dowel with screws

α ≔ 90 ° Screw orientation

ds ≔ 20 mm screw diameter

ls ≔ 180 mm screw length

S ≔ 300 mm distance between screws

n≔1 number of screws in width direction

-Due to lack of experiment and not being sure about the influence length of the notch one eight of span
length to be in the safe side was considered.

L
Snotch ≔ ―= 287.5 mm
8

fyd ≔ 240 MPa Design yield strength

Slip modulus for fastener per shear plane and per fastener, [Design_of_timber_structure_Vo2.pdf, Table
9.2]
⎛⎛ m3 ⎞
1.5
1 ⎞
n ⋅ 2 ⎜⎜ρCLT_mean ⋅ ―― ⎟ ⋅ ds ⋅ ―― ⎟
⎝⎝ kN ⎠ mm ⎠ N N
Ks_b ≔ ―――――――――――⋅ ―― = 14.969 ――
23 mm mm

Smeared slip modulus(Connection stiffness per unit length):

N
――
mm
ks_n ≔ 1500 ―― for depth of the notch larger than 30mm
mm

kN
――
Ks_b mm
ks_b ≔ ―― = ⎛⎝4.99 ⋅ 10 -5⎞⎠ ――
S mm

⎛ π 2 Ecm ⋅ Ac ⎞
-1

⎜
γ1 ≔ 1 + ―――― ⎟ = 0.449
⎜⎝ ks_n ⋅ L 2 ⎟⎠

Non-Commercial
v Use Only
 ⎛ π 2 ⋅ Ei_0mean ⋅ At ⋅ ht ⎞
-1

γ3 ≔ ⎜1 + ―――――― ⎟ = 0.709
⎜⎝ Gi_rmean ⋅ L 2 ⋅ b ⎟⎠

γ2 ≔ 1

⎛ hc ht ⎞ ⎛ ht ht ⎞
γ1 ⋅ Ecm ⋅ Ac ⋅ ⎜―+ ―⎟ - γ3 ⋅ Ei_0mean ⋅ At ⋅ ⎜―+ ht + ―⎟
⎝2 2⎠ ⎝2 2⎠
d_NL1 ≔ ――――――――――――――――― = 11.796 mm
γ1 ⋅ Ecm ⋅ Ac + γ2 ⋅ Ei_0mean ⋅ At + γ3 ⋅ Ei_0mean ⋅ At

Since d_NL1 is positive it means the neutral axis is above the center of gravity of the second layer and
it is 11.796 mm above the center of the first layer of CLT.

⎛ hc ht ⎞
ac ≔ ⎜―+ ―- d_NL1⎟ = 38.204 mm
⎝2 2 ⎠

ht ht
at ≔ ―+ ht + ―+ d_NL1 = 91.796 mm
2 2
3
EIeff ≔ ∑ Ei ⋅ Ii + γi ⋅ Ei ⋅ Ai ⋅ ai 2
n=1

In case of composite action from the Analytical method( γ - ((method)) :

hc 3 ht 3
EIeff_sh ≔ Ecm ⋅ b ⋅ ―― + γ1 ⋅ Ecm ⋅ b ⋅ hc ⋅ ac 2 + Ei_0mean ⋅ b ⋅ ―― + γ2 ↲ = ⎛⎝2.345 ⋅ 10 3 ⎞⎠ kN ⋅ m 2
12 12
ht 3
⋅ Ei_0mean ⋅ b ⋅ ht ⋅ d_NL1 2 + Ei_0mean ⋅ b ⋅ ―― + γ3 ⋅ Ei_0mean ⋅ b ⋅ ht ⋅ at 2
12

In case of assumption of No composite action:

γ1_N ≔ 0
hc 3 ht 3
EIeff_N ≔ Ecm ⋅ b ⋅ ―― 2
+ γ1_N ⋅ Ecm ⋅ b ⋅ hc ⋅ ac + Ei_0mean ⋅ b ⋅ ―― + γ2 ↲ = ⎛⎝1.699 ⋅ 10 3 ⎞⎠ kN ⋅ m 2
12 12
3
ht
⋅ Ei_0mean ⋅ b ⋅ ht ⋅ d_NL1 2 + Ei_0mean ⋅ b ⋅ ―― + γ3 ⋅ Ei_0mean ⋅ b ⋅ ht ⋅ at 2
12

Non-Commercial
vi Use Only
In case of assumption of Full composite action:

γ1_F ≔ 1
hc 3 ht 3
EIeff_F ≔ Ecm ⋅ b ⋅ ―― + γ1_F ⋅ Ecm ⋅ b ⋅ hc ⋅ ac 2 + Ei_0mean ⋅ b ⋅ ―― + γ2 ↲ = ⎛⎝3.137 ⋅ 10 3 ⎞⎠ kN ⋅ m 2
12 12
ht 3
⋅ Ei_0mean ⋅ b ⋅ ht ⋅ d_NL1 2 + Ei_0mean ⋅ b ⋅ ―― + γ3 ⋅ Ei_0mean ⋅ b ⋅ ht ⋅ at 2
12

Comparison between effective bending stiffness from γ -method, Full and No composite
action:

EIeff_N ≤ EIeff_sh ≤ EIeff_F = 1

Checking of the SLS criteria in short term :

For checking maximum deflection of the mid-span and compare it with finite element modeling result,
since the load applied in numerical modeling is by displacement-control the applied displacement is
compared with Eurocode 5 requirement.

The deflection in SLS(short-term):

5 ⋅ qSurf_k ⋅ L 4 ⋅ b
Δy ≔ ―――――= 0.423 mm Short term deformation
384 ⋅ EIeff_sh

Deflection limit:
L
vlimit ≔ ―― = 7.667 mm
300

Δy ≤ vlimit = 1 Passed

Non-Commercial
vii Use Only
Vibration check:
kg kg
mself_weight ≔ qSurf_k ⋅ b ⋅ 1 ――= 272.2 ―
10 N m

π 2 ‾‾‾‾‾‾‾‾‾
EIeff_sh
f1 ≔ ―― ⋅ ―――― = 27.563 Hz
2 ⋅ L2 mself_weight

f1 ≥ 8 Hz = 1 Passed

Calculation for stress in short-term :

a ≔ 700 mm
ht
zend ≔ ―+ d_NL1 = 0.032 m
2
Checking that the ULS load bearing capacity is exceeded or not.
qSurf_k ⋅ L 2 ⋅ b
Mu ≔ ――――― = 1.8 kN ⋅ m
8

L
Vu ≔ qSurf_k ⋅ b ⋅ ―= 3.13 kN
2

Normal stresses due to bending of the layer in concrete layer:

Mu hc
σc.bending ≔ ――― ⋅ Ecm ⋅ ―= 0.756 MPa
EIeff_sh 2

Normal stresses due to composite behavior in concrete layer:

Mu
σc.normal ≔ ――― ⋅ γ1 ⋅ Ecm ⋅ ac = 0.433 MPa
EIeff_sh
σc.top ≔ ||-σc.bending - σc.normal|| ≤ fcd = 1 passed

σc.Bottom ≔ σc.bending - σc.normal ≤ fctd = 1 passed

Non-Commercial
viii Use Only
Normal stresses due to bending of the layer in timber part:
Mu ht
σt.bending ≔ ――― ⋅ Ei_0mean ⋅ ―= 0.169 MPa
EIeff_sh 2

Normal stresses due to composite behavior in timber layers:

Mu
σt.normal_3 ≔ ――― ⋅ γ3 ⋅ Ei_0mean ⋅ at = 0.549 MPa
EIeff_sh

Mu
σt.normal_1 ≔ ――― ⋅ γ2 ⋅ Ei_0mean ⋅ d_NL1 = 0.1 MPa
EIeff_sh

σt.bending σt.normal_1
Degreeofuse ≔ ―――+ ―――― ≤1=1 passed
fmd ft0d

σt.bending σt.normal_3
Degreeofuse ≔ ―――+ ―――― ≤1=1 passed
fmd ft0d

Stress results:

σc.top ≔ -σc.bending - σc.normal = -1.189 MPa

σc.Bottom ≔ σc.bending - σc.normal = 0.323 MPa

σt1.top ≔ -σt.bending - σt.normal_1 = -0.268 MPa

σt1.Bottom ≔ σt.bending + σt.normal_1 = 0.268 MPa

σt3.top ≔ -σt.bending + σt.normal_3 = 0.381 MPa

σt3.Bottom ≔ σt.bending + σt.normal_3 = 0.718 MPa

Ai ≔ b ⋅ ht = 0.02 m 2

Non-Commercial
ix Use Only
Maximum shear stress: (only for layers containing the neutral axis are valid)

Vu ⋅ ⎛⎝γ3 ⋅ Ei_0mean ⋅ Ai ⋅ at + γ2 ⋅ Ei_0mean ⋅ b ⋅ ⎛⎝zend⎞⎠ 2 ⋅ 0.5⎞⎠

τmax ≔ ――――――――――――――――― = 0.046 MPa
EIeff_sh ⋅ b

Shear stress at the joint:

Vu ⋅ γ1 ⋅ Ecm ⋅ Ac ⋅ ac
τjoint_1_2 ≔ ――――――= 0.045 MPa
EIeff_sh ⋅ b

Vu ⋅ ⎛⎝γ1 ⋅ Ei_0mean ⋅ Ac ⋅ at⎞⎠

τjoint_2_3 ≔ ―――――――― = 0.036 MPa
EIeff_sh ⋅ b

※ In joint between layer 1 and 2 it is assumed dominate connection was the Notch and its effective
length for spacing is considered.

F1 ≔ τjoint_1_2 ⋅ b ⋅ Snotch = 6.49 kN

F3 ≔ τjoint_2_3 ⋅ b ⋅ Snotch = 5.223 kN

τjoint_2_3 ≤ fvd = 1 So by fulfilling this criteria the connections in timber can withstand
shear forces.
τjoint_1_2 ≤ fcd = 1

Design value of shear stress in the interface between concrete and timber layer.(Eurocode 2, section
6.2.5(1))

vED ≔ τjoint_1_2

Non-Commercial
x Use Only
To define the design shear resistance at the interface equation (6-25) of Eurocode 2 is used.

Ÿ Since prefabricated concrete is utilized there is no cohesion between layers, and c=0 was
considered.
Ÿ in case of wet-dry connection friction coefficient according to Dias et al., experiment is 0.5.
Ÿ However for case of dry-dry connection this value would be so smaller by considering probable
executive errors like erecting issues. So in order to be in the safe side one third of this value was
considered.

μ ≔ 0.4

σn ≔ σc.normal = 0.433 MPa stress per unit area

σn ≤ 0.6 ⋅ fcd = 1 Passed

ds 2
As ≔ n ⋅ π ⋅ ―― = ⎛⎝3.142 ⋅ 10 -4⎞⎠ m 2
4
Ai ≔ b ⋅ ⎛⎝3 ⋅ ht + hc⎞⎠ = 0.09 m 2
As
ρ ≔ ―= 0.003
Ai
VRdi ≔ μ ⋅ σn + ρ ⋅ fyd ⋅ ((μ ⋅ sin ((α)) + cos ((α)))) = 0.508 MPa
⎛ fck ⎞
v ≔ 0.6 ⋅ ⎜1 - ―――― ⎟ = 0.528
⎝ 250 MPa ⎠
0.5 v ⋅ fcd = 5.28 MPa

VRdi ≤ 0.5 v ⋅ fcd = 1 Passed

Check of interface:
τjoint_1_2 ≤ VRdi = 1 Passed, it means the concrete at interface can also withstand against
shear.

Non-Commercial
xi Use Only
-Calculation of Bending stiffness from Finite element modeling(Abaqus):

PMax ≔ 33 kN Maximum vertical reaction force of the support (FEM)

w ≔ 4.25 mm Mid-span deflection after loading (FEM)

PMax ⋅ a ⋅ ⎛⎝3 ⋅ L 2 - 4 ⋅ a 2 ⎞⎠
EIeff_FEM ≔ ―――――――― = ⎛⎝3.15 ⋅ 10 3 ⎞⎠ kN ⋅ m 2
24 ⋅ w

EIeff_sh
ratio ≔ ―――― = 0.745
EIeff_FEM

Non-Commercial
xii Use Only