J.

Stal

Transverse Shear Capacity

of Deep Composite Slabs
Based on a Finite Element Analysis of ComFlor 210
 Transverse Shear Capacity
of Deep Composite Slabs
Based on a Finite Element Analysis of ComFlor 210

by

F.J. Stal

in partial fulfilment of the requirements for the degree of

Master of Science
in Civil Engineering

at the Delft University of Technology,

to be defended publicly on Thursday February 13, 2020 at 4:00 PM.

Supervisor: Ir. J.M. Houben TU Delft

Thesis committee: Dr. ir. R. Abspoel TU Delft
Dr. ir. Y. Yang TU Delft

An electronic version of this thesis is available at http://repository.tudelft.nl/.

Preface

Writing a thesis may be experienced as a period in life with a lot of ups and downs. At least, this was
the case for me. After defining the subject, I found it very difficult to start. I read multiple papers and
gathered a lot of information, but this only made me worried about how to make it to the end. As an
advice to all my fellow graduation students: relax and take one step at a time. Sometimes, finite element
software gives unexpected results. Other times, you may come up with questions even your supervisors
don’t have clear answers to. Be creative, self-reflective and discuss your problems. In the end, you’ll
find satisfaction.

First of all, I would like to thank Lambert Houben for taking over the role as supervisor from Dick Hordijk,
as he retired from the TU Delft. I haven’t experienced any adverse consequences of this change.
Secondly, I would like to thank Paul Lagendijk for giving me the opportunity to graduate on this topic,
since he came up with the subject from his daily work as structural engineer. The various meetings with
him and Roland Abspoel really helped me getting through the process. Thank you both for all your
valuable feedback. Since Paul also left the University, I would like to thank Yuguang Yang for taking
over his role in the thesis committee and helping me finishing the project with his expertise on the
transverse shear behaviour of concrete. Furthermore, I would like to thank Michele Longo for helping
me with the questions that arose from finite element modelling. At last, special thanks go to:
▪ my parents, for the support and giving me their full trust in succeeding;
▪ my girlfriend, for dealing with my complaints every now and then;
▪ my father-in-law, for letting me borrow his desktop when my own laptop failed in computational
power.

F.J. Stal
Dordrecht, February 2020
Contents

ABSTRACT ....................................................................................................................................................... I
1. INTRODUCTION....................................................................................................................................... 1
1.1. STEEL-CONCRETE COMPOSITE ELEMENTS .............................................................................................. 1
1.2. OBJECTIVE .......................................................................................................................................... 4
1.3. RESEARCH QUESTIONS ......................................................................................................................... 6
2. LITERATURE REVIEW............................................................................................................................. 7
2.1. COMPOSITE SLABS IN GENERAL ............................................................................................................. 7
2.1.1. Materials .................................................................................................................................... 7
2.1.2. Origin ......................................................................................................................................... 8
2.1.3. Production process .................................................................................................................... 8
2.1.4. Types......................................................................................................................................... 9
2.1.5. Applications ............................................................................................................................. 11
2.1.6. Advantages.............................................................................................................................. 11
2.1.7. Disadvantages ......................................................................................................................... 12
2.1.8. Fire safety ................................................................................................................................ 12
2.1.9. Deep composite slabs: ComFlor 210 and ComFlor 225 ............................................................ 12
2.2. COMPOSITE BEHAVIOUR ..................................................................................................................... 14
2.3. FAILURE MECHANISMS OF COMPOSITE SLABS ........................................................................................ 18
2.4. LONGITUDINAL SHEAR RESISTANCE OF COMPOSITE SLABS ACCORDING TO EUROCODE 4 ............................ 20
2.5. TRANSVERSE SHEAR RESISTANCE OF COMPOSITE SLABS ACCORDING TO EUROCODE 4 (REFERRING TO
EUROCODE 2: CONCRETE STRUCTURES) .......................................................................................................... 23
2.6. TRANSVERSE SHEAR RESISTANCE OF STEEL DECKS ACCORDING TO EUROCODE 3 ..................................... 25
2.7. TRANSVERSE SHEAR RESISTANCE OF COMPOSITE SLABS ACCORDING TO LITERATURE ............................... 25
3. FINITE ELEMENT ANALYSIS ................................................................................................................ 35
3.1. GEOMETRY AND LOADING CONDITIONS ................................................................................................. 35
3.2. FEA CONCRETE SECTION OF COMFLOR 210 ......................................................................................... 38
3.2.1. Material properties ................................................................................................................... 39
3.2.2. Analysis procedure .................................................................................................................. 40
3.2.3. Results and discussion ............................................................................................................ 42
3.3. FEA COMFLOR 210 ........................................................................................................................... 47
3.3.1. Material properties ................................................................................................................... 49
3.3.2. Analysis procedure .................................................................................................................. 53
3.3.3. Results and discussion ............................................................................................................ 54
3.4. FEA COMFLOR 210: A LOWER BOUND VALUE FOR THE TRANSVERSE SHEAR CAPACITY .............................. 61
3.4.1. Material properties & analysis procedure .................................................................................. 62
3.4.2. Results and discussion ............................................................................................................ 64
4. DISCUSSION.......................................................................................................................................... 69
5. CONCLUSION ........................................................................................................................................ 77
BIBLIOGRAPHY ............................................................................................................................................ 79
APPENDIX A.................................................................................................................................................. 81
APPENDIX B.................................................................................................................................................. 84
APPENDIX C.................................................................................................................................................. 85
APPENDIX D.................................................................................................................................................. 87
Abstract

For calculation of the resistance of a composite slab against the transverse shear force, the
Eurocode 4 (composite structures) simply refers to the calculation procedures of the Eurocode 2
(concrete structures). It is assumed that the composite slab consists out of a consecutive range of
concrete ribs in its width direction, which are solely responsible for resisting the transverse shear force.
To calculate the transverse shear capacity of these concrete ribs, an empirical formula is used that was
originally derived for regular reinforced concrete beams (without stirrups). However, the concrete ribs
of the composite slab are created by the profile of the steel deck, making that each concrete rib is
accompanied by two steel webs on the sides. According to the Eurocode 3 (steel structures), these
webs of the steel deck have their own transverse shear capacity, which is neglected by the current
design approach defined in the Eurocode 4. Besides, the interaction between the steel deck and the
concrete may lead to an even higher transverse shear capacity of the composite slab. In this thesis, the
aforementioned two aspects, which are currently overlooked by the design principle of the Eurocode 4,
are further studied by means of non-linear finite element modelling.

The validation of this empirical formula of the Eurocode 2 for calculating the transverse shear capacity
of the concrete ribs is the first point of interest. From the finite element analysis (FEA) of the concrete
section of ComFlor 210, it is concluded that the prediction of the transverse shear capacity by the
Eurocode 2 is unnecessarily conservative. The study suggests to use the mean width of the concrete
rib (b0) in calculation, instead of the minimum width in the tensile area of the concrete rib (bw), as an
improvement to the method of the Eurocode 2.

In the next stage, the contribution of the steel deck to the transverse shear capacity of the composite
slab is studied. The exact bonding properties between the steel deck and the concrete (at the interface)
were not clear when the finite element model was developed, so some assumptions had to be made.
When assuming that the steel deck can’t separate from the concrete and the relative slip is restrained
in longitudinal direction by the embossments, an increase of 131.6% in transverse shear capacity is
found. Because of the assumed interface properties, the steel deck contributes to the total transverse
shear capacity in the following ways: it resists a part of the transverse shear force in its webs; it acts as
reinforcement to the concrete like a longitudinal rebar; it acts as reinforcement to the concrete like
stirrups. However, whether this stirrup-functioning of the steel deck’s webs is representative for the
actual transverse shear behaviour of deep composite slabs is being questioned, because it relies on
the assumption of no separation at the interface. Therefore, a second FEA of ComFlor 210 is executed
in which the interaction between the steel deck and the concrete is neglectable. Still, an increase of
51.4% in transverse shear capacity is found, which can be considered as a lower bound value.

At last, from the FEA results of this thesis, it can indeed be concluded that the current Eurocode 4
provides a unnecessarily conservative calculation method for the transverse shear capacity of
ComFlor 210. However, using a simple engineering model that adds up the partial resistances of the
concrete ribs and the steel deck’s webs, gives a better prediction while still being safe. For the partial
resistance of the concrete ribs, the empirical formula of the Eurocode 2 is used, but this parameter bw
is substituted by b0 as already mentioned in the foregoing. For the partial resistance of the steel deck’s
webs, the procedures of the Eurocode 3 are followed.

I
II
1. INTRODUCTION
As may have become clear from the title: this thesis is about the transverse shear capacity of deep
composite slabs. A small introduction on composite elements is provided within this chapter, whereby
it is explained why verification of the transverse shear capacity of deep composite slabs is the main
point of interest. The research questions are formulated afterwards, which will be tried to be answered
at the end of this thesis.

1.1. Steel-concrete composite elements

It is generally known that concrete needs reinforcement to restrain tensile stresses. This is traditionally
taken care of by using steel rebars. This combination of steel and concrete is known as “reinforced
concrete” and is globally used in designing structural safe buildings and bridges. However, other kinds
of systems with the use of steel and concrete are also possible, e.g. circular steel tubes filled with
concrete or steel girders connected to a concrete slab on top. These kinds of combinations are relatively
new compared to the traditional reinforced concrete and are labelled differently as “steel-concrete
composite elements” (or just “composite elements” as used in the sequel of this thesis).

Despite that steel and concrete are essentially two different materials, they still are completely
compatible and complementary to each other[1]:
▪ One important aspect is that both materials have almost the same thermal expansion coefficient:
12∙10-6/K for steel[30] and 10∙10-6/K for concrete[24]. This means that internally generated stresses
due to fluctuations in temperature are limited, because both materials expand and contract in good
accordance;
▪ When the concrete (partly) envelops the steel section, it provides corrosion protection and increased
fire resistance to the steel. This later can be an import aspect in the design of a building, because
the strength and stiffness properties of steel easily reduce at strongly elevated temperatures;
▪ The concrete part can give increased resistance to common instability phenomena of steel sections,
e.g. local buckling or lateral torsional buckling;
▪ The use of these two materials gives an ideal combination of strengths: concrete is the most efficient
in compression, while steel is in tension.

Above-mentioned reasons have played a major part in the success of composite elements. However,
a critical aspect is the connection between the different materials (especially for composite slabs and
beams). To explain this, imagine two identical linear-elastic timber beams with dimensions b x h laying
on top of each other. This is illustrated in Figure 1.1. When a load is applied on top and there is no
interaction between the beams, both separate beams will experience positive and negative strains. This
will give tensile and compressive stresses respectively. Additionally, a relative slip will be visible at the
ends of the beams. Now, when the timber beams are connected to each other by gluing them together
for instance, longitudinal shear stresses will develop at the interface. When these shear stresses won’t
be bigger than the capacity of the glue and the glue behaves very stiff, then this relative slip will not
occur. The beams fully operate together, which gives an increase in the total section modulus of the
“composite” beam.

1
 1 2 1 4
𝑊𝑛𝑜 𝑖𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑖𝑜𝑛 = 2 ∙ 6 ∙ 𝑏 ∙ ℎ2 = 6 ∙ 𝑏 ∙ ℎ2 𝑊𝑓𝑢𝑙𝑙 𝑖𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑖𝑜𝑛 = 6 ∙ 𝑏 ∙ (2ℎ)2 = 6 ∙ 𝑏 ∙ ℎ2

4 2
𝑊𝑓𝑢𝑙𝑙 𝑖𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑖𝑜𝑛 6 ∙ 𝑏 ∙ ℎ
= =2
𝑊𝑛𝑜 𝑖𝑛𝑡𝑒𝑟𝑎𝑐𝑡𝑖𝑜𝑛 2 2
6 ∙ 𝑏 ∙ ℎ

From these simple calculations, it may be established that the section modulus increases with a factor
of 2 due to application of the glue as longitudinal shear connection. As a result, the stresses in the glued
composite beam will also drop with a factor of 2 (σ = M/W). Not only the stresses will be lower, but also
the deflection since the composite beam becomes much stiffer due to application of the glue. From
Figure 1.1, it can be established that the bending stiffness EI is 4 times bigger in case of full interaction.
Furthermore, it can also be seen that for full interaction, the top beam is completely in compression as
the lower beam is in tension. Applying concrete on top and steel at the bottom would therefore be
effective in this case. This explains in a very simple manner the power of composite elements.

No interaction

Longitudinal shear connection;

Full interaction

Figure 1.1 – The effect of longitudinal shear connections[31]

It may be concluded that it is of importance to ensure interaction between the separate elements.
Sufficient longitudinal shear capacity should be provided at the interface. In case of steel-concrete
composite beams, this is usually taken care of by applying mechanical shear connectors. A typical
composite beam exists of a hot-rolled steel profile with a concrete slab on top (see Figure 1.2). Headed
studs are often used as connectors, which are welded to the upper flange of the steel section after
which the concrete is casted on top. The headed studs are surrounded with concrete after hardening,
which provides a certain degree of longitudinal shear connection. Based on the former example (Figure
1.1), it may be established that this connection gives a significant improvement in bending stiffness of
the total system. With the deflection criteria usually being normative for the steel section, it makes that
less steel is necessary to support the concrete slab. This reduction in steel weight can overcome the
additional costs for application of the headed studs. Furthermore, an important aspect of the connection
is that it has a ductile behaviour. A brittle connection can lead to a total loss of composite behaviour at
early stage of overloading, which is not desired. Figure 1.3 shows a typical test of headed studs as
used in composite beams. It is obvious to see that the use of headed studs provides a ductile
longitudinal shear connection.

Figure 1.2 – Typical composite beam: steel section connected to a concrete slab on top by headed
studs[1]

2
 Figure 1.3 – Shear test on headed studs[1]

Figure 1.4 – Main categories of composite elements[1]

Composite elements can be divided into 3 categories: beams (or girders), columns and slabs. Examples
of these kinds of elements are shown in Figure 1.4. Composite beams and columns are outside the
scope of this thesis: the focus lays on the structural behaviour of composite slabs. In general, composite
beams and slabs do have the most agreement on structural behaviour, since they are both mainly
subjected to bending and shear. Therefore, not only for composite beams, but also for composite slabs
it is of importance that the steel deck and the concrete topping act together. This can only be achieved
by ensuring the development of longitudinal shear stresses at the interface between the elements. In
composite slabs this is generally taken care of by applying embossments in the steel deck. These
embossments improve the longitudinal shear capacity at the interface due to mechanical interlock
between the steel deck and the concrete.

Embossments in
flanges and webs

Figure 1.5 – Steel decking (ComFlor 210) is put in position between supporting beams[37]

Composite slabs are generally applied as floor systems. It has turned out to be a safe way of building
in which the speed of construction is greatly improved [2]. It owes its performance to the smart way of
3
utilizing the intrinsic properties of steel and concrete in the same way as for composite beams. Profiled
steel decking is put in position between the supporting beams and columns (Figure 1.5), whereby it
serves as formwork on which the concrete can be casted. After hardening of the concrete, both
components act together in carrying the load. This construction method makes that in case of sagging
bending moments, the concrete is mainly in compression and the steel deck in tension. Both materials
have good strength properties for these types of stresses, which makes that the load can be carried
with a minimum use of materials. Since the concrete is spread on the steel deck, no external formwork
is needed, and it offers a possibility for unpropped construction. On basis of this description of
composite floor systems, it may be concluded that the construction method is mainly developed for the
best way of carrying the bending moment. It is the curvature, generated by the load, that causes the
concrete to be in compression and the steel in tension. This distribution of stresses mostly leads to
shallow designed slabs. Application of deep composite slabs can be necessary when the load or span
becomes big. A composite slab is labelled as deep when its steel deck is over 200 mm high[9].

1.2. Objective
Looking at the Eurocode 4 for verification of composite slabs on structural safety, it can be established
that the calculation procedures for the bending moment capacity and longitudinal shear capacity are
well described. The bending moment capacity of a composite slab is its resistance against the tensile
and compressive stresses caused by the bending (Figure 1.6). With the longitudinal shear capacity it
is meant to what extent the composite slab is able to prevent extensive slipping at the interface between
the concrete and the steel deck (Figure 1.7). However, for calculation of the transverse shear capacity
of a composite slab, which is the resistance against the transverse shear force (Figure 1.8), the
Eurocode 4 (composite structures) simply refers to the Eurocode 2 (concrete structures) without
providing additional information. The transverse shear capacity of a composite slab is therefore totally
assigned to its concrete part. Figure 1.9 shows the transverse shear failure of a reinforced concrete
beam, with its distinctive inclined cracks from support to loading point. For reinforced concrete
members, this type of failure is normally referred to as shear failure. But for composite slabs, a
distinction is made between longitudinal shear failure and transverse shear failure (or vertical shear
failure as used in other literature[17][18][21]).

Bending moment Stress distribution

Figure 1.6 – Bending of a composite slab[23]

Exceedance of bond strength

Figure 1.7 – Longitudinal shear failure[23][16]

4
 Figure 1.8 – Transverse shear force[23]

Figure 1.9 – Transverse shear failure of a reinforced concrete beam[33]

With the Eurocode 4 referring to Eurocode 2, it makes that the transverse shear capacity of a composite
slab is based on empirical formulas that were derived for reinforced concrete members. The
Eurocode 4 lacks in information about the influence of the steel deck on the total transverse shear
resistance and therefore it is supposed that this causes the current calculation method to be
unnecessarily conservative. So, there are formulas to predict the transverse shear capacity of the
concrete section (Eurocode 2); there are formulas to predict the transverse shear capacity of the steel
deck (Eurocode 3), but there is no good prediction method for the combined resistance (Eurocode 4).
In this thesis, it is tried to review the current standard and find a more realistic calculation method for
the transverse shear capacity of deep composite slabs.

Figure 1.10 – Typical cross-section deep composite slab (ComFlor 210)[32]

It has been chosen to focus on deep composite slabs in this thesis. A typical cross-section of a deep
composite slab is shown in Figure 1.10. The following elements can be distinguished: a thin steel deck
at the bottom with concrete casted on top (concrete section); additional rebars in the concrete ribs and
a reinforcement mesh in the top layer. Due to the shape of the steel deck, the deep composite slab
consists out of concrete ribs which are positioned at a relatively large mutual distance (600 mm for
ComFlor 210). As will become clear in the next chapter, these concrete ribs provide the transverse
shear resistance of the composite slab according to the Eurocode 4. Therefore, it may be expected that
deep composite slabs are generally more vulnerable to transverse shear compared to shallow
composite slabs, in which the concrete ribs are spaced at a much smaller mutual distance. Furthermore,
it can be seen in Figure 1.10 that the steel deck contains large webs, which are oriented nearly vertically.

5
These webs provide the transverse shear resistance of the steel deck according to the Eurocode 3, but
this resistance is totally neglected in calculation of the total transverse shear capacity of the composite
slab. Therefore, it is expected that considerably more transverse shear capacity is present than
currently predicted. It should be noticed that due to the high web slenderness of the steel deck, it is
expected to be more susceptible to local buckling, which reduces its capacity.

1.3. Research questions

The objective of this thesis (as described in the former section) is translated into short research
questions, which will be tried to be answered at the end of this thesis. The sub-questions are meant to
help answering the main question.

Main question
▪ How to calculate the transverse shear capacity of deep composite slabs?

Sub-questions
▪ How big is the transverse shear capacity of the concrete section and how does this relate to the
current Eurocode 2: Concrete structures (EN 1992)?
▪ What is the influence of the steel deck on the transverse shear capacity and how does this relate to
the current Eurocode 4: Composite structures (EN 1994)?

In order to give answers to these questions, the thesis will be started with a review of current literature.
More in-depth knowledge of composite slabs will be provided. After the literature review, the results of
the Finite Element Analysis will be discussed. These results will be used to answer the research
questions and make a conclusion. The structure of the thesis is illustrated in Figure 1.11.

Literature review

Finite Element Analysis

Discussion

Conclusion

Figure 1.11 – Structure of the thesis

6
2. LITERATURE REVIEW
This chapter presents a review of literature which will help to get more insight into the subject. It is
started with some general aspects about (deep) composite slabs and their structural behaviour. After
distinguishing the different failure modes of composite slabs, this chapter will mainly focus on the
transverse shear capacity. Calculation procedures according to standards and literature will be
discussed at the end of this chapter.

2.1. Composite slabs in general

To give this thesis a proper basis, some general aspects of composite slabs are being discussed in this
section. It is explained where it is made of, where it originates from and which variants are used in
current building techniques. Furthermore, the production process is explained, different types of
applications are provided and advantages/disadvantages are mentioned. At last, information is given
on the fire safety of composite slabs and two variants of deep composite slabs are shown that are
mainly used in the Netherlands[4] and the United Kingdom[23].

2.1.1. Materials
A composite slab is a structural element that is composed out of different materials. Steel-concrete
composite slabs are discussed in this thesis. As clarified by the name: it is a combination of (cold-
formed) steel and concrete.

Cold-formed steel
The profiled metal deck at the bottom of the slab is made of thin-walled cold-formed steel. This deck
serves as both formwork and reinforcement to the concrete. The steel deck is characterized by its
average yield strength and is the most effective in tension. This especially holds for thin-walled steel,
where compressive stresses can easily lead to local buckling. Commonly used steel grades are S280
and S350[4]. Higher grades rarely give significant advantages due to stiffness limitations[10]. The steel
deck is usually fabricated from hot-dipped galvanized plates with protective zinc layers (approximately
0.02 mm on each surface)[6]. This is normally sufficient to protect the steel from hazards when applied
as internal floors in non-aggressive environments[6].

Concrete
Concrete is a non-homogeneous material that in its most standard form is composed out of aggregates,
cement and water. The combination of cement and water glues the aggregates together due to a
hydration process. Aggregates are ranged from fine to coarse in the concrete to prevent the inclusion
of big areas of cement only. Concrete is characterized by its compressive strength, because this is its
most outstanding property. The tensile strength of concrete is only about 10% of its compressive
strength, so for most of its applications reinforcement is needed to resist the tensile stresses. In
composite slabs, the concrete is applied as topping on the profiled steel deck. This steel deck acts as
reinforcement to the concrete. Concrete can have multiple additives/adjustments to improve its
characteristics. In the United Kingdom, a frequently applied variant in composite slabs is lightweight
concrete to reduce the self-weight of the slab[9].

7
2.1.2. Origin
As mentioned in literature[5], Loucks and Gillet (1926) firstly patented a structural solution in which a
steel deck was used in combination with concrete. The resistance of this “composite” slab was only
obtained by the load-carrying capacity of the steel deck. The function of the concrete was mainly to
protect for possible fire and to level the decking for functionality. Some decades later, Granco Steel
Products Company (1950) marketed the first steel deck for composite slabs, called Cofar. This deck
was provided with transverse wires welded to the top of profile to ensure composite behaviour. Freiberg
(1954) investigated the Cofar’s resistance and published an important article on the design of composite
slabs. Later in 1967, the so-called Hibond was introduced. This contained a new trapezoidal profile with
embossments and re-entrant parts, and was the precursor of currently used steel decks in composite
slabs.

2.1.3. Production process

The production of a composite slab starts with the manufacturing of the profiled steel deck. This deck
is made in a process of cold forming. The process starts with the production of raw steel in a Basic
Oxygen Furnace (BOF) or an Electric Arc Furnace (EAF). The steel is poured into a slab and then
reduced into thinner strips of steel (so-called “hot band”)[7]. In the finishing process, the hot band is
further reduced into thin “cold rolled steel” [7]. Protective zinc layers are added through a galvanization
process and the sheet is coiled. This coil with steel is the start-product for the second process to
produce the desired cold-formed section. This process is shown in Figure 2.1. First, the coil is unwound,
after which the plate goes through a flattener to remove possible imperfections. These imperfections
can be caused by transportation of the coil for instance. After the plate has been flattened,
embossments are pushed into the plate (if demanded). The steel plate is now put into the desired shape
by means of roll forming: the plate is drawn through a series of rollers, which bend the plate into the
desired profile. No heat is required in this process (so the name cold forming) and the shape can
therefore only be maintained by plastic deformation of the steel. In the end, the profile is cut into the
desired length and collected for transportation. The profile of the steel deck is generally such that it is
completely nestable to make one stack. The steel decks are therefore effectively transported to the
building site, as it takes little space to transport a lot of decks.

Figure 2.1 – Production process of cold-formed steel[16]

Once arrived at the building site, the steel decks are put in position by the construction workers by
hand. This is possible due to the low self-weight. The last step is to lay down additional reinforcement
on the steel deck, after which the concrete is casted on top. The steel deck has to carry the wet concrete
in the beginning, but the composite slab is obtained after hardening of the concrete.

Effects of cold forming

Since the plate is bent at room temperature, plastic strains will develop in the corners of the profile.
Cold forming therefore involves strain hardening effects which locally influence the yield stress, the
ultimate strength and the ductility[8]. The degree of influence is dependent on the radius of curvature,

8
the thickness of the plate, the type of steel and the forming process[8]. Figure 2.2 shows an example
of the influence on the yield stress. The yield stress has become substantially higher near the corners
due to strain hardening. The average yield stress of the section has also become higher.

Figure 2.2 – Effect of cold forming on the yield stress[8]

2.1.4. Types
Composite slabs can roughly be divided into two categories based on the profile of the decking. This is
shown in Figure 2.3. A distinction is made between the re-entrant profile (left) and the trapezoidal profile
(right). In the Netherlands, the steel deck usually has a height in the range from 46 to 225 mm, with a
sheet thickness varying from 0.9 to 1.25 mm[4]. For shallow steel decks (height < 200 mm), typical total
slab thicknesses vary in the range from 100 mm to 250 mm[9]. For deep steel decks (height > 200 mm),
this is in the range from 280 mm to 320 mm[9].

Figure 2.3 – Two types of profiled steel decks in composite slabs[2]

The choice of the profile is strongly dependent on the span and construction method. Short span
composite slabs are generally constructed in an unpropped way, meaning that the steel deck has to
carry the weight of the wet concrete and other construction loads. This phase controls the design of the
steel deck. Due to the short span, the stresses in the composite slab after hardening of the concrete
are generally low. For such slabs, trapezoidal profiled steel decks are most often used with limited
longitudinal shear resistance[6]. This is because they have the lowest self-weight per square metre of
floor area[6]. If the span becomes big (e.g. > 6m), props can be necessary to support the steel deck.
9
After hardening of the concrete, the composite slab is more highly stressed due to the longer span. This
final stage may become governing in design, whereby a good longitudinal shear resistance will be
required between the concrete and steel for transfer of the stresses. Re-entrant profiles are often used
leading to greater self-weights per square metre of floor area[6].

The profiled steel deck at the bottom of the composite slab can be labelled as very thin. Local buckling
can therefore become a critical aspect. During the construction phase, the wet concrete doesn’t fulfil
any role in carrying the load. This makes that the steel deck is not only loaded in tension, but also in
compression. Local buckling due to these compressive stresses must be resisted, which makes that
stiffeners are applied in the profile as shown in Figure 2.4. In the upper flange, a re-entrant form has
been applied which improves the resistance for local buckling. The same holds for the embossments
in the webs, which also play an important role in improvement of the longitudinal shear resistance of
the composite slab.

Figure 2.4 - Typical 80 mm trapezoidal steel deck with stiffeners[2]

For composite slabs it holds that the slip at the interface must be prevented partly or completely, as
well as the vertical separation between the steel and concrete. The importance of this has been
explained in section 1.1. In composite beams, this is most often assured by making use of headed
studs. In composite slabs, there are more ways to ensure the composite behaviour. These are
illustrated in Figure 2.5 (provided in EN 1994-1-1[3]).

Figure 2.5 – Options to ensure bonding between the steel deck and the concrete [3]

10
2.1.5. Applications
Composite slabs are frequently applied as floors in steel-framed buildings. These buildings often have
one of the following applications[9]:
▪ Office
▪ Industrial building
▪ Warehouse
▪ Leisure building
▪ Stadium
▪ Hospital
▪ School
▪ Cinema
▪ Housing
Furthermore, composite slabs offer a good solution in refurbishments projects, due to the reduced
weight compared to regular concrete slabs. The steel deck can easily be brought inside the building
and can get modified into the desired shape.

2.1.6. Advantages
Composite slabs contain a lot of advantages which are listed below[9].
▪ Speed of construction
Bundles of steel decks can be lifted to desired position by one crane movement which makes the crane
time minimal. The individual sheets can be installed by hand due to the low self-weight. The use of the
steel deck as formwork and working platform speeds up the construction process for the next phases.
For most of the cases minimal additional reinforcement is required and large areas of flooring can be
poured with concrete quickly.
▪ Safe method
The steel decks provide a safe platform for the construction workers to perform their activities. It also
offers a protection for falling objects.
▪ Better use of material properties
In ordinary concrete slabs, the concrete outside the compressive zone has low contribution to the
strength of the slab due to the low tensile resistance of the concrete. In composite slabs, due to the
shape of the steel deck, the concrete outside the compressive zone is reduced which makes it more
effective in material use.
▪ Lightweight and shallower construction
Due to the effectiveness of composite slabs, the weight and size of the primary structure can be reduced
significantly. Consequently, the costs and size of the foundation can also be reduced.
▪ Structural stability
The composite slab can act as a lateral restraint for the supporting beams, when it is sufficiently
connected to these members.
▪ Sustainable
Due to savings in transport and trying to minimize the use of concrete, composite slabs are generally
more sustainable than many other solutions. Furthermore, the steel in composite slabs can be recycled
repeatedly without reducing its inherent properties.
▪ Easy installation of services
Pipes and cables for services like electricity, telephone and other network cabling can easily be hung
to the bottom of the composite slab. This is done by making use of special equipment that can be
attached to the steel deck due to its particular shape.

11
2.1.7. Disadvantages
Composite slabs also contain some disadvantages as listed below.
▪ Less effective for hogging bending moments
Composite slabs are less effective for hogging bending moments, since in this case the concrete is in
tension and the steel deck is in compression. Necessarily, reinforcement has to be applied in the top
layer of the concrete section and local buckling of the thin steel deck should be avoided.
▪ One-way span
Due to the shape of the composite slab with ribs in one direction, the load-carrying capacity in
transverse direction is much lower.

2.1.8. Fire safety

Fire safety generally is an important aspect in the design of structural elements. The influence of fire
safety on the design process is mainly dependent on the application of the building: a fire resistance of
30, 60, 90 or 120 minutes can be required. When composite slabs are applied as floors in buildings,
special attention should be paid to guarantee the required fire resistance. This is because of the steel
deck at the bottom of the composite slab, which can easily be heated in case of fire. Steel loses much
of its strength and stiffness at strongly elevated temperatures, so this becomes very critical when there
is nothing present in the concrete to carry the tensile stresses. Therefore, additional reinforcement is
often needed and rebars are positioned in concrete ribs (called rib reinforcement in this thesis). These
rebars can take over the tensile stresses in case of fire, whereby they are protected by the surrounding
concrete.

2.1.9. Deep composite slabs: ComFlor 210 and ComFlor 225

In the Netherlands and United Kingdom (UK), composite floor systems are generally known by the
tradename “ComFlor®”. The main supplier of these floor systems in the Netherlands is Dutch
Engineering[4]. They offer composite floors in the range from ComFlor 46 up to ComFlor 225, in which
the number indicates the height of the steel deck. As shown earlier in section 2.1.4, the steel deck can
have a trapezoidal or re-entrant profile. ComFlor 210 and 225 are labelled as deep composite
floors/slabs and both these types contain a trapezoidal profiled steel deck (see Figure 2.6). The steel
decks are made of grade S280GD (minimum yield strength of 280 MPa) and are provided with
protective zinc layers of a total mass of 275 g/m2 (approx. 0.02 mm on each surface)[4]. The total sheet
thickness of ComFlor 210 can be 1.0 mm or 1.25 mm, while ComFlor 225 is only offered with a 1.25
mm sheet thickness[4]. It should be noticed that there are some inconsistencies in the use of the name
“ComFlor®” between the Netherlands and the UK. For example, ComFlor 75 in the Netherlands refers
to ComFlor 60 in the UK. Furthermore, differences in applied steel grades can also be found, e.g. deep
composite slabs in the UK are normally offered with a steel grade of S350GD[23].

For composite slabs with a shallow steel profile (e.g. ComFlor 46), the steel deck is normally put on top
of the supporting beams. It is connected to these beams with headed studs, after which the composite
slab and supporting beams also act compositely. The composite slab is continuously supported over
the beams (shown in Figure 2.3). In contrast, ComFlor 210 and 225 are usually supported within the
beam depth, e.g. on the bottom flanges of the supporting beams as shown in Figure 2.7. This provides
a very shallow floor zone. The steel deck is simply supported, but the concrete topping is still continuous
over the supporting beams. As also shown in Figure 2.7, the steel decks are provided with end
diaphragms. These diaphragms are of importance for the structural integrity of the deck and to ensure
that the concrete doesn’t flow away when it is poured. Typical spans for deep composite slabs extending
up to 5.6 metres and 9 metres for unpropped and propped construction respectively [4].

12
 Figure 2.6 – Steel profile for deep decks: ComFlor 210 (left) and ComFlor 225 (right)[4]

Figure 2.7 – Composite slabs supported on lower flange: ComFlor 210 (left) and ComFlor 225 (right)[23]

Minimal mesh reinforcement is always applied in the top layer of the slab to reduce the crack width due
to shrinkage[4]. This reinforcement also distributes the effects of localised loads and provides a certain
hogging moment capacity. Mainly based on this later two aspects, it can be necessary to increase the
applied reinforcement in the top layer by using additional reinforcement bars. This especially holds for
the regions at the supports in case of hogging moment conditions. Furthermore, for deep composite
slabs, the resistance is always increased by using additional rib reinforcement[4]. This rib reinforcement
is positioned at the bottom of the slab (with some concrete cover) and is necessary in case of fire.
However, the benefit of this reinforcement is considered for both normal (room temperature) and fire
conditions. Figure 2.8 shows the way this rib reinforcement is put in position. Furthermore, from this
figure it may also be observed that the steel deck has an overlap at the bottom of the ribs, caused by
the assemblance of the separate parts of the steel deck. The sheet thickness is therefore two times
bigger at the bottom of the ribs. Connectors are applied with some mutual distance over the full length
of this bottom flange. These connectors consist out of “shear clips” and self-drilling screws (which go
through these clips and the separate plates of the bottom flange) [4].

Shear clips

Rib reinforcement

Double thickness
bottom part steel deck

Figure 2.8 – Positioning of rib reinforcement in ComFlor 210[4]

13
It needs to be noticed that the considered deep composite slabs of this thesis aren’t officially covered
by the Eurocode 4. This is because they don’t meet the requirement on narrowly spaced webs as
prescribed in clause 9.1.1 of EN 1994-1-1[3]. This requirement is defined by an upper limit of 0.6 for
the ratio br/bs. Figure 2.9 shows the definitions of these parameters br and bs. For ComFlor 210, the
ratio is equal to 425/600 = 0.71. For ComFlor 225, the ratio is equal to 400/600 = 0.67. As may be
established, both profiles don’t meet the requirement since these values are bigger than 0.6.

Figure 2.9 – Definition of different parameters for trapezoidal (left) and re-entrant (right) steel decks[3]

2.2. Composite behaviour

The structural behaviour of a composite slab depends on the properties of the separate materials and
the connection between them. Figure 2.10 shows the distribution of forces and stresses in a deep
composite slab due to a distributed load. At midspan, the stress distribution is such that the concrete
top layer is in compression, while the steel is in tension. Concrete in tension is neglected due to its
weak tensile properties. The transverse shear force is zero at midspan where the bending moment is
at its maximum. At the support this is visa-versa, making that the shear stresses are significantly higher
at this location. This holds for both the longitudinal and transversal shear stresses. Longitudinal shear
stresses are generated between the steel deck and the concrete section (bonding). Slip is likely to occur
due to the non-rigidity of this connection. The bond capacity is improved by the embossments in the
steel deck due to mechanical interlock. The shape of the steel deck is also of importance: a re-entrant
profile has better connection properties than a trapezoidal profile, because it offers more resistance
against vertical separation between the steel and concrete (opening of the interface).

Figure 2.10 – Stresses in a deep composite slab[23]

As also described by Schuurman[12], the three main important aspects of the interface are its strength,
stiffness and ductility. This distinction in properties is quite straight-forward, as these aspects are
generally important for structural elements. The meaning of these properties can be related to the
stress-slip diagram of Figure 2.11. The strength of the interface is the maximum shear stress τu that
14
can be reached, the ductility is the maximum slip δ//,u that can be reached and the stiffness Eint is the
tangent to the diagram. Two ultimate cases of interaction can be distinguished: full and no interaction.
For no interaction, the longitudinal shear stress is zero for every value of slip. For full interaction, there
is no slip for every value of longitudinal shear stress (below τu). Cases in between can be referred to as
partial interaction. The influence of the mentioned properties on the composite behaviour will now be
discussed.

Figure 2.11 – Qualitative shear-slip curve[12]

Strength of the interface

Consider the simplistic composite beam model as shown in the introduction of this thesis in Figure 1.1.
A part of this composite beam is also shown in Figure 2.12. For each separate beam (h x b) at the top
and bottom, it holds that its stress distribution can be replaced by a normal force N and a bending
moment Mp*. The total bending moment capacity of the composite beam MR,comp equals the summation
of these bending moments and the product of the normal force N times the internal lever arm z (equation
2.1). The longitudinal shear stresses at the interface will result in a shear force T (equation 2.2). This
longitudinal shear force T (acting over a length x from the left support) is in equilibrium with normal force
N at the considered cross-section. When there is no interaction between the beams, T equals zero,
meaning that the normal force N also has to be zero. Assuming elasto-plastic behaviour, both separate
beams can now be loaded to their full plastic capacities Mp as shown in Figure 2.12 (above the y-axis).
When there is interaction between the beams, T doesn’t equal to zero, resulting in a normal force N in
each beam. Due to the presence of the normal forces, the separate beams can’t be loaded to their full
plastic capacities, but only to a reduced bending moment capacity Mp* (equation 2.3). When the normal
force N equals its plastic capacity Np, the ultimate bending moment capacity MRu,comp of the composite
beam is reached.

Figure 2.12 – Development of the ultimate moment capacity due to composite behaviour (N = T)[12]
15
 𝑀𝑅,𝑐𝑜𝑚𝑝 = 2 ∗ 𝑀𝑃∗ + 𝑁 ∗ 𝑧 2.1
𝑥
𝑇 = 𝑏 ∫ 𝜏 𝑑𝑥 2.2
0
2
𝑁 𝑏∙ℎ 2
𝑀𝑃∗ = 𝑀𝑝 ∗ (1 − (𝑁 ) ), with 𝑀𝑝 = 4
∙ 𝑓𝑦 𝑎𝑛𝑑 𝑁𝑝 = 𝑏 ∙ ℎ ∙ 𝑓𝑦 2.3
𝑝

What may be concluded from the former is that the strength of the interface is of importance for
development of the ultimate capacity of the composite element. As shown, the greatest benefit in
bending moment capacity MR,comp is when both separate parts can be loaded to their full plastic normal
capacities. This can only be reached when the interface is strong enough to carry the corresponding
longitudinal shear force T. This is not only dependent on the ultimate value for the longitudinal shear
stress τu, but also on the distance x from the support as the value of T is also dependent on this later
variable. The value of τu therefore determines at which distance from the support the ultimate capacity
MRu,comp of the composite beam can be reached. A higher longitudinal shear strength leads to a lower
distance from the support. This length is called LSF and can be calculated with equation 2.4.

𝑁𝑝
𝐿𝑆𝐹 = 2.4
𝜏𝑢

Composite slabs are much more complex than the fictitious composite beam as used in above’s
explanation, but still it can be established that the strength of the longitudinal shear connection is an
important factor for the capacity of the composite slab.

Stiffness of the interface

The stiffness of the connection Eint influences the bending stiffness of the composite slab and its load-
carrying capacity. This can be explained with the help of Figure 2.13, in which again the same
composite beam is shown. The situation on the left of the figure corresponds to no interaction, which
means that the stiffness of the connection equals to zero (x-axis in Figure 2.11). This results in a
relatively large slip and a discontinuity in strains/stresses at the location of the interface. The situation
on the right corresponds to full interaction as a result of an infinite stiffness (y-axis in Figure 2.11),
leading to no slip at the interface. Both normal stresses and transverse shear stresses are now
continuous over the height of the composite beam. Formulas are provided below to compare both
ultimate cases of interaction. From these formulas it may be concluded that the moment of inertia
increases with a factor of 4 (I2 = 4I1) when the stiffness of the connection is increased from zero to
infinity. As a result, the total deformation reduces proportionally (δ2 ~ EI2-1). This directly shows how the
bending stiffness of the composite beam is affected by the stiffness of the connection. Furthermore, it
can also be seen that the normal stresses reduce with a factor of 2, which means that a two times larger
force is needed to cause the same stress level.

16
 Figure 2.13 – Comparison between no, partial and full interaction in a composite beam[16]

Ductility
Ductile behaviour of the interface is an important aspect in preventing brittle failure of the whole
composite element. This has been well explained by Schuurman[12], who described three types of
longitudinal shear connections in a composite beam. Figure 2.14 shows the properties of these
connections (A, B and C) together with the boundary conditions of the composite beam. Connection A
and B both have an infinite stiffness, with B having a ductile behaviour in contrast to A. Connection C
has a ductile behaviour and a finite stiffness. The composite beam is made of a linear-elastic material
with an infinite strength.

Figure 2.14 – Three types of longitudinal shear connections: A) brittle and infinite stiffness, B) ductile
and infinite stiffness, C) ductile and finite stiffness.[12]

From classical mechanics it is known that the transverse shear force in the composite beam is at its
maximum near the supports in case of a distributed loading. With an infinite stiffness of the interface as
in connection A and B, full interaction is obtained leading to a maximum shear stress at the location of
the interface (shown in Figure 2.13). This holds for both the transversal and longitudinal shear stress.
17
When the load on the composite beam increases, the longitudinal shear stress τ increases, reaching
its maximum first at the location of the supports. In case of connection A, after reaching the strength of
the connection τu, it fails in a very brittle way. The stress release causes the whole interface to fail
progressively like a zipper, without additional increase of the load[12]. With this happening, the bending
stiffness of the composite beam reduces with a factor of 4 (from EI2 to EI1, see Figure 2.13). In other
words, the composite beam quickly changes from full interaction to no interaction. This also causes that
the normal stresses will increase with a factor of 2, which could lead to total failure of the beam when
the material contains a finite strength. Since no strength is specified in this example, additional loading
is possible and will be carried by the two separate beams.

For connection B and C, it holds that after reaching τu at the supports, the longitudinal shear stress
doesn’t drop to zero, but stays at a constant level. This introduces some slip, causing the bending
stiffness to reduce at these locations. This reduction in bending stiffness only propagates towards the
midspan when the load on the composite beam is increased. By doing so, τu will be reached at more
locations of the interface. In the end, τu is reached over the full length of the interface and the bending
stiffness of the total composite beam is also reduced with a factor of 4. The end-situation is thus the
same as for connection A. The only difference is that additional loading is needed to cause this end-
situation for connection B and C.

Introducing a finite stiffness as in connection C makes that the slip at the interface is introduced before
reaching the strength of the interface τu. This causes the bending stiffness to drop quicker for connection
C than B, resulting in a bigger deflection at the same load level. In the ultimate phase, when τu is
reached over the full length of the interface, every case (A, B and C) contains the same bending stiffness
EI1. Figure 2.15 gives an overview of the load-displacement behaviour for all three different
connections. The marked points with index 1 indicate the occasions when τu is reached at the supports,
while the points with index 2 indicate the occasions when τu is reached over the full interface.

Figure 2.15 – Load-displacement curves for three different types of longitudinal shear connections[12]

2.3. Failure mechanisms of composite slabs

Composite slabs are susceptible to three possible failure mechanisms[12]:
1. Flexural failure
2. Longitudinal shear failure
3. Transverse shear failure (or vertical shear failure)
Figure 2.16a gives indications for the locations of these failure mechanisms in case of a four-point
bending test. The index of the location corresponds to the index of the failure mechanism as given
above. Flexural failure occurs when the bending moment capacity of the composite slab is exceeded.

18
This failure mechanism is characterised by yielding of the steel deck (and rib reinforcement), with the
concrete on top in compression. This failure mechanism is most often ductile and frequently occurs in
the middle region of the span (between the localized loads caused by force W in Figure 2.16a). In order
to utilize the bending moment capacity of the slab, it is needed to provide enough bonding between the
concrete and steel at the interface. Longitudinal shear failure is therefore a secondly common noticed
failure mechanism in composite slabs. This failure mode is limited by the shear capacity of the
connection (τu, see Figure 2.11) and is characterized by excessive slip between the steel and concrete.
Diagonal cracks near the loading point are also often visible[13]. Most research in composite slabs has
been conducted to this failure mechanism in order to characterize its corresponding resistance. To date,
it is still not possible to determine it barely with design formulas. The longitudinal shear capacity is
therefore provided by the supplier and determined by testing. At last, transverse shear failure is a third
possible failure mechanism which mainly occurs closely to the supports. This mechanism is known by
its brittle character and should be avoided. The failure mode is generally characterised by opening of
the cracks in the concrete which are oriented diagonally from the support to the loading point.

Figure 2.16 – a) Four-point bending test of composite slab with locations for failure modes;
b) Failure load Vt (local load at the end of Ls) as a function of the inversed shear span (Ls-1)[13]

The governing failure mode in a composite slab is influenced by the shear span Ls. The shear span is
defined as the distance between the local load and the support, as also visible in Figure 2.16a. The
dependency of the failure mode on this shear span is shown in Figure 2.16b. In this figure, the failure
load Vt is qualitatively plotted against the inversed shear span Ls-1. Keep in mind that this inversion
causes long spans to be close to zero. The quantities are normalized by constants b, dp and Ap, which
are respectively the width of the slab, the effective depth of the slab and the cross-sectional area of the
steel deck.

What can be seen in Figure 2.16b is that the load-carrying capacity for flexural failure increases for
shorter shear spans. A short shear span makes that the bending moment in the composite slab is lower
compared to a large shear span loaded by the same force W. Therefore, for large shear spans the
ultimate bending moment capacity is reached at a lower load-level than for short shear spans. For that
reason, flexural failure mainly occurs when the shear span is large and especially for slender slabs
having a relatively low bending moment capacity.

Furthermore, it can be seen in Figure 2.16b that the load-carrying capacity for transverse shear failure
is independent of the shear span. The four-point bending test causes the shear force to be constant
within the shear span Ls. Assuming that the composite slab has an equal transverse shear resistance
over its whole length, indeed leads to the fact that the failure load for transverse shear is independent
of the shear span. Increasing or decreasing the shear span only leads to movement of the point loads,
19
but this movement doesn’t increase the shear force in the slab. It should be noticed that this is off course
a purely theoretical way of thinking, as research has shown that the transverse shear capacity of
concrete sections is dependent on the shear span [28].

Assuming a short shear span and neglecting longitudinal shear failure, the load can be increased until
the transverse shear capacity of the slab is reached. When this failure load is now kept constant, but
moved towards the midspan, makes that the transverse shear force in the slab remains the same while
the bending moment increases. At a critical value for the shear span, not the transverse shear capacity,
but the bending moment capacity will become governing and the slab fails in bending. This is also
visible in Figure 2.16b as the intersection of the dashed lines for flexural and transverse shear failure.

Due to the presence of the interface, longitudinal shear failure should be considered. The interface
should be strong enough in order to reach the ultimate bending moment capacity of the composite slab.
This has been explained in section 2.2, in which it was shown that the longitudinal shear force T needs
to be big enough to reach the ultimate moment capacity of the composite beam (as also for slabs). This
force T is dependent on the strength of the connection τu and the length x from the support (see equation
2.2). This length x equals the shear span Ls, since the longitudinal shear stresses only develop within
this part of the composite slab. Let’s assume that flexural failure is the governing mode for a given
(long) shear span and this shear span will be reduced. At some critical point, the longitudinal shear
force T becomes too big in order to maintain the ultimate bending moment capacity, because the shear
span Ls (or distance x) becomes too short. In this case it is demanded that the connection provides a
shear stress above its ultimate strength τu, but since this is not possible, longitudinal shear failure
occurs. This transition from flexural failure to longitudinal shear failure is visible in Figure 2.16b as the
intersection of the corresponding lines. The resistance of a composite slab against longitudinal shear
failure (as expressed in Figure 2.16b) can be determined by testing with the so-called m-k method. The
meanings of these symbols are also visible in the figure. It can be seen that the resistance against
longitudinal shear failure increases for shorter shear spans. This may seem contradictory, as it is
expected that longitudinal shear failure occurs at an earlier stage when the shear span (= length which
can carry the longitudinal shear stresses) is reduced. However, the load-carrying capacity should be
related to the load that shall cause flexural failure for the same shear span Ls. When this is done and
the diagram for longitudinal shear failure is assessed with respect to the diagram for flexural failure, it
can be established that the difference between the two diagram gets bigger when the shear span is
reduced. This indeed confirms the expectation that longitudinal shear failure occurs at an earlier stage
for shorter shear spans.

This thesis focusses on the transverse shear failure of deep composite slabs. What can be concluded
from the former is that transverse shear failure is the closest related to longitudinal shear failure
according to the theory. Therefore, both corresponding resistances are further explained in the following
sections.

2.4. Longitudinal shear resistance of composite slabs according to Eurocode 4

NEN 1994-1-1[3] specifies two methods for determining the resistance against longitudinal shear
failure, namely the m-k method and the partial shear method. This later method may only be applied to
composite slabs with a ductile longitudinal shear behaviour. The behaviour may be considered as
ductile if the failure load for longitudinal shear exceeds the load that causes a relative end slip of 0.1
mm by more than 10%. If this failure load is reached at a midspan deflection of more than L/50, the
failure load should be taken as the load that causes this deflection of L/50.

20
Ductile behaviour isn’t a requirement for application of the m-k method and can therefore be used for
every composite slab that is governed by EN 1994-1-1[3]. With this method, m and k are empirical
constants which have to be determined by testing as shown in Figure 2.17. The composite slabs are
brought to failure in a four-point bending test (with line loads). Before this procedure, the slab is first
cyclically loaded between 20% and 60% of the expected failure load. With this cyclic load it is intended
to represent the load over a period of time and to minimize the influence of chemical/adhesive bonding
between the steel and concrete[15]. Typically, two sets of three tests (or three sets of two tests) are
carried out, in which each set has a different shear span Ls. The shear span for each set is chosen such
that it is as long or short as possible (not less than 3ht) for every tested slab to still fail in longitudinal
shear. After gathering the test results, m and k are obtained by drawing a linear line at a specific
distance below the lowest failure load in each set (to allow for scatter in the test results). The distance
between this line and the test results is prescribed by the Eurocode 4.

Figure 2.17 – Test method to determine the values for m and k[3]

As can be seen in Figure 2.17, m relates to the slope of the line and k relates to the offset of the line at
the y-axis. They are used in equation 2.5 to determine the resistance of the composite slab against
longitudinal shear failure Vl,Rd. As explicitly mentioned in clause 9.7.3 of EN 1994-1-1[3], the nominal
cross-sectional area Ap of the steel deck is used in equation 2.5, as it is normally used in the test
evaluations to determine m and k. To be consistent, dp is therefore defined as the distance between
the top fibre of the slab and the centroidal axis of this nominal cross-section Ap of the steel deck. When
the effective cross-section Ape is used, dp should be related to the centroidal axis of this effective cross-
section. The effective cross-section Ape is obtained by reducing the nominal cross-section Ap with the
area that includes the embossments (according to Eurocode 4).

𝑏 𝑑𝑝 𝑚 𝐴𝑝
𝑉𝑙,𝑅𝑑 = 𝑉𝑡 = ( + 𝑘) 2.5
𝛾𝑣𝑠 𝑏 𝐿𝑠

With:
𝑏 width of the slab [mm];
𝑑𝑝 effective depth of the slab [mm];
𝛾𝑣𝑠 safety factor for the ultimate limit state [-];
𝐴𝑝 nominal cross-section of the steel deck [mm2];
𝐿𝑠 shear span [mm];
𝑚, 𝑘 empirical factors [MPa]
21
Because of full-scale testing, the m-k method is a time-consuming and costly process. When a new
type of steel deck is developed, the values for constants m and k in principal have to be determined for
every thickness, coating and grade of the steel deck; for every total slab depth to be used and for a
range of concrete strengths and densities[3][15]. To reduce the number of tests, some simplifications
for these requirements are given in EN 1994-1-1[3].

The partial shear method may only be applied to composite slabs with ductile longitudinal shear
behaviour, because it is based on a simple engineering model with rigid stress-blocks for the steel and
concrete. This is shown in Figure 2.18. With this method it is possible to determine the longitudinal
shear resistance τu of the connection between the concrete and steel. The longitudinal shear stress is
assumed to be equally distributed within the shear span. At least four tests have to be carried out, in
which for three tests the shear span has to be as long as possible while still providing longitudinal shear
failure. In the other test, the shear span must be as short as possible (not less than 3ht) with also
providing longitudinal shear failure. This later test is to confirm the ductility requirement for the partial
shear method. In each test, the bending moment Mtest at longitudinal shear failure should be divided by
the calculated bending moment capacity Mp,Rm based on a full shear connection (stress distribution on
the top-right of the diagram in Figure 2.18). With this ratio, a horizontal line from A to B can be drawn,
after which the degree of shear connection ηtest can be determined at point C. Subsequently, the shear
resistance τu for every test can be determined with the help of equation 2.6.

Figure 2.18 – Partial shear method[3]

𝜂 𝑁𝑐𝑓
𝜏𝑢 = 2.6
𝑏 (𝐿𝑠 + 𝐿0 )

With:
𝜂 degree of shear connection [-];
𝑁𝑐𝑓 compressive force in concrete for full shear connection [N];
𝑏 width of the slab [mm];
𝐿𝑠 shear span [mm];
𝐿𝑜 length of the overhang [mm].

As discussed in section 2.1.9, EN 1994-1-1[3] doesn’t cover ComFlor 210 (br/bs > 0.6). Despite this,
Van Erp[16] recently determined the longitudinal shear resistance of this composite slab with the
methods provided in EN 1994-1-1[3]. In his thesis, he tested multiple ComFlor 210 type of composite
slabs with a total height of 280 mm. The slabs contained Ø12 mm longitudinal rebars as rib

22
reinforcement and a Ø8-150 mm mesh in the top layer; both made of steel with a minimal strength of
500 MPa. The thickness of the steel deck was 1 mm and had a minimal strength of 350 MPa. The
concrete had an expected nominal strength of 20 MPa. He found values of m = 154.3 MPa and k =
0.0369 MPa. In the determination of these constants he used an (effective) cross-section Ap of 1140.5
mm²/m for the steel deck. Furthermore, he found an average value of τu = 0.103 MPa, with
corresponding characteristic value τu,Rk = 0.0716 MPa and design value τu,Rd = 0.0573 MPa. In the
determination of these values, he assumed an effectiveness of 12.5% for the parts of the embossments.
This means that for calculation, the effective thickness of the steel deck at these locations is assumed
to be 0.125 mm instead of 1 mm.

2.5. Transverse shear resistance of composite slabs according to Eurocode 4

(referring to Eurocode 2: concrete structures)
Transverse shear failure is most likely to occur for highly concentrated loads near the supports and in
general for composite slabs with a low L/ht ratio (i.e. span/depth)[15]. The resistance of a composite
slab against transverse shear failure (Vt,Rd) is prescribed in clause 9.7.5 of EN 1994-1-1[3] (Eurocode
4), which simply refers to clause 6.2.2 of EN 1992-1-1[24] (Eurocode 2). This later is about the
transverse shear capacity of concrete members without shear reinforcement (stirrups). From this, it
may be established that the transverse shear resistance of a composite slab according to the
Eurocode 4 is completely based on its concrete part (i.e. concrete ribs of the composite slab). The
standard indirectly suggests that the steel deck doesn’t give any additional resistance to the composite
slab in case of transverse shear failure. Equation 2.7 can be used to calculate the transverse shear
resistance Vt,Rd of a composite slab per rib in accordance with Eurocode 4 and 2 (see also Figure 2.20).
The effective width bw has to be taken as the smallest width of the concrete rib in the tensile area
according to EN 1992-1-1[24]. In the old Dutch code NVN-ENV 1994-1-1[26], the effective width is
specified differently as b0 (see Figure 2.19). In case of a trapezoidal profiled deck, b0 equals the mean
width of the concrete rib; for a re-entrant deck, b0 is equal to the minimum width at the top of the trough
of the steel deck. The use of this parameter b0 instead of bw can also be found in other
literature[15][17][18].

Trapezoidal steel deck

Re-entrant steel deck

Figure 2.19 – Definition parameter b0 in composite slabs

Furthermore, it may be concluded that there are some ambiguities in this calculation method, caused
by the fact that Eurocode 4 simply refers to Eurocode 2 without providing additional information. For
example, it is not perfectly clear whether it is allowed to take into account the area of the steel deck in
calculation of the longitudinal reinforcement ratio ρl, which positively contributes to the transverse shear

23
resistance according to the formula. Since the steel deck acts as tensile reinforcement in almost the
same way as ordinary rebars (ULS: plastic stress distribution), the author tends to assert that this is
allowed. On the contrast, since equation 2.7 is empirically determined for reinforced concrete members,
it is simply not known how the area of the steel deck can be implemented into this formula for calculation
of the composite slab’s resistance. Furthermore, no guidance is given on how to take the effective depth
of the composite slab in case of presence of rib reinforcement. To comply with both norms EN 1992-1-
1[24] and EN 1994-1-1[3], the following is recommended by the author and applied in this thesis:
▪ No rib reinforcement → Effective depth d is taken as the distance between the top fibre of the
slab and the centroidal axis of the effective steel deck (i.e. dp in EN 1994-1-1[3]); reinforcement
ratio ρl is set to zero.
▪ Including rib reinforcement → Effective depth d is taken as the distance between the top fibre
of the slab and the centroidal axis of the longitudinal rebar; reinforcement ratio ρl is based on
the tensile area of the rebar.

1
𝑉𝑡,𝑅𝑑 = 𝑉𝑅𝑑,𝑐 = 𝐶𝑅𝑑,𝑐 𝑘 (100𝜌𝑙 𝑓𝑐𝑘 )3 𝑏𝑤 𝑑 ≥ 0.035 𝑘 3/2 𝑓𝑐𝑘 1/2 𝑏𝑤 𝑑 2.7

With:
𝑉𝑡,𝑅𝑑 transverse shear resistance of the composite slab per rib [N];
𝑉𝑅𝑑,𝑐 transverse shear resistance of the concrete part per rib [N];
𝐶𝑅𝑑,𝑐 empirical coefficient [-];
200
𝑘 =1+ √ ≤ 2.0, with d in mm [-];
𝑑
𝐴𝑠𝑙
𝜌𝑙 = ≤ 0.02 [-];
𝑏𝑤 𝑑
𝐴𝑠𝑙 total area of the tensile reinforcement, that is minimally anchored with a length of
(𝑙𝑏𝑑 + 𝑑) beyond the cross-section considered [mm2];
𝑙𝑏𝑑 design anchorage length [mm];
𝑓𝑐𝑘 characteristic concrete compressive strength [MPa];
𝑏𝑤 smallest width of the cross-section in the tensile area [mm];
𝑑 effective depth composite slab [mm].

The general mission of the Eurocode is to ensure structural safety with efficiently making use of
materials. Prediction of the strength should therefore be as close as possible to reality, but
literature[17][18][21] suggests that EN 1994-1-1[3] gives too conservative results when it comes to the
transverse shear resistance of a composite slab. The most obvious clarification is that the influence of
the steel deck is not considered.
Transverse shear resistance of
the composite slab per rib (Vt,Rd)

VRd,c

Figure 2.20 – Transverse shear resistance of a composite slab per rib (Vt,Rd) according to
Eurocode 4: it is totally based on the capacity of the concrete section (VRd,c)

24
2.6. Transverse shear resistance of steel decks according to Eurocode 3
During construction of a composite slab, the steel deck has to carry the wet concrete and other
construction loads in an often unpropped way as explained in section 2.1.3. As a result, transverse
shear stresses will develop in the webs of the steel deck. These stresses must be resisted to guarantee
safety. The transverse shear resistance Vb,Rd of a steel deck is prescribed in clause 6.1.5 of EN 1993-
1-3[25] and can be calculated with equation 2.8. It has been chosen to express the transverse shear
resistance of the steel deck in Newtons for one “rib” (see Figure 2.21). This has been done to comply
with the former section about the transverse shear capacity of a composite slab and to be consistent
throughout the whole thesis. One rib contains two webs, which clarifies the factor of 2 in equation 2.8.

ℎ𝑤
𝑠𝑖𝑛(𝜑) 𝑡 𝑓𝑏𝑣 2.8
𝑉𝑏,𝑅𝑑 = 2∙
𝛾𝑀0

With:
𝑉𝑏,𝑅𝑑 transverse shear capacity of a steel deck (expressed per “rib”) [N];
𝑓𝑏𝑣 shear strength considering buckling [MPa];
ℎ𝑤 height of the web between the centrelines of the flanges, measured perpendicularly to
these flanges (Figure 2.24) [mm];
𝑡 thickness of the web (excluding coatings) [mm];
𝜑 angle of the web relative to the horizontal axis [rad];
𝛾𝑀0 material factor for the ULS [-].

Vb,Rd

Figure 2.21 – Transverse shear resistance steel deck Vb,Rd per “rib” (bounded by red dashed lines)

Since the steel deck might be sensitive to local buckling due to its high plate slenderness, shear
buckling is considered by introducing a shear buckling strength fbv in equation 2.8. Normally the shear
strength of steel is taken as fy/√3, but depending on the web slenderness the value for fbv may be below
this shear strength. More information is provided in Appendix A.

2.7. Transverse shear resistance of composite slabs according to literature

In section 2.5 it has been explained that the transverse shear resistance of a composite slab is only
depended on the resistance of its concrete ribs according to Eurocode 4. The contribution of the steel
deck to this resistance is totally neglected. Someone may use the tensile area of the steel deck in
calculation of the longitudinal reinforcement ratio ρl, but it is doubted if this is allowed since the Eurocode
4 doesn’t specifically mention this. Suppose it is allowed, then it would still be remarkable that the
contribution of the steel deck to the transverse shear resistance is based on its area rather than its
shape[17]. The webs of the steel deck ensure that this profile has a significant transverse shear
resistance on its own as explained in section 2.6 (according to Eurocode 3). Therefore, it may be
contradictory to neglect the resistance of these webs in calculation of the transverse shear resistance
of the composite slab.
25
Stark[18] tried to validate a new design model for the transverse shear resistance of composite slabs.
He distinguished three potential contributions to this resistance as shown in Figure 2.22. These
contributions can be described as follows:
I. Transverse shear resistance of the concrete ribs;
II. Transverse shear resistance of the steel deck;
III. Additional transverse shear resistance from the concrete above the flanges of the steel deck.

Figure 2.22 – Potential contributions to the transverse shear resistance of a composite slab[18]

The first contribution to the transverse shear resistance of a composite slab has been well explained in
section 2.5. According to Stark[18], the effective depth d has to be taken as the distance between the
top fibre of the slab and the centroidal axis of the rib reinforcement. Furthermore, the mean width of the
concrete rib b0 has to be used in calculation as the effective width of the concrete rib (instead of bw).
The rebars are assumed to be fully anchored, while the contribution of the steel deck to the longitudinal
reinforcement ratio is conservatively neglected.

The contribution of the steel deck (II) is based on the assumption that each separate web provides a
transverse shear resistance equal to the vertical component of the shear yield capacity Vlu of that web
(see Figure 2.23). This shear yield capacity Vlu can be calculated with equation 2.9, by which the
transverse shear resistance VRd,p of the steel deck can be calculated with equation 2.10. It can be
established that Stark’s model is in great agreement with Eurocode 3 (see section 2.6), although there
are some differences:
▪ no reduced shear strength is used to consider shear buckling;
▪ the height of the web is taken as the part between the radii to the flanges (h l, see Figure 2.24).

𝑓𝑦
𝑉𝑙𝑢 = ℎ𝑙 ∙ 𝑡 ∙ 2.9
√3
cos(𝛼 ) ∙ 𝑉𝑙,𝑢
𝑉𝑅𝑑,𝑝 = 2 ∙ 2.10
𝛾𝑀0

With:
𝑉𝑅𝑑,𝑝 contribution of the steel deck to the transverse shear resistance of a composite slab for
one rib (according to Stark[18])[N];
𝑉𝑙𝑢 shear yield capacity of a steel deck’s web [N];
ℎ𝑙 height of the web, specified as the distance between the radii to the flanges (Figure 2.24)
[mm];
𝑡 thickness of the web (excluding coatings) [mm];
𝑓𝑦 yield strength [MPa];
𝛼 angle of the web with respect to the vertical axis [rad].

26
 Figure 2.23 – Transverse shear resistance of the web Vlu (example for ComFlor 60(UK))[18]

Figure 2.24 – Definition of hw and hl

The contribution of the concrete part above the flanges (III) is based on the same principles as for the
contribution of the concrete ribs (I). Prerequisite for this contribution is that a double reinforcement mesh
is present in the upper part of the composite slab as also can be seen in Figure 2.22[18]. The transverse
shear resistance VRd,c2 of this concrete part can be calculated with equation 2.11. Visualization of the
model is given in Figure 2.25.

1 3 1
𝑉𝑅𝑑,𝑐2 = 𝐶𝑅𝑑,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑘 )3 𝑏𝑤2 𝑑2 ≥ 0.035 𝑘 2 𝑓𝑐𝑘 2 𝑏𝑤2 𝑑2 2.11

With:
𝑉𝑅𝑑,𝑐2 contribution of the concrete part above the flanges according to Stark [18] [N];
𝐶𝑅𝑑,𝑐 empirical coefficient [-];
200
𝑘 = 1+ √𝑑 ≤ 2.0 with 𝑑2 in mm [-];
2
𝐴𝑠𝑙
𝜌𝑙 = ≤ 0.02 [-];
𝑏𝑤2𝑑2
𝐴𝑠𝑙 cross-sectional area of the lower reinforcement mesh in the concrete part above the
flange [mm2];
𝑓𝑐𝑘 characteristic concrete compressive strength [MPa];
𝑏𝑤2 = 𝑏𝑠 − 𝑏0 [mm];
𝑏𝑠 centre-to-centre distance concrete ribs [mm];
𝑏0 mean width of the concrete rib [mm];
𝑑2 effective depth of the concrete part above the flange [mm];

27
 Figure 2.25 – Area of the concrete part above the flange that contributes to the transverse shear
resistance of the composite slab[18]

According to the model of Stark[18], the total transverse shear resistance of the composite slab is
obtained by simply adding the contributions of the three different parts. In order to validate the model,
series of experiments were conducted by Popo-Ola[19] on ComFlor 60(UK) type of composite slabs.
The tested specimens were 600 mm wide and contained two complete concrete ribs. The steel deck
was made of S350. Different parameters were varied in the tests: the concrete strength, the total slab
height, the reinforcement mesh, the sheet thickness and the rib reinforcement. Based on comparison
of test results with calculated values, it has been concluded by Stark[18] that the model provides safe
values for all tests: the maximum ratio of calculated resistance to corresponding test result is equal to
0.86. For calculation of the transverse shear resistance, the contribution of the concrete ribs (I) has only
been considered when rib reinforcement was present in the corresponding test. The same holds for the
contribution of concrete parts above the flanges (III) with respect to the double reinforcement mesh in
the top layer of the composite slab. This is based on the observation in tests without rib reinforcement,
that after crack initiation in the concrete ribs, the load was transferred to the steel deck which carried
almost all of the load[18]. This led to a considerable uplift of the concrete part with respect the steel
deck (vertical separation)[18]. Most important conclusion of the experiments and corresponding report
is that the steel deck considerably contributes to the transverse shear resistance of the composite slab.
For almost all tests this contribution could be calculated with equation 2.10 according to Stark[18].

Since the use of a double reinforcement mesh is not very common in practice, the contribution of the
concrete part above the flange (III) may be neglected in the calculation of the total transverse shear
resistance Vt,Rd. This leads to the proposed calculation method as shown in equation 2.12, in which the
transverse shear resistance of the composite slab equals the summation of the capacity of the concrete
rib (VRd,c) and the capacity of the steel deck (VRd,p). This is illustrated in Figure 2.26.

𝑓𝑦
1 3 1
𝑉𝑡,𝑅𝑑 = 𝑉𝑅𝑑,𝑐 + 𝑉𝑅𝑑,𝑝 = max (𝐶𝑅𝑑,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑘 )3 𝑏0 𝑑 ; 0.035 𝑘 2 𝑓𝑐𝑘 2 𝑏0 𝑑) + 2 ∙ cos(𝛼) ∙ ℎ𝑙 ∙ 𝑡 ∙ √3 2.12
𝛾𝑀0

Transverse shear resistance of

the composite slab per rib (Vt,Rd)

VRd,c

VRd,p

Figure 2.26 – Transverse shear resistance of composite slab per rib according to Stark[18]

28
Summing-up the partial resistances of the concrete rib and the steel deck for the total transverse shear
resistance of the composite slab is in great accordance with the Brazilian standard (clause Q.3.1.3.1 of
ABNT NBR 8800[20]). This is in contrast to the European standard in which the partial resistance of the
steel deck is totally neglected. It should be mentioned that in the Brazilian standard, the contributions
of the concrete and steel are determined according to different formulas. The effective area of the
concrete rib is also specified differently. This especially holds for trapezoidal profiles as can be seen in
Figure 2.27.

Figure 2.27 – Effective area of the concrete rib according to Brazilian standard:
trapezoidal profile (left) and re-entrant profile (right)[20]

Abspoel, Stark & Prins[21] carried out a research program in close cooperation with Dutch
Engineering[4] and Tuls[22] to gain more insight into the influence of the transverse shear force on the
hogging moment capacity of composite slabs. This is of importance for continuously supported
composite slabs and the current European standard lacks in information about this interaction.
Therefore, they tested three ComFlor 210 slabs and two ComFlor 75 slabs (NL, refers to ComFlor 60
in the UK). Second point of interest was to validate the contribution of the steel deck to the transverse
shear capacity of the composite slabs. The specimens were made of grade S350 for the steel deck,
grade B500 for the reinforcement bars and mesh, and class C30/37 for the concrete section. The
ComFlor 210 specimens contained one rebar Ø12 (rib reinforcement) and all specimens were provided
with a reinforcement mesh and additional rebars on top. The ComFlor 210 and ComFlor 75 specimens
had a nominal concrete topping of 70 mm and 90 mm respectively. No significant influence of the
transverse shear force on the hogging moment capacity was found according to the experiments. None
of the specimens failed in transverse shear, so quantitative determination of this resistance was not
possible. Still, from the measured transverse shear forces it was concluded that the current European
standard provides too conservative procedures for determination of the transverse shear capacity of
composite slabs. This applied to both ComFlor 75 (NL) and ComFlor 210. A maximum cylinder load of
147.18 kN was measured for which the ComFlor 210 specimen (two ribs) still didn’t fail in transverse
shear (including hogging moment conditions). This load was directly equal to the transverse shear force
acting on the specimen due to the used test setup (neglecting self-weight).

Pereira, Simões & Duarte[17] performed two tests on shallow composite slabs to evaluate the
transverse shear resistance according to equation 2.13 (visualised in Figure 2.28). Comparison with
equation 2.12 (Stark[18]) shows that VRd,p is substituted by Vb,Rd, which makes that the contribution of
the steel deck is now calculated according to the formula that is provided by the Eurocode 3.
Consequently, the effect of shear buckling is taken into account for determination of the transverse
shear capacity of the composite slab.

29
 ℎ𝑤
1 3 1 𝑠𝑖𝑛(𝜑) 𝑡 𝑓𝑏𝑣 2.13
𝑉𝑡,𝑅𝑑 = 𝑉𝑅𝑑,𝑐 + 𝑉𝑏,𝑅𝑑 = max (𝐶𝑅𝑑,𝑐 𝑘 (100𝜌𝑙 𝑓𝑐𝑘 )3 𝑏0 𝑑 ; 0.035 𝑘 2 𝑓𝑐𝑘 2 𝑏0 𝑑) + 2∙
𝛾𝑀0

Transverse shear resistance of

the composite slab per rib (Vt,Rd)

VRd,c

Vb,Rd

Figure 2.28 – Transverse shear resistance of composite slab per rib according to Pereira et al.[17]

The first tested slab collapsed due to the transverse shear force, since vertical cracks appeared near
the loading zone and the values of measured slip were low. For this test, the transverse shear
resistance according to the Eurocode 4 (section 2.5) was only 21.09% of the achieved capacity, while
the resistance according to equation 2.12 was 98.61% of the achieved capacity. The proposed model
therefore predicted the transverse shear resistance much closer to reality than the Eurocode 4 did. The
second tested slab failed in longitudinal shear for the reason that ductile slip behaviour was observed
with a maximum end-slip of approximately 13 mm. With this test it was therefore not possible to
determine the transverse shear resistance quantitatively. However, it could again be concluded that
Eurocode 4 is unnecessarily conservative: the calculated transverse shear resistance according to this
standard corresponded to 38.74% of the achieved capacity. This means that according to procedures
of Eurocode 4, the composite slab should already have failed in transverse shear instead of longitudinal
shear. Statistically speaking, these two tests don’t prove the rightness of equation 2.13, but they do
support the supposition that the Eurocode 4 doesn’t predict the transverse shear capacity accurately.

Hartmeyer & Kurz[27] also tried to come up with a new transverse shear model for composite slabs
based on the observation that it is not accurately described by the Eurocode 4. Research has been
carried out on composite slabs with shallow re-entrant profiled decks in combination with both normal
and lightweight concrete. Only normal weight concrete is considered in this thesis. They have looked
into more detail at the cracked state at failure and distinguished three different components that
contribute to the transverse shear capacity in the ultimate limit state:
▪ The vertical force component of the uncracked compressive zone Vc,cz;
▪ The transverse shear capacity of the metal sheet Vp,R;
▪ The local aggregate interlock Vc,ct (related to the tensile forces transmitted at the crack tip).
These components are shown in Figure 2.29. In experiments it was observed that across the major
inclined ‘bending shear crack’[27], transverse shear forces can be transferred to the metal deck up to
its capacity. By reaching a critical load, local buckling of the metal deck was observed at the location of
this critical crack. This is shown in Figure 2.30. The buckling allows a relative displacement of the
crack’s surfaces which activates aggregate interlock. The effectiveness of the interlock is mainly
dependent on the concrete’s properties, the bonding properties between the steel and concrete and the
existing crack width at the moment of local buckling [27]. It was also observed that the inclined crack
extended almost horizontally underneath the compression zone, until the section of maximum bending
moment was reached. This led to reaching the ultimate capacity of the composite slab.

30
 Figure 2.29 – Components that contribute to the transverse shear resistance of composite slabs
according to Hartmeyer & Kurz[27]

Figure 2.30 – Buckling of the webs at the location of the inclined crack[27]

The height of the compression zone xm depends on the composite action. The activated longitudinal
shear stress at the interface determines the effective tensile force in the steel deck which is transferred
into the concrete. Horizontal equilibrium of this tensile force (η∙Apm∙fypm) and the compressive force in
the concrete gives equation 2.14 to calculate xm. The factor η has to be determined in accordance with
Appendix B.3.6 of EN 1994-1-1[3] (see also the end of section 2.4). The height of the compressive zone
xm is calculated in the section of maximum moment (underneath the local line load) and can be used in
equation 2.15 to calculate the contribution Vc,cz to the total transverse shear resistance[27].

31
 𝜂 ∙ 𝐴𝑝𝑚 ∙ 𝑓𝑦𝑝𝑚
𝑥𝑚 = 2.14
𝑏𝑐 ∙ 𝑓𝑐𝑚
𝑥𝑚 𝑏𝑐
2
𝑉𝑐,𝑐𝑧 = ∫ ∫ 𝜏𝑥𝑧 (𝑧) ∙ 𝑑𝑦 ∙ 𝑑𝑧 = ∙𝑥 ∙ 𝑏 ∙ 𝑓 2.15
0 0 3 𝑚 𝑐 𝑐𝑡𝑚

With:
𝑥𝑚 mean height of the compressive zone [mm];
𝑉𝑐,𝑐𝑧 shear capacity of the uncracked compressive zone [N];
𝜂 degree of shear connection according to partial shear method of EN 1994-1-1[3] [-];
𝑓𝑦𝑝𝑚 yield strength steel deck [MPa]
𝐴𝑝𝑚 area of the steel deck [mm2];
𝑏𝑐 width of the concrete at the top [mm];
𝑓𝑐𝑚 mean cylindric compressive strength concrete [MPa];
𝑓𝑐𝑡𝑚 mean tensile strength concrete [MPa].

For the effect of the aggregate interlock Vc,ct an energetic approach was chosen by Hartmeyer &
Kurz[27] based on investigations by Hillerborg, Modéer & Petersson (1976), and Walraven & Reinhart
(1981) on the fracture process at the tip of the crack. Equation 2.16 can be used to calculate this
contribution to the total transverse shear resistance of the composite slab. Since this part of the model
is related to the tensile stresses at the crack tip, there is no need for description of the crack path[27].

𝐺𝑓 ∙ 𝐸𝑐𝑚
𝑉𝑐,𝑐𝑡 = 0.12 ∙ 𝑙𝑐ℎ ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑡𝑚 = 0.12 ∙ ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑡𝑚 2.16
𝑓𝑐𝑡𝑚 2

With:
𝑉𝑐,𝑐𝑡 the local aggregate interlock [N];
𝑙𝑐ℎ characteristic length of the crack [mm];
𝑏𝑐 width of the concrete at the top [mm];
𝑓𝑐𝑡𝑚 mean tensile strength of the concrete [MPa];
𝐺𝑓 fracture energy: 𝐺𝑓 = 0.024 + 0.026 ∙ 𝑓𝑐𝑡𝑚 (empirical relation as used in [27] for normal
weight concrete) [N/mm];
𝐸𝑐𝑚 Young’s modulus [MPa].

For the contribution of the steel deck Vp,R, Hartmeyer & Kurz[27] referred to information provided by the
supplier of the steel decks. According to their model, the total mean transverse shear resistance Vt,Rm
of a composite slab is obtained by summing up the separate contributions as shown in equation 2.17.
This method has been compared with experimental values and showed a good conformity[27]. To come
to a design value, mean values have been converted to characteristic values and a safety factor has
been added as shown in equation 2.18. The presented model is only valid for re-entrant composite
slabs (with embossments) in combination with a safety factor of 1.35[27].

2 𝐺𝑓∙𝐸𝑐𝑚
𝑉𝑡,𝑅𝑚 = 𝑉𝑝,𝑅𝑚 + 𝑉𝑐,𝑐𝑧 + 𝑉𝑐,𝑐𝑡 = 𝑉𝑝,𝑅𝑚 + (3 ∙ 𝑥𝑚 + 0.12 ∙ 𝑓𝑐𝑡𝑚 2
) ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑡𝑚 2.17

1 1 2 𝜂∙𝐴𝑝 ∙𝑓𝑦𝑝𝑘 𝐺𝑓∙𝐸𝑐𝑚

𝑉𝑡,𝑅𝑑 = 𝛾 ∙ 𝑉𝑡,𝑅𝑘 = 1.35 ∙ (𝑉𝑝,𝑅𝑘 + (3 ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑘
+ 0.12 ∙ 𝑓𝑐𝑡𝑘 2
) ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑡𝑘 ) 2.18
𝑉

32
Trapezoidal profiles are generally manufactured with bigger heights of the webs, leading to larger
transverse shear capacities[27]. Their bond capacities are usually worse than for re-entrant profiles,
which makes that bigger crack widths may be expected at the time-moment of reaching the capacity of
the steel deck[27]. It is therefore suggested by Hartmeyer & Kurz[27] that aggregate interlock shouldn’t
be considered for trapezoidal profiled composite slabs. They described a fundamental value for the
transverse shear capacity Vt,0,Rm in which the effect of the aggregate interlock is neglected (see equation
2.19). This value has been compared with tests on trapezoidal profiled composite slabs with lightweight
concrete and showed a good correspondence[27]. For normal weight concrete, no comparison was
made. In their paper it is proposed to put more effort in further investigation of the effect of additional
rib reinforcement, as this should reduce the crack width and may activate the aggregate interlock for
trapezoidal profiled composite slabs as well.

2
𝑉𝑡,0,𝑅𝑚 = 𝑉𝑝,𝑅𝑚 + 𝑉𝑐,𝑐𝑧 = 𝑉𝑝,𝑅𝑚 + 3 ∙ 𝑥𝑚 ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑡𝑚 2.19

As mentioned in the introductory part, this thesis focuses on the transverse shear capacity of deep
composite slabs. Table 2.1 on the next page gives an overview of the different models that have been
found in literature to calculate this capacity (mean values, Vt,Rm). A point of interest is how these models
relate to the finite element results of this thesis, because the models aren’t officially valid (Eurocode,
br/bs > 0.6) or derived (Stark[18]; Pereira et al.[17]; Hartmeyer & Kurz [27]) for deep composite slabs.

33
 Table 2.1 – Overview of the different models to calculate the mean transverse shear capacity per rib
Eurocode 4 (refers to the procedure of Eurocode 2 as given below)

Transverse shear capacity composite slab:

𝑉𝑡,𝑅𝑚 = 𝑉𝑅𝑚,𝑐 Parameters
VRm,c α = angle of web
Partial resistance concrete section: relative to the
1 3 1
𝑉𝑅𝑚,𝑐 = max (𝐶𝑅𝑚,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑚 )3 𝑏𝑤 𝑑 ; 0.035 𝑘 2 𝑓𝑐𝑚 2 𝑏𝑤 𝑑) vertical axis
b0 = mean width
𝑘 ≤ 2.0 ; 𝜌𝑙 ≤ 0.02
concrete rib
Partial resistance steel deck: bw = minimum
None width concrete rib
in tensile area
Stark[18] bc = width of the
Transverse shear capacity composite slab: composite slab at
𝑉𝑡,𝑅𝑚 = 𝑉𝑅𝑚,𝑐 + 𝑉𝑅𝑚,𝑝 the top
VRm,c CRm,c = emperical
Partial resistance concrete section: coefficent
1 3 1
𝑉𝑅𝑚,𝑐 = max (𝐶𝑅𝑚,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑚 )3 𝑏0 𝑑 ; 0.035 𝑘 2 𝑓𝑐𝑚 2 𝑏0 𝑑) d = effective depth
fbv = shear
𝑘 ≤ 2.0 ; 𝜌𝑙 ≤ 0.02 strength steel
VRm,p Please notice the use of b0 instead of bw considering
buckling
Partial resistance steel deck:
𝑓𝑦 fcm = mean
𝑉𝑅𝑚,𝑝 = 2 ∙ cos(𝛼) ∙ ℎ𝑙 ∙ 𝑡 ∙ compressive
√3 strength concrete
Pereira et al.[17] fctm = mean tensile
strength concrete
Transverse shear capacity composite slab:
fy = yield strength
𝑉𝑡,𝑅𝑚 = 𝑉𝑅𝑚,𝑐 + 𝑉𝑏,𝑅𝑚
VRm,c steel
hl = height web,
Partial resistance concrete section:
1 3 1 see Figure 2.24
𝑉𝑅𝑚,𝑐 = max (𝐶𝑅𝑚,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑚 )3 𝑏0 𝑑 ; 0.035 𝑘 2 𝑓𝑐𝑚 2 𝑏0 𝑑) hw = height web,
𝑘 ≤ 2.0 ; 𝜌𝑙 ≤ 0.02 see Figure 2.24
Vb,Rm Please notice the use of b0 instead of bw k = size factor
φ = angle of web
Partial resistance steel deck: relative to the
ℎ𝑤 horizontal axis
𝑉𝑏,𝑅𝑚 = 2 ∙ ∙ 𝑡 ∙ 𝑓𝑏𝑣
𝑠𝑖𝑛(𝜑) ρl = reinforcement
ratio
Hartmeyer & Kurz [27]
t = thickness web
Transverse shear capacity composite slab: xm = height of
𝑉𝑡,𝑅𝑚 = 𝑉𝑐,𝑐𝑧 + 𝑉𝑝,𝑅𝑚 compressive zone
Vc,cz concrete
Partial resistance concrete section:
2
𝑉𝑐,𝑐𝑧 = ∙ 𝑥𝑚 ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑡𝑚
3

Partial resistance steel deck:

Vp,Rm 𝑉𝑝,𝑅𝑚 is the transverse shear capacity of the
steel deck as provided by the supplier

34
3. FINITE ELEMENT ANALYSIS
In this chapter the results of the Finite Element Analysis (FEA) of ComFlor 210 will be discussed. It has
been chosen to analyse this type of composite slab, since most of the found research on deep
composite slabs was also conducted to ComFlor 210 (Van Erp[16], Abspoel et al.[21]). It has been tried
to quantitively determine its transverse shear capacity with the finite element software Diana[14]. The
FEA is started by merely modelling the concrete section of ComFlor 210 (neglecting the steel deck at
the bottom), after which the total composite slab will be analysed to show the influence of the steel deck
on the transverse shear capacity.

3.1. Geometry and loading conditions

To give the most accurate prediction of the transverse shear capacity, it is obvious that the FEA should
be in 3D. In the Eurocode 4, no specific test method is given to determine the transverse shear capacity
of a composite slab. Procedures for determination of the longitudinal shear capacity are well described
though. As explained in section 2.4, with both the m-k method and partial shear method, four-point
bending tests have to be carried out. It has been chosen to also simulate a four-point bending test for
determination of the transverse shear capacity of ComFlor 210. This is shown in Figure 3.1. The
corresponding distributions of transverse shear force and bending moment are also provided in this
figure. The loading conditions make that the transverse shear force is constant over the shear span Ls
and equal to the applied load Vt (i.e. line load integrated over its width). Therefore, when transverse
shear failure occurs, it can be established that the transverse shear capacity of the slab is directly equal
to the applied load Vt. Furthermore, the bending moment is constant between the loading points and its
maximum value is only dependent on Vt and Ls (so not the span L). The part of the span between these
loading points is called the region of constant bending moment.

Figure 3.1 – Four-point bending test with corresponding bending moment and transverse shear force
distribution

The setup for the finite element model is given in Figure 3.2 (side view). Due to symmetry around mid-
span of the four-point bending test, only half of the span needs to be modelled. Proper boundary
conditions are applied on the left side to account for this symmetry. As also shown in the figure, it has
been chosen to model a steel support plate (80 mm x 56 mm x 20 mm) at the right lower side. With
35
this, it is tried to prevent compressive failure at the support due to the concentrated stresses in vertical
direction, especially since the width of the support is relatively small (56 mm). Furthermore, the span L
is chosen to be equal to 2000 mm with an overhang length Lo of 355 mm (see Appendix B for
calculation). Both values are kept constant throughout the whole FEA. A span of 2000 mm is off course
rarely applied in practice for deep composite slabs, but choosing a bigger span will unnecessarily
increase the computation time of the FEA.

Figure 3.2 – Test setup of the FEA (side view)

FEA results are usually compared to real-life test results for sake of verification of the model. However,
clear measurements of the transverse shear capacity of ComFlor 210 haven’t been found. The research
of Abspoel et al.[21] has been the only one in which some conclusions were made about the transverse
shear capacity of ComFlor 210 based on real test values. In their research, a maximum transverse
shear force of 147.18 kN was measured for which the ComFlor 210 still failed due to the hogging
bending moment (concrete in tension, steel in compression). This indirectly implies that the transverse
shear capacity of the specimen was even bigger than this measured 147.18 kN. Since this is the only
test value of ComFlor 210 that could be found, it has been decided to start the FEA by analysing the
same cross-section as in the research of Abspoel et al.[21].

Figure 3.3 – Symmetry of cross-section with two ribs (ComFlor 210)[32]

In the research of Abspoel et al.[21], the tested ComFlor 210 contained two complete ribs. Due to the
symmetry of two ribs around their middle (Figure 3.3), only one of them needs to be modelled. This
makes that in total, due to double symmetry of the composite slab, only one quarter has been
implemented in the FEA. Table 3.1 shows a very schematic representation of dimensions of ComFlor
210 (for one rib). The steel deck is drawn with straight parts for sake of simplicity. A distinction is made
between the two types of analyses that will be executed in this thesis. First, only the concrete section
of ComFlor 210 is analysed, so neglecting the steel deck at the bottom (but including all other
reinforcement). This is shown in the upper image of Table 3.1. Then, the steel deck is added to the
model, to investigate its influence on the transverse shear capacity of ComFlor 210. This is shown in
36
the lower image. As mentioned, all given dimensions are taken from the research of Abspoel et al.[21].
When any of these dimensions are changed in the FEA, it will be specifically mentioned. In the top layer
of the slab, a reinforcement mesh and additional reinforcement are present. These additional rebars
and the part of the mesh that reinforces in longitudinal direction, are simply assigned to the plane “mesh
x-direction”. The part of the mesh that reinforces in the other direction is assigned to the plane “mesh
y-direction”. The inclusion of these additional rebars in the top layer is the direct result of trying to
analyse the same cross-section as in the research of Abspoel et al. [21].

Table 3.1 – Schematic representation of the dimensions for the two types of FEA’s (one rib)

FEA concrete section of ComFlor 210 (steel deck not included)

FEA ComFlor 210

i.e. concrete section + steel deck

b1 600 mm c 40 mm

0.96 mm
b2 175 mm t
(i.e. 1 mm excl. coatings)
Ø8-150 +
b3 56 mm Mesh x-direction
Ø12-150 (mm)

ht 280 mm Mesh y-direction Ø8-150 (mm)

hp 210 mm dx-direction 29 mm

Ø 12 mm dy-direction 37 mm

37
3.2. FEA concrete section of ComFlor 210
Although the main goal of the FEA is to determine the transverse shear capacity of ComFlor 210, the
first point of interest is to validate this capacity for merely the concrete part of the composite slab.
According to the Eurocode 4, the transverse shear capacity of a composite slab is only determined by
its concrete part. The potential additional resistance provided by the steel deck is not considered.
Therefore, determining the transverse shear capacity of the concrete section of ComFlor 210, actually
equals the same procedure that is followed by the Eurocode 4 for calculation of the total transverse
shear resistance of the composite slab. Figure 3.4 gives an overview of the finite element model for the
analysis. Solid elements with quadratic interpolation have been used for the concrete and the support
plate. All reinforcement has been modelled with embedded elements. The general dimensions of the
cross-section are just as given in Table 3.1 (upper image).

Concrete

Boundary conditions Distributed (line) load

(symmetry)

Support plate

Support

a) 3D view

Mesh / additional rebars

No steel deck is present, but the
bottom of the concrete section still Rib reinforcement
follows the profile that is normally
created by this steel deck.

b) Front view (mesh not visible)

Figure 3.4 – Finite element model concrete section

38
3.2.1. Material properties
Table 3.2 gives an overview of the material properties for the FEA of the concrete section. The strength
values are taken from the research of Abspoel et al.[21]. Figure 3.5, Figure 3.6 and Figure 3.7 are
provided in addition to clarify some of the mentioned terms in Table 3.2. For the tensile diagram of
Hordijk, the area underneath this graph is equal to Gf/h i.e. the fracture energy divided by the crack
bandwidth. The crack bandwidth h is used to calculate the crack width from the crack strain (w = ε * h).
It defines over what length the crack width w is smeared out. As also mentioned in Table 3.2, it has
been chosen to use a rotating crack model. This means that the orientation of a crack within an element
is always perpendicular to the tensile principal direction, so that no shear stress is present along this
crack. Any change in principal direction, due to redistribution of stresses for instance, leads to rotation
of the crack. This is shown in Figure 3.6. It may be clear that rotation of an existing crack sounds
somewhat unphysical. However, the big advantage of this rotating crack model is that there is no need
for specifying relationships for shearing along the crack.

Table 3.2 – Material properties FEA concrete section (as specified in DIANA[14])
Concrete
Mean compressive strength (fc) 41.8 MPa
Compressive curve Thorenfeldt (Figure 3.5)
Mean tensile strength (ft) 3.1 MPa
Tensile curve Hordijk (Figure 3.5)
Mode-I tensile fracture energy (Gf) 0.1046 N/mm
Young’s modulus (Ec) 33789 MPa
Poisson’s ratio (v) 0.2
Crack model Rotating (Figure 3.6)
Rib reinforcement and mesh
Shape of diagram Bilinear (Figure 3.7)
Yield strength (fy) 536 MPa (εy = 2.68∙10-3)
Ultimate strength (fu) 574 MPa (εu = 5∙10-2)
Young’s modulus (Es) 200000 MPa
Support plate
Young’s modulus (Es) 210000 MPa
Poisson’s ratio (v) 0.3

Figure 3.5 – Compressive (left) and tensile (right) behaviour of concrete in the FEA[14]

39
 A crack develops in an element by exceeding the tensile strength of the material in the principal direction;
When this principal direction rotates, the crack rotates too;
No shearing occurs along the crack.
Figure 3.6 – Rotating crack model

Figure 3.7 – Bilinear stress-strain diagram for steel

3.2.2. Analysis procedure

As discussed in section 2.3, transverse shear failure may become the governing failure mode for short
shear spans. It has been explained that, when assuming an equal transverse shear capacity over the
full span of the slab, reduction of the shear span will lead to transverse shear failure. However, as also
described by the Eurocode 2, more of the applied load is directly transferred to the support when the
shear span becomes smaller. To reduce this effect of direct load transfer and accurately measure the
transverse shear capacity, the shear span Ls should be bigger than a certain limit value. In the Eurocode
2, the effect of direct load transfer is considered as follows: when the shear slenderness ratio (av/d) is
smaller than 2.0, the measured shear force Vt should be multiplied with β = av/(2d). This reduction takes
into account the effect of direct load transfer. The distance av is taken as the horizontal distance
between the applied load and the edge of the support plate, in accordance with clause 6.2.2 of EN
1992-1-1[24] (see Figure 3.8). In this thesis, av is related to the shear span Ls in the following way: Ls =
av + 40 mm.

Figure 3.8 – Definition of different parameters

However, Rombach & Henz[29] concluded in their research that direct load transfer is still possible up
to shear slenderness ratios of 3.0. This is based on test results of concrete slabs without shear

40
reinforcement (stirrups). For that reason, it has been chosen to start the FEA with a shear slenderness
ratio of 3.0. When transverse shear failure doesn’t occur, the shear span Ls (and so the shear
slenderness ratio av/d) is further reduced. With this, it is expected that either transverse shear failure
occurs, or that the effect of direct load transfer becomes decisive and makes that bending failure is the
governing mode for every shear slenderness ratio.

Because of the non-linear material properties that have been assigned to the concrete, a non-linear
analysis is required as well to solve the problem. The load is applied in a force-controlled incremental
way, in combination with the so-called arc-length method to overcome convergence problems due to
softening behaviour of the concrete (cracking). This is illustrated in Figure 3.9. To overcome a limit point
in a regular force-controlled process, negative load steps should be specified after this limit point to find
convergence. This means that the location of the limit point should be known a priori, which is generally
not the case. By using the arc-length method, this problem is solved because the method doesn’t have
difficulties with finding convergence after limit points (during softening, see Figure 3.9). Regarding the
iteration process within every load step, regular Newton-Raphson method has been used. This method
is shown in Figure 3.10. With this method, the stiffness is adapted every iteration to find convergence
of a certain load step. Convergence is found when both energy and displacement norms are satisfied
with a tolerance of 0.005 and 0.05 respectively. When convergence isn’t found after 30 iterations, the
analysis moves on with the next load step instead of stopping the analysis. This is because quite often,
when the cracking process in the concrete is somewhat unstable, the analysis can’t find convergence
for these specific load steps. This can lead to a consecutive series of non-converging load steps as
also shown in Figure 3.11 (red bars). Off course, the range of this series shouldn’t be too big (like a
maximum of 3 load steps as shown in the figure) and it should also not happen over the whole analysis.
When this occurs, the size of the load step should be reduced. Post-processing of the results regarding
this non-convergence is very important to check whether the FEA results are still reliable. In the FEA,
only physical non-linearity has been considered.

Limit point

Softening

Force controlled Arc-length method

Figure 3.9 – Incremental loading[14]

Figure 3.10 – Iterative process: Regular Newton-Rapson method[14]

41
 Figure 3.11 – Convergence of the FEA with respect to crack development in concrete

3.2.3. Results and discussion

Figure 3.12 shows the load-displacement curves of the FEA for 4 different shear slenderness ratios
av/d. The shear slenderness ratio defines at which distance the force Vt is located from the support. For
instance, a shear slenderness ratio of 2 means that the distance between the load and the edge of the
support plate (av) is 2 times the effective depth of the slab (d). For the ratios 2.50 and 1.54, the ultimate
load-carrying capacity wasn’t reached, because too few load steps had been specified in the FEA
(values written in red in Figure 3.12). Still, it can be concluded that bending failure was the governing
mode for every analysis. Quite ductile behaviour has been observed, which occurred due to yielding of
the rib reinforcement at the locations of these vertically oriented bending cracks. This is shown in Figure
3.13. Inclined cracking is also clearly visible, but opening of the vertically oriented cracks at the left from
the applied load (region of constant bending moment) caused failure in every simulation. In Table 3.3,
the bending moment capacity of the concrete section has been derived from two different analyses.
Both values show good conformity when being compared. It can be seen in Figure 3.12 that the failure
load increases when the shear slenderness ratio is reduced: when the load is close to the support,
more load is needed to reach the bending moment capacity.

Table 3.3 – Derivation of bending moment capacity from FEA results

Shear slenderness ratio av/d 2.99 2.05
Ultimate load Vt 23.65 kN 33.28 kN
Effective depth d 280 – 40 – 6 = 234 mm 280 – 40 – 6 = 234 mm
av (see Figure 3.8) 2.99 ∙ 234 = 700 mm 2.05 ∙ 234 = 480 mm
Shear span Ls (= av + 40 mm) 700 + 40 = 740 mm 480 + 40 = 520 mm
Bending moment capacity MR 23.65 ∙ 0.740 = 17.50 kNm 33.28 ∙ 0.520 = 17.30 kNm

42
 av/d

Figure 3.12 – Load-displacement curves FEA concrete section for different shear slenderness ratios

av/d = 2.99 av/d = 2.50

Inclined crack

av/d = 2.05 av/d = 1.54

Vertical crack due to pure bending;
yielding of rib reinforcement.
Figure 3.13 – Crack strain plots at failure for FEA concrete section
Since bending failure occurred in all previous analyses, it has been chosen to adjust the cross-section
by doubling the diameter of the rib reinforcement. The rib reinforcement now has a diameter of 24 mm
instead of 12 mm. It is expected that this will mainly increase the bending moment capacity of the
concrete section, so that transverse shear failure may become governing. Keep in mind that a rib
reinforcement diameter of 24 mm may be rarely seen in practice, but it is not too unrealistic for
experimental purposes. The increased rib reinforcement diameter leads to a reduced concrete cover c
at the bottom (34 mm instead of 40). All other dimensions and parameters are kept the same as
described before in Table 3.1.

Failure of the finite element model has been simulated for shear slenderness ratios 3.50, 2.99, 2.50
and 2.05. Figure 3.14 shows the load-displacement curves of the analyses. Figure 3.15 shows the
corresponding cracking patterns at failure. It is clearly visible that the concrete section failed in
transverse shear for every shear slenderness ratio. The cracking patterns match with what is known as
“flexural shear failure”[28]. In short, this failure mode is characterized by the formation of bending cracks
within the shear span, which start to curve towards the loading point under influence of the transverse
shear force. Opening of these inclined cracks within the shear span led to total failure of the concrete
element.
43
 av/d

Figure 3.14 – Load-displacement curves FEA concrete section for different shear slenderness ratios
(rib reinforcement: Ø24 mm instead of Ø12 mm)

av/d = 3.50 av/d = 2.99

av/d = 2.50 av/d = 2.05

Figure 3.15 – Crack strain plots at failure for FEA concrete section
(rib reinforcement: Ø24 mm instead of Ø12 mm)

From the load-displacement curves in Figure 3.14 it may be established that the failure load increases
when the shear slenderness ratio is reduced. This is the result of direct load transfer to the support and
can be explained on basis of the following. Figure 3.16 shows the crack formation in the concrete
element for the FEA analysis with shear slenderness ratio av/d = 2.50. Different snapshots are given,
marked with corresponding dots in the load-displacement curve. Starting from the beginning of the
analysis, it may be established that the load keeps increasing until the first major crack arises in the
concrete. This occasion is marked with a green dot (1) and leads to a setback in the load-displacement
graph. The crack is located in the section underneath the loading point. What follows next is a clear
reduction in overall stiffness as a result of the proceeding crack formation. At the occasion
corresponding to the yellow dot (2), the most right inclined crack start to arise. This crack has been fully
developed at the moment of the orange dot (3), after which horizontal cracking starts to occur along the
rib reinforcement (dowel cracking[28]). Although it may be expected that this horizontal cracking
initiates total failure of the concrete section, still significant more load is needed to cause this. The
difference between the starting of this dowel cracking and reaching the failure load is more than 10 kN.
This additional load-carrying capacity is provided by the sort of arch-like structure[28] that arises in the
concrete section after crack formation.

44
 4
Crack starts at the 2 3
bottom by exceedance 1
of the tensile strength Crack curves towards
due to the bending the loading point
moment under influence of the
transverse shear force

1 2

3 4

“Dowel cracking”: cracking along the rib reinforcement

Figure 3.16 – Flexural shear failure of the concrete section (FEA av/d = 2.50 ; Ø24 mm)

The functioning of this arch structure is illustrated in Figure 3.17. It shows that even after full crack
formation, the load P can still be carried by the diagonally oriented compressive struts, for which the
reinforcement acts as a tensile chord[28]. For high shear slenderness ratios, formation of the critical
inclined crack often leads to total failure of the reinforced concrete member, because the arch structure
isn’t able to carry the load[28]. When the shear slenderness ratio is reduced, it generally holds that the
capacity of the arch structure is improved due to the bigger inclination of the compressive strut with
respect to the longitudinal axis (because the load is closer to the support). Due to the bigger inclination
of the strut, more compressive load is taken by this strut for the same load P. This can even give a
different kind of transverse shear failure mode, namely “shear compression failure”[28]. This failure
mode is characterized by compressive failure of the diagonal strut (i.e. splitting cracks). This haven’t
been observed in any analysis of this thesis.

Figure 3.17 – Arch-like structure in reinforced concrete members[28]

45
 Figure 3.18 – Comparison of FEA results with calculated resistances
(rib reinforcement: Ø24 mm instead of Ø12 mm)

Figure 3.18 shows the transverse shear capacity VRm,c of the concrete section for every shear
slenderness ratio that has been analysed. This picture clearly shows the dependency of VRm,c on the
shear slenderness ratio av/d. A difference of 21.25 kN in capacity can be found between the two ultimate
shear slenderness ratios (3.50 and 2.05). The FEA shows that even for shear slenderness ratios bigger
than 3.0, direct transfer of the load to the support is still possible. For shear slenderness ratio 3.50, a
12.3% lower capacity has been found compared to that of 2.99. Looking at Figure 3.18, it seems that
an exponential curve could be drawn through the results of the FEA. Assuming that this is true, it can
be established that for big shear slenderness ratios, the transverse shear capacity approaches a
constant value (asymptote) that excludes any effect of direct load transfer.

The transverse shear capacity according to the Eurocode 2 is also given in Figure 3.18. The calculation
procedure according to this standard is provided in the first row of Table 2.1 at the end of chapter 3. A
factor CRm,c of 0.15[28] has been used to calculate the mean transverse shear capacity VRm,c of the
concrete section. The calculation is shown below.

200 200
𝑘 = min (1 + √ 𝑑
; 2.0) = min (1 + √234 ; 2.0) = min(1.925 ; 2.0) = 1.925

𝐴𝑠𝑙 0.25 ∙ 𝜋 ∙ (24)2

𝜌𝑙 = min ( ; 0.02) = min ( ; 0.02) = min(0.0345 ; 0.02) = 0.02
𝑏𝑤 ∙ 𝑑 56 ∙ 234
1 3 1
𝑉𝑅𝑚,𝑐 = max (𝐶𝑅𝑚,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑚 )3 𝑏𝑤 𝑑 ; 0.035 𝑘 2 𝑓𝑐𝑚 2 𝑏𝑤 𝑑)
1 3 1
= max (0.15 ∙ 1.925 ∙ (100 ∙ 0.02 ∙ 41.8)3 ∙ 56 ∙ 234 ; 0.035 ∙ 1.9252 ∙ 41.82 ∙ 56 ∙ 234)

= max(16540 ; 7917)
= 16540 𝑁 = 16.54 𝑘𝑁

The calculated capacity according to the Eurocode 2 is quite conservative. The lowest transverse shear
capacity that has been found in the FEA is equal to 37.80 kN and this is still 2 times bigger than the
calculated resistance of 16.54 kN. According to the Eurocode 2, it is only allowed to use the smallest
width in the tensile area bw (56 mm) for calculation of VRm,c. However, the concrete cross-section has a
great varying width over its height as also illustrated in Table 3.1: starting at 56 mm at the bottom (b3),
46
175 mm at top of the tapered rib (b2) and finally a width of 600 mm at the top (b1). It is suggested to
replace this parameter bw in calculation by the mean width of the concrete rib b0, as also proposed by
Stark[18] and Pereira et al.[17]. For calculation of this parameter b0, it is assumed that the concrete rib
smoothly transits from 56 mm at the bottom to 175 mm at the top. With this, the mean width of the rib
is simply calculated as (175+56)/2 = 115.5 mm. This b0 is located at mid-height of the steel deck (i.e.
0.5∙hp = 0.5∙210 = 105 mm). This is illustrated in Figure 3.19.

Figure 3.19 – Mean width of the concrete rib b0 (ComFlor 210)

Replacing bw by b0 makes that the transverse capacity VRm,c is calculated as follows:

𝑘 = 1.925 (𝑛𝑜𝑡 𝑎𝑓𝑓𝑒𝑐𝑡𝑒𝑑)

𝐴𝑠𝑙 0.25 ∙ 𝜋 ∙ (24)2
𝜌𝑙 = min ( ; 0.02) = min ( ; 0.02) = min(0.0167 ; 0.02) = 0.0167
𝑏0 ∙ 𝑑 115.5 ∙ 234
1 3 1
𝑉𝑅𝑚,𝑐 = max (𝐶𝑅𝑚,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑚 )3 𝑏0 𝑑 ; 0.035 𝑘 2 𝑓𝑐𝑚 2 𝑏0 𝑑)
1 3 1
= max (0.15 ∙ 1.925 ∙ (100 ∙ 0.0167 ∙ 41.8)3 ∙ 115.5 ∙ 234 ; 0.035 ∙ 1.9252 ∙ 41.82 ∙ 115.5 ∙ 234)

= max(32149 ; 16328)
= 32149 𝑁 = 32.15 𝑘𝑁
This value approaches the lowest found resistance of 37.80 kN much more closely, and therefore
supports the proposal to use b0 instead of bw in the Eurocode’s formula (as shown above). The
improvement of this adjusted calculation is also clearly visible in Figure 3.18. The calculated transverse
shear capacity of 32.15 kN seems to be provide a reasonable asymptote to the FEA results of this
thesis.

3.3. FEA ComFlor 210

After verification of the transverse shear capacity of the concrete section, this thesis will now focus on
the influence of the steel deck on this capacity. On basis of the literature review it may be expected that
the steel deck will give a significant improvement of the transverse shear capacity. A FEA of ComFlor
210 has been executed to investigate this expectation. Figure 3.20 gives an overview of the finite
element model. Information about the geometry and loading conditions has already been provided in
section 3.1. With respect to the dimensions given in Table 3.1, please be aware that the rib
reinforcement in the finite element model contains a diameter Ø of 24 mm instead of 12 mm (all other
47
dimensions are still the same). This has been applied to be consistent with the last FEA of the concrete
section, so that comparison of the results is possible. A consequence of this decision is that the
analysed cross-section of ComFlor 210 isn’t completely in in accordance with the research of Abspoel
et al.[21], which was initially suggested. Solid elements with quadratic interpolation have been used for
the concrete, steel deck and support plate. For the steel deck, at least 2 elements over its thickness
have been implemented to account for local bending of the plate. Furthermore, a small transition zone
has been included between the different thicknesses at the bottom of the steel deck, to reduce the
stress concentration at this location (Figure 3.21). Interface elements have been used for
implementation of the slip behaviour between the steel deck and concrete.

Concrete

Boundary conditions Distributed (line) load

(symmetry)

Support plate

Support

Steel deck

a) 3D view

Mesh / additional rebars

Rib reinforcement
(Ø24 mm)

b) Front view (mesh not visible)

Figure 3.20 – Finite element model ComFlor 210

48
 Single thickness (t)

Transition zone

Double thickness (2t)

Figure 3.21 – Small transition zone between the different thicknesses at the bottom of the steel deck

3.3.1. Material properties

The material properties of the concrete part and reinforcement have already been provided in section
3.2.1. These parameters are kept the same for the FEA of ComFlor 210. As known from the literature
review, the profile of the steel deck contains embossments which are of importance for the longitudinal
shear capacity due to the mechanical interlock between the concrete and steel. These embossments
also influence the cross-sectional area of the steel deck that is effective in transmission of stresses.
Because of the wrinkling of the steel sheet due to the embossments, it flattens out when it is loaded in
tension. This initial reduced stiffness makes that the embossments in the steel deck are less effective
in the transmission of stresses than the adjacent parts. Therefore, in EN 1994-1-1[3] it is prescribed
that the parts with embossments should be neglected when determining the effective area of the steel
deck. This is off course very conservative since some tensile stresses will be transmitted through these
parts. Especially since the concrete partly prevents the embossments from flattening out (illustrated in
Figure 3.22). As mentioned in section 2.4, Van Erp[16] therefore assumed an effectiveness of 12.5%
for the parts of the steel deck with embossments, based on his findings. A 1 mm thickness will then be
reduced to 0.125 mm in design. A visual representation of the effective cross-section for ComFlor 210
is given in Figure 3.23.

Figure 3.22 – Concrete preventing the embossments from flattening out[16]

Figure 3.23 – Effective cross-section ComFlor 210: bold parts are fully effective; other parts have a
reduced effectiveness due to the embossments.[16]

49
The use of an effective cross-section with thickness reduction (as shown in Figure 3.23) is commonly
used in the design of composite slabs or other thin-walled cold-formed steel sections. This reduction in
thickness is off course not representative for the real-life behaviour of the embossments, which makes
it also not very suitable for application in the FEA. To account for the reduced effectiveness of the
embossments, it is therefore chosen to change the material properties of these parts as also proposed
by Veljković[34]. Figure 3.24 shows the stress-strain relations for different parts of a steel deck based
on tensile tests[34]. Comparing the dimpled sheeting (spec. 3) with the flat sheeting (spec. 6), shows
that the initial Young’s modulus is reduced as well as the yield strength. Veljković[34] assumed a
reduction of 53% for both material parameters. The dimpled sheeting also has a steeper hardening
plateau. For FEA purposes, Veljković[34] used the approximated stress-strain relationships as shown
in Figure 3.25. From these relationships it can be established that the ultimate strength is reduced as
well (approx. 27%).

Figure 3.24 – Stress-strain diagrams for different parts of a steel deck based on tensile tests[34]

-27 %

-53 %

Figure 3.25 – Approximated stress-strain relationships for the steel deck in the FEA of Veljković[34]

Figure 3.26 shows the cross-section of the steel deck as how it has been implemented in the FEA. The
longitudinal stiffeners in the webs and upper flanges are included in the model, but they are approached
by straight parts. Furthermore, all rounded corners are modelled as straight corners. The geometry of
the lower flange is also approached by straight parts, with a double thickness to account for the overlap.
As can be seen in the figure, large parts of the upper flange and webs are less effective due to the
embossments. Reduced material properties have been assigned to these parts of the cross-section. All
material parameters for the steel deck in the FEA are shown in Table 3.4, of which the non-reduced
parameters again are in accordance with the research of Abspoel et al.[21].

50
 Figure 3.26 – Cross-section steel deck as used in the FEA (unit: mm)

Table 3.4 – Material properties FEA ComFlor 210 (as specified in DIANA[14])
Concrete; reinforcement; support plate.
See Table 3.2
Steel deck
Shape of diagram Bilinear (Figure 3.7)
Yield strength (fy) 409 MPa (εy = 1.95∙10-3)
Ultimate strength (fu) 490.8 MPa (εu = 16∙10-2)
Young’s modulus (Es) 210000 MPa
Reduced yield strength (fy,red) 192.15 MPa (εy = 1.83∙10-3)
Reduced ultimate strength (fu,red) 306.75 MPa (εu = 16∙10-2)
Reduced Young’s modulus (Es,red) 105000 MPa
Poisson’s ratio v 0.3

Regarding the provided material properties in Table 3.4, please notice the following:
▪ The applied reduction in yield strength is equal to 53% in accordance with Veljković [34];
▪ The Young’s modulus has been reduced with 50% and the ultimate strength has been reduced with
37%. As mentioned earlier, these values should have been 53% and 27% (derived from the paper
of Veljković[34]). Some small calculation errors were made by the author in determining the reduced
material properties which led to this deviation in percentages. However, since these reduced
properties are already quite an assumption, the effect of these calculation errors is neglected.
▪ An ultimate strain value εu of 16% has been assumed (for reduced and non-reduced properties) in
accordance with NEN 10346[38].

The reduced material properties for the embossments are mainly based on the reduced effectiveness
of these parts in the longitudinal direction (local x-direction). This is illustrated in Figure 3.27a. However,
whether these reduced material properties correctly describe the behaviour of the embossments in any
direction may be doubted. It is illustrated in Figure 3.27b and -c that for these other loading conditions,
the material seems to be less affected by the wrinkling of the plate. Therefore, it may be expected that
the that the steel deck behaves more or less orthotropic at the locations of the embossments. With
respect to this orthotropy, detailed information is missing. For that reason, it has been chosen to apply
an isotropic material model for the embossments with reduced properties, to be on the conservative
side.

51
 z

x
a) Top view embossments; tensioned in local x-direction

b) Sideview embossments; sheared in local xy-plane

c) Sideview embossments; compressed in y-direction

Figure 3.27 – Behaviour embossments for different loading conditions

Figure 3.28 shows the type of interface element that has been applied in the finite element model to
simulate the interaction between the steel deck and concrete. This element can describe the behaviour
in normal direction (local z-axis) and shear directions (local x- and y-axis). The local x-axes of all
interface elements are aligned with the global x-axis of the finite element model. The bonding in this
direction is the strongest due to the orientation of the embossments (shown in Figure 3.29). The bond-
slip diagram for this x-direction is shown in Figure 3.30. The shape of the diagram is based on research
by Ríos, Cifuentes, Martínez-De La Concha & Medina-Reguera[35] on numerical modelling of shear-
bond behaviour in composite slabs. As can be seen, the diagram exists out of 3 different linear
segments: starting with (almost) full interaction, then a branch of partial interaction and finally a
descending part[35]. This shape of the diagram can generally be observed in the load-displacement
curves of composite slabs for longitudinal shear failure[16][35]. The provided values that characterise
this bond-slip relation are based on the experiments by Van Erp[16]. Calculation of these values is given
in Appendix C. In contrast to the slip behaviour in x-direction, simple linear material properties have
been assigned to the local y- and z-direction of the interface elements. To this y-direction, a 1000 times
smaller stiffness (0.000154 N/mm3) has been implemented compared to that of the x-direction (0.154
N/mm3), since it is assumed that the embossments don’t effectively restrain the slip in this y-direction
(Figure 3.29). To the z-direction, which represents the contact between the steel and concrete, a very
high linear stiffness has been applied (154000 N/mm3). Separation between the concrete and steel will
therefore not occur.

52
 Figure 3.28 – 3D surface interface element in Diana[14]

Figure 3.29 – Orientation of the local directions with respect to the embossments

Figure 3.30 – Bond-slip behaviour for local x-direction of interface elements[16][35]

3.3.2. Analysis procedure

It has been chosen to use a shear slenderness ratio av/d of 2.99 (shear span Ls = 740 mm). From the
former analysis of the concrete section, it was concluded that direct transfer of the load to the support
is still possible for bigger shear slenderness ratios. However, this will also increase the chance on
longitudinal shear failure or bending failure as known from section 2.3. Both physical and geometrical
non-linearity have been considered in the FEA. The presence of this slipping interface has been the
main reason to also include geometrical non-linear behaviour in the analysis. For further information
about the solution procedure, please see section 3.2.2. Convergence is found when both energy and
force norms are satisfied with a tolerance of 0.005 and 0.05 respectively.

53
3.3.3. Results and discussion
Figure 3.31 shows the load-displacement curve of ComFlor 210 (blue), together with the load-
displacement curve of the former analysis of the concrete section (orange). The rib reinforcement
diameter and shear slenderness ratio were the same in both analyses.

Increase in load-
carrying capacity
due to the steel
deck

Figure 3.31 – Results FEA ComFlor 210 and FEA concrete section
(rib reinforcement Ø24 mm; av/d = 2.99)

The composite slab (ComFlor 210) failed at a load of 99.86 kN. Unfortunately, bending failure was the
governing mode. The composite slab failed by crushing of the concrete top layer (compressive zone),
just underneath the application of the line load. This is shown in Figure 3.32 at the locations of the red
dashed circles, where the compressive stresses in longitudinal direction start to drop to zero. Figure
3.33 shows the Von Mises stresses in the steel deck at maximum load. Yielding at the bottom of the
steel deck did occur within the region of constant bending moment. The effect of this yielding isn’t
directly visible in the load-displacement diagram of Figure 3.31. This is caused by the fact that the stress
in the rib reinforcement (455.86 MPa) stayed below its yield strength (536 MPa), before failure of the
composite slab occurred. Figure 3.34 shows the cracks in the composite slab at maximum load. It may
be established that opening of the cracks has been very limited. From Figure 3.32 it may also be
observed that a rather big longitudinal crack was present in the top part of the composite slab (location
where the principal stresses are almost zero). This crack resulted from the secondary bending in y-
direction, for which the mesh in the top layer acted as reinforcement. The maximum value of slip in
longitudinal direction was equal to 1.02 mm, so longitudinal shear failure didn’t occur.

The FEA of the concrete section reached a maximum load of 43.11 kN as a consequence of transverse
shear failure. When both analyses are being compared, it may be concluded that the steel deck does
significantly improve the transverse shear resistance of the composite slab. Adding the steel deck to
the finite element model led to an increase in the load-carrying capacity from 43.11 kN to 99.86 kN.
Although this later failure load characterizes the bending moment capacity of the composite slab (=
99.86∙0.74 = 73.90 kNm), it also provides a minimum value for its transverse shear resistance (>99.86
kN). So, according to the FEA of ComFlor 210, the composite slab has a minimal transverse shear
capacity of 99.86 kN. For barely the concrete part of this composite slab, the transverse shear capacity
is equal to 43.11 kN. So, according to the analyses, the additional transverse shear resistance provided
by the steel deck is minimally equal to 99.86 – 43.11 = 56.75 kN.

54
Longitudinal crack
due to bending in
y-direction

Compressive failure

Figure 3.32 – Principal stresses in concrete at maximum load (FEA ComFlor 210)

Yielding

Figure 3.33 – Von Mises stresses in steel deck at maximum load (FEA ComFlor 210, plotted in
integration points)

Figure 3.34 – Crack strain plot at maximum load (FEA ComFlor 210)

Figure 3.35 shows the principal stresses in the steel deck at maximum load (FEA ComFlor 210). Only
one side of the steel deck is shown in this figure, but it can be assumed that the stress distribution in
the steel deck is the same for the other side (i.e. the other web). It can be seen that, within the shear
span, the webs of the steel deck are mainly subjected to tensile stresses in the principal directions.
Most of these principal stresses even show sort of uniaxial state: tensile stress in one direction, (almost)
no stress in the other direction. This stress state is quite unexpected when we consider the steel deck
to be barely responsible for carrying a part of the transverse shear force. This difference in expectation
55
 versus observation is also shown in Figure 3.36. The most obvious explanation for this, is that the
stresses in the steel deck’s webs are significantly influenced by the cracking in the concrete, making
that the webs are subjected to tensile stresses in both longitudinal direction (σx) and transverse direction
(σy) within the shear span Ls. Off course, the steel deck’s webs are subjected to shear stresses as well
(τxy). With the tensile stresses bridging the cracks, the steel deck acts as reinforcement to the concrete.
As shown in Figure 3.35, the tensile (principal) stresses in the webs are mostly below their yield strength
(i.e. 192.15 MPa for embossments; 409 MPa for normal parts). A small part of the embossments in the
webs is above its yield strength (green tensors), but since the embossments do have a quite steep
hardening plateau, opening of the cracks is still prevented. From the FEA of the concrete section, it is
known that transverse shear failure would have been observed in case of absence of the steel deck. It
has been shown that opening of the inclined cracks within the shear span caused transverse shear
failure of the concrete element. By including the steel deck to the finite element model, opening of these
inclined cracks is restrained. The steel deck acts as transverse shear reinforcement to the concrete,
which makes that the transverse shear capacity is increased.

Embossments

Shear span Ls

“Uniaxial” tension
Figure 3.35 – Principal stresses in steel deck at maximum load (FEA ComFlor 210)

equivalent
(local coordinates
of web)
Expectation Observation
Figure 3.36 – Principal stresses around mid-height of steel deck’s webs (FEA ComFlor 210)

From this FEA of ComFlor 210, it may also be questioned whether shear buckling is expected to occur
when considering transverse shear failure of a deep composite slab. The dark blue tensors in Figure
3.35 represent the compressive stresses in the steel deck. It can be seen that the upper flanges of the
steel deck are mostly in compression, and so are the webs closely above the support. However, within
the shear span, no significant dark blue tensors can be observed in the webs. The webs are mainly
subjected to tensile stresses in one diagonal direction; the compressive stresses in the other diagonal
direction are very low or absent (Figure 3.36). Since these compressive stresses are the reason for
shear buckling of the webs (see Appendix A), it is doubted whether this kind of buckling may be
observed in case of transverse shear failure of a deep composite slab.
56
In Figure 3.37 it is shown that opening of an inclined crack generally leads to a horizontal displacement
in global x-direction and a vertical displacement in global z-direction. This horizontal opening of the
crack is definitely restrained by the steel deck’s webs, since the transfer of stresses between the
concrete and steel is assured by the embossments in this direction. However, for vertical opening of
the crack, the local transfer of stresses between the steel deck’s webs and concrete is assumed to be
nihil. It has been explained in section 3.3.1 that, due to the orientation of the embossments, low linear
stiffness properties have been assigned to the interface elements in local y-direction. Slip between the
concrete and the steel deck’s webs is therefore free to occur in this direction, making that the steel
deck’s webs aren’t locally reinforcing the cracks for this vertical opening. This seems a bit contradictory
with the observation in the FEA that the steel deck’s webs are subjected to tensile stresses in both
longitudinal direction (σx) and transverse direction (σy) within the shear span. However, this can be
explained on basis of the following. In the FEA, detachment between the steel deck and concrete
couldn’t occur. With the ends of the webs being fully attached to the concrete, vertical opening of the
cracks (global z-direction) makes that the webs can still be stressed in the local y-direction. So, due to
the chosen interface properties, the steel deck’s webs are able to reinforce the concrete by resisting
the tensile stresses from the inclined cracks in both longitudinal direction and vertical direction.

Global coordinates xlocal = xglobal

ylocal ≠ zglobal (webs do have an angle
with respect to z-axis)

Orientation of
embossments with
respect to the crack;
local coordinates.
Figure 3.37 – Opening of an inclined crack[28]

Separation between the steel deck and concrete is not considered,

so the ends of the webs are fully attached to the concrete.

Cross-section
inclined crack

Crack tries to open in vertical

Steel deck’s webs are subjected
direction (global z-direction), so
to tensile stresses in local y-
the concrete part below this crack
direction due to crack opening.
wants to move downwards.

Figure 3.38 – How vertical opening of the inclined cracks is restrained by the webs in the finite element
model

57
With the statement that the steel deck’s webs are able to reinforce the inclined cracks in a vertical
sense, it may be established that the functioning of this is comparable to stirrups in ordinary reinforced
concrete beams. These stirrups are positioned at a certain mutual distance in the concrete beam,
whereby they bridge the existing transverse shear cracks (Figure 3.39a). The stirrups take over the
tensile stresses that can’t be resisted by the cracks, whereby they prevent these cracks from opening.
For the transverse shear resistance of such reinforced concrete members, it is assumed by the
Eurocode 2 that the internal distribution of stresses can be described with a strut-and-tie model as
shown in Figure 3.39b and -c. The upper chord of this model represents the concrete compressive
zone, as the lower chord represents the tensile rebars. The vertical and diagonal members of the model
represent the stirrups and concrete compressive struts respectively. The process for transfer of the
transverse shear force F from loading point to support can be explained as follows. First, the external
force F flows from the top of the beam through the concrete to the bottom of the adjacent stirrup. A
diagonally oriented compressive strut arises in the concrete. This strut is loaded by a force F√2. At the
bottom of the stirrup, this force is guided to the top of the beam due to the tensile forces that arise in
the stirrup. The total tensile force that is taken by the stirrup is equal to F. Once arrived at the top, again
the force flows diagonally from the top of the beam to the bottom of the next stirrup. The process is
repeated until the support is reached. It may be established that the total transverse shear resistance
of the concrete beam is determined by the partial resistances of the struts (compressive failure) and
stirrups (yielding). In clause 6.2.3 of EN 1992-1-1[24], formulas are provided to calculate these partial
resistances.

Cross-section Side view

Stirrup
Rebar

a) Stirrup bridges the existing crack and is loaded in tension

b) Strut-and-tie model for reinforced concrete member[36]

c) Forces in strut-and-tie model[36]

Figure 3.39 – Strut-and-tie model for reinforced concrete beams with stirrups

58
With respect to the transverse shear capacity of deep composite slabs, it is wondered whether the
same kind of strut-and-tie model can be applied. The proposal for such a strut-and-tie model is shown
in Figure 3.40 (for one rib). The diagonal members represent the compressive struts in the concrete rib
and the vertical members represent the tensile stresses in the steel deck’s webs. It is assumed that
these tensile stresses in the steel deck can be represented by one vertical member, in the middle of
the shear span Ls. This assumption is based on the stress distribution of Figure 3.41, which shows the
normal stresses in global z-direction in the steel deck for one web (FEA of ComFlor 210). It clearly
shows a peak in the middle of the shear span. The choice for one vertical member within the shear
span restricts the model to two concrete compressive struts (see Figure 3.40). When the strut-and-tie
model is applied to the finite element model of this thesis, it gives that the compressive struts have an
angle θ of 34.3° with respect to the longitudinal axis. The Eurocode 2 gives restrictions for this angle,
namely 21.8° ≤ θ ≤ 45°, which is satisfied. The compressive principal stresses in the concrete are shown
in Figure 3.42. The orientation of these compressive stresses corresponds well with the compressive
struts in the model of Figure 3.40.

Upper chord: concrete compressive zone

Lower chord: steel deck’s lower flange
Diagonal members: concrete compressive struts
Vertical members: tensile stresses in steel deck’s webs

Figure 3.40 – Strut-and-tie model for load transfer in deep composite slabs

Figure 3.41 – Normal stress in global z-direction in steel deck at mid-height

(for one web; maximum load (failure); FEA ComFlor 210)

59
 Figure 3.42 – Principal compressive stresses in concrete at maximum load (FEA ComFlor 210)

The validity of the proposed strut-and-tie model (Figure 3.40) as calculation method for the transverse
shear capacity of deep composite slabs, haven’t been confirmed by the analysis. The description of the
load transfer in the composite element by this 2D model can be assessed as a big simplification. This
especially holds for the stress state in the steel deck’s webs: the strut-and-tie model assumes that the
steel deck’s webs can only be stressed in vertical direction (i.e. global z-direction = vertical component
of local y-direction of the webs), but this is off course never the case. Furthermore, it is also not clear
how to convert the steel deck’s webs into such linear members for the strut-and-tie model. Using an
effective length (in x-direction) may be the most obvious solution to this.

However, when considering that the steel deck’s webs act as transverse shear reinforcement in a
similar way as stirrups, does provide a valid argument for the significant increase in transverse shear
capacity like observed in the FEA. The efficiency of the load-transferring mechanism (strut-and-tie
model) seems to be the main reason for the increase. However, the model relies on a good collaboration
between the concrete and steel deck. Using a strut-and-tie model as proposed, with different materials
for the various members, is only valid when transfer of load between the different materials is assured.
For that reason, accurate description of the interface behaviour between the concrete and steel deck
plays an important role as well.

With respect to this interface behaviour, it is being questioned wheter the adhesive capacity between
the steel deck and concrete is big enough, so that the assumption of no separation makes a good
approach. Suppose both materials to be completely free to separate from eachother, then the steel
deck loses its function as stirrup. With the embossments only restraining the slip in longitudinal direction
and the ends of the webs not being attached to the concrete, opening-up of the cracks can no longer
result in tensile stresses in the steel deck’s webs in local y-direction (σy). On the other hand, due to the
possible mechanical interlock of the embossments in this y-direction (Figure 3.43), the steel deck’s
webs may still be stressed in this direction when restraining the inclined cracks from opening.

Mechanical interlock
in y-direction
z

Figure 3.43 – Local directions of both the interface and the steel deck’s web

60
So, when considering the transverse shear capacity of deep composite slabs, the following may be
established. The steel deck’s webs have the potential to act as transverse shear reinforcement in the
same way as stirrups in reinforced concrete. The load is assumed to be mainly transferred by
compressive struts in the concrete ribs and vertical tensile stresses in the steel deck’s webs. These
tensile stresses in the webs restrain the cracks in the concrete from opening, so good collaboration
between the separate materials is necessary. In the finite element model, this collaboration has been
assured by assuming that the steel deck and concrete couldn’t separate from each other. In reality, for
the steel deck’s webs to effectively restrain the crack opening in a vertical sense, most probably relies
on the mechanical interlock of the embossments in the local y-direction. Therefore, accurate description
of the total interface behaviour between the steel deck and concrete is a critical aspect to determine
whether this stirrup-functioning of the webs really occurs. With respect to this interface behaviour,
detailed information is missing as listed below:
▪ How does the embossments restrain the slip in local y-direction?
▪ How big is the adhesive strength between the steel deck and concrete in local z-direction?
▪ Wat is the influence of separation between the steel deck and concrete (in local z-direction) on
the slip behaviour (in local x- and y-directions)?

So, the assumption that the steel deck and the concrete can’t separate from each other, is considered
to be the biggest simplification of the finite element model. However, due to the above-mentioned
missing links, accurate implementation of the interface properties is not possible. Still, in the next
section it is tried to find a lower bound value for the transverse shear capacity of ComFlor 210, by using
other interface properties.

3.4. FEA ComFlor 210: a lower bound value for the transverse shear capacity
From the FEA results in former section 3.3, it could be concluded that the steel deck does give a
significant improvement of the transverse shear capacity. Due to the chosen properties of the interface,
the steel deck’s webs were able to reinforce the concrete in a similar way as stirrups in regular
reinforced concrete beams. However, it is doubted whether these interface properties are correct and
if this stirrup-functioning of the steel deck’s webs is representative for the actual transverse shear
behaviour of deep composite slabs.

So, it is wondered what will happen to the transverse shear resistance when the steel deck and concrete
are free to separate from each other. This behaviour is quite easy to implement in the finite element
model by specifying a no-tension relationship for the interface elements in local z-direction. In reality, it
could be that some adhesive strength is present between the steel deck and concrete. But since this is
not known, assuming that the interface contains no tensile strength in local z-direction may be
considered as a conservative approach. Implementation of this no-tension behaviour for the interface
in local z-direction, also affects the slip behaviour in longitudinal direction (local x-direction of the
interface elements). It isn’t hard to imagine that the effect of mechanical interlock from the embossments
becomes less when the steel deck separates from the concrete. However, the crosslink between this
behaviour in local z-direction and local x-direction isn’t known. In the finite element model of section
3.3, it has been assumed that the embossments mainly restrain the slip in local x-direction, so that
slipping in local y-direction is (nearly) free to occur. Opening of the interface therefore doesn’t affect the
behaviour in local y-direction. However, when it is assumed that the embossments also restrain the slip
in local y-direction, and a more accurate behaviour is implemented in the finite element model (like for
the local x-direction), then opening of the interface affects the behaviour in both x- and y-directions.

61
To overcome this ignorance about an appropriate longitudinal shear relationship (local x-direction)
when considering separation between the steel deck and concrete (in local z-direction), it is suggested
to assume that slipping in this longitudinal direction is almost free to occur (like for the local y-direction).
A full sliding connection for the interface in longitudinal direction is likely to give numerical problems, so
some slip restraining properties should be implemented. However, by setting a low linear stiffness for
the interface in local x-direction (so the slip is not effectively restrained) and including separation
between the steel deck and concrete (for tensile conditions in local z-direction), it is tried to investigate
if the steel deck is still possible to provide additional transverse shear resistance.

3.4.1. Material properties & analysis procedure

The no-tension behaviour for the local z-direction of the interface elements can be specified in Diana[14]
with a predefined function. It is needed to specify a stiffness, a critical interface opening and a stiffness
reduction factor. For this stiffness, a value of 1540 N/mm3 has been applied. This stiffness mainly
defines the compressive behaviour of the interface. Compressive failure of the concrete occurs at a
stress value of 41.8 MPa, which means that the interface has the potential to show a negative relative
displacement of 41.8/1540 = 0.027 mm before failure of the concrete occurs. In real life, this isn’t
possible since the steel deck and concrete can’t move through each other. However, this inaccuracy is
neglected because the relative displacement is very small. Increasing the stiffness may improve this
inaccuracy, but it tried to prevent large stiffness differences between the different directions of the
interface elements to overcome convergence problems. The critical interface opening is set to 1e-6 mm
and the stiffness reduction factor is set to 1e-6. This means that when the critical interface opening is
reached (for tensile conditions), the stiffness is reduced to 1e-6∙1540 = 0.00154 N/mm 3.

For the longitudinal shear behaviour (local x-direction of the interface elements), a linear stress-slip
relationship has been applied with a stiffness that is 100 times smaller than the initial stiffness in the
finite element model of section 3.3. The longitudinal shear behaviour for the current analysis and the
former analysis are shown in Figure 3.44 with the orange and blue lines respectively. The reduction in
stiffness has been chosen carefully: when the stiffness becomes very low, it is likely to give
convergence problems.

100 times smaller stiffness

Figure 3.44 – Differences in longitudinal shear behaviour for FEA in section 3.3 and 3.4

So, regarding the interface properties, the following can be summed up:
▪ For the behaviour in local z-direction, the stress-displacement graph follows a linear diagram with a
stiffness of 1540 N/mm3 for compressive conditions. As soon as the relative displacement reaches
a positive value of 1e-6 mm (opening), this stiffness is reduced to 0.00154 N/mm3 for tensile loading
conditions. With this low tensile stiffness and high compressive stiffness, it is tried to simulate that
the steel deck and concrete can separate from each other, but can’t move through each other at the

62
 interface. However, with this compressive stiffness being non-infinite, some small negative relative
displacement is allowed for the interface elements. Also, with the tensile stiffness being larger than
zero, some tensile stresses will always be transmitted between the steel deck and concrete when
the interface opens. It is assumed that both inaccuracies can be neglected.
▪ The longitudinal slip behaviour in x-direction follows a linear diagram with a stiffness of 0.00154
N/mm3. Longitudinal slip is not effectively restrained due to this small stiffness, but some composite
behaviour is still guaranteed to prevent convergence problems.
▪ The slip behaviour in local y-direction of the interface elements is left unchanged with respect to the
former FEA in section 3.3 (linear diagram with stiffness of 0.000154 N/mm3).

Please be aware that the assumed interface properties are totally non-realistic parameters and can’t
be used for accurate simulation of composite behaviour of ComFlor 210. However, it is tried to simulate
an imaginary experiment in which the interface is barely capable in preventing the concrete and steel
deck to move through each other. So, when an external load is applied on top of the composite slab, it
gives that the steel deck deflects together with the concrete section. Due to this combined deflection,
both elements are stressed as shown in Figure 3.45. Equilibrium at the considered cross-section makes
that Vt = Vc + Vwebs. So, a part of the load is carried by the concrete section (Vc) and a part of the load
is carried by the transverse shear force in the webs of the steel deck (Vwebs). The described calculation
methods in section 2.7 all rely on the same kind of superposition. It is expected that including the steel
deck to the finite element model still gives an increase in transverse shear capacity, even when the
composite behaviour is limited.

Figure 3.45 – General forces in composite slab due to external loading

For more information about the material properties of the concrete and steel deck, see section 3.2.1 &
3.3.1. Again, in this FEA of ComFlor 210, the rib reinforcement diameter is equal to 24 mm and a shear
slenderness ratio of 2.99 (shear span Ls = 740 mm) is used. For more information about the analysis
procedure, see section 3.3.2.
63
3.4.2. Results and discussion
The load-displacement graph of the executed FEA of ComFlor 210 is plotted in Figure 3.46 with a green
line. The results of the other two FEA’s (section 3.2 and 3.3) are also shown in this figure (orange and
blue lines). The rib reinforcement diameter and shear slenderness ratio were the same in all analyses.

Figure 3.46 – Load-displacement graphs FEA’s ComFlor 210 and FEA concrete section
(rib reinforcement Ø24 mm; av/d = 2.99)

In section 3.3, the composite slab failed in bending at a load of 99.86 kN (blue graph in Figure 3.46).
By changing the interface properties of the composite slab (see section 3.4.1) and redoing the FEA,
this failure load is reduced to 65.27 kN (green graph in Figure 3.46). The composite slab now failed in
transverse shear instead of bending. Figure 3.47 shows the formation of this clear inclined crack within
the shear span. Compressive failure of the concrete’s compressive zone underneath the location of the
external load became decisive in the same way as in the former analysis of ComFlor 210 (section 3.3,
Figure 3.32). However, yielding of the bottom flange of the steel deck haven’t been observed this time.
The same holds for yielding of the rib reinforcement. The combination of this and the cracking pattern
in the concrete, makes that the failure mode is assessed as transverse shear failure.

Compressive
failure

Figure 3.47 - Crack strain plot at maximum load

So, allowing the steel deck and concrete to separate from each other and reducing the composite
behaviour between the two elements, led to a reduction in load-carrying capacity of ComFlor 210. Still,
by making the comparison with the FEA of the concrete section (orange graph in Figure 3.46), it can
be concluded that the steel deck does give a significant increase in transverse shear capacity.
According to both analyses, the additional resistance provided by the steel deck is equal to
65.27 – 43.11 = 22.16 kN. This confirms the expectation that, even when the steel deck and concrete
are not effectively acting together due to these bad interface properties, the steel deck is still able to
64
provide additional transverse shear resistance. The steel deck deflects together with the concrete due
to the external load, making that it is stressed and takes part in resisting this external load.

Figure 3.48 shows the compressive (blue) and tensile (red) principal stresses in the steel deck just after
the first load step of the FEA. The analysis is clearly in the linear-elastic stage and the stress distribution
in the steel deck is as expected. The upper flange of the steel deck is in compression as the lower
flange is in tension. Within the shear span, the principal stresses around mid-height of the webs are
rotated at an angle of approx. 45° due to the shear stresses. By considering the structural behaviour of
the steel deck in the ULS as an extension of this initial stress state, it is easy to explain the observed
buckling behaviour of the steel deck as shown in Figure 3.49. This figure actually shows the opening of
the interface in local z-direction, but this representative for the buckling behaviour of the steel deck. By
allowing the steel deck and concrete to separate from each other, the steel deck becomes susceptible
to local flange buckling, shear buckling and buckling of the webs above the support due to the
introduced compressive stresses. Comparing Figure 3.49 with Figure 3.48 shows that the observed
buckling pattern in the steel deck corresponds well with these initial compressive stress trajectories.
However, the local buckling of the steel deck makes that the stress distribution in the ULS is very
irregular and difficult to assess. This is shown in Figure 3.50. From this figure, it may also be seen that
the maximum principal stress values (649.23 MPa and -575.47 MPa) are beyond the specified material
strengths, which shouldn’t be possible. However, this is due to the extrapolation of the stresses from
the integration points to the outer surfaces of the elements. The combination of local buckling and the
limited number of elements over the plate thickness of the steel deck is likely to cause this inaccuracy.
Figure 3.50 shows the Von Mises stresses in the steel deck at maximum load, plotted in the integration
points of the elements. From this it may be concluded that the steel deck is hardly at yielding, except
for this location above the support at the inside of the webs. This yielding occurred due to the very local
bending at this kink in the steel deck’s webs.

Rotation of principal
stresses due to shear
stresses in the webs

Figure 3.48 – Principal stresses in steel deck just after the first load step:
blue is compression, red is tension.

Flange buckling
Shear buckling
Buckling of the webs
above the support

Figure 3.49 – Relative displacement of interface elements in local z-direction at maximum load

65
 a) Blue is compression, red is tension

b) Discrete colour scale

Figure 3.50 – Principal stresses in steel deck at maximum load
(stresses at the outer surfaces of the elements)

Yielding above support

due to very local bending
at this kink in the steel
deck’s webs

Figure 3.51 – Von Mises stresses in steel deck at maximum load

(values in integration points of elements)

In former section 3.3, it is mentioned that shear buckling isn’t expected to occur when considering the
transverse shear capacity of ComFlor 210. This statement is based on the stress distribution in the
steel deck’s webs during failure of the composite slab in the FEA (Figure 3.35). However, by changing
the interface properties of the finite element model, it is shown that the steel deck becomes more

66
susceptible to local buckling, including shear buckling of the webs. So, the interface behaviour plays an
important role for this local buckling of the steel deck. Of course, when the steel deck can’t separate
from the concrete at the interface, like in the FEA of section 3.3, local buckling of the steel deck becomes
nearly impossible: the steel deck first has to detach from the concrete by reaching the concrete’s tensile
strength in the elements close to the interface. Therefore, the inclusion of this no-tension relationship
for the interface elements in local z-direction makes it easier for the steel deck’s webs to buckle. In
addition, due to the chosen interface properties, it can be assumed that only function of the steel deck’s
webs is to resist a part of the transverse shear force. When there are no other factors that influence the
stress distribution in the steel deck’s webs (like cracking in the concrete for instance), it may give that
the steel deck becomes more susceptible to shear buckling as well. However, regarding the observed
buckling behaviour of the steel deck, it may be question whether the finite element model accurately
covers this. It has been explained in section 3.3.1 that reduced material properties have been applied
to the parts of the steel deck with embossments. But, when considering local buckling of these parts, it
is expected that the embossments give an increased resistance due to the actual geometry: the moment
of inertia for out-of-plane bending is increased by the embossments.

67
68
4. DISCUSSION
• How big is the transverse shear capacity of the concrete section and how does this relate to the
current Eurocode 2: Concrete structures (EN 1992)?
As explained, the transverse shear capacity of a composite slab according to the Eurocode 4 is only
determined by its concrete part. The Eurocode 4 (composite structures) prescribes that the calculation
procedures of Eurocode 2 (concrete structures) should be used when it comes to the transverse shear
resistance of the total composite slab. Nothing is mentioned about the potential influence of the steel
deck. So, for calculation of the transverse shear capacity of the composite slab, only the resistance
provided by the concrete section is being considered. For that reason, the first point of interest is to
validate the calculation procedures of Eurocode 2 for the transverse shear resistance of this concrete
section. The formulas that are provided by the Eurocode 2 are empirically derived for reinforced
concrete members, which generally have a different lay-out compared to the concrete ribs in deep
composite slabs.

In section 3.2, the results of the FEA of the concrete section have been discussed. Applying a Ø24 mm
bar as rib reinforcement resulted in transverse shear failure of the finite element model for every
analysed shear slenderness ratio av/d. This rebar diameter may be rarely applied in practice (for deep
composite slabs), but it has been used in the FEA to make sure that bending failure didn’t become
decisive, as this was the case with a Ø12 mm rebar. The reinforcement ratio ρl of the finite element
model can be calculated at 0.0345 (= 0.25π∙Ø2/(bw∙d)) when using a Ø24 mm rebar. According to the
Eurocode 2, only a maximum reinforcement ratio ρl of 0.02 is allowed to use in calculation of the
transverse shear capacity. Therefore, instead of doubling the rebar diameter like done in the FEA, it
may had been better to double the reinforcement ratio ρl in order to control this parameter a bit more.

Figure 4.1 – Direct transfer of the load to the support by this compressive strut in the concrete rib

The results of the FEA clearly showed that the effect of direct load transfer must be excluded for
accurate determination of the transverse shear capacity of the concrete element. The load shouldn’t be
too close to the support, so that the beneficial effect of this arising arch-like structure (concrete
compressive strut and tensile rib reinforcement) in the concrete element is eliminated. This is shown in
Figure 4.1. For that reason, the shear slenderness ratio av/d (or shear span Ls) should be bigger than
a certain limit value. Using a shear slenderness ratio of 3.0 resulted in a transverse shear capacity of
43.11 kN. The analysis with a shear slenderness ratio of 3.5 gave a lower capacity of 37.80 kN. So, it
69
may be established that there is still some dependency of the transverse shear capacity on the shear
slenderness ratio av/d for values bigger than 3.0.

When following the procedures of Eurocode 2, the transverse shear capacity can be calculated at 16.54
kN for the analysed cross-section (one concrete rib of ComFlor 210). This is not even half of the lowest
found value in the FEA (37.80 kN). As mentioned before, the reinforcement ratio ρl of the analysed
cross-section is equal to 0.0345, but only a maximum value of 0.02 may be used in the calculation
according to the Eurocode 2. However, when this limitation is neglected and the actual reinforcement
ratio (0.0345) is used to calculate the transverse shear capacity, a resistance of 19.84 kN is obtained
(instead of 16.54 kN). This is still far below the found capacity in the FEA. Therefore, it is proposed to
change the Eurocode’s calculation method by adjusting the effective width that is used in calculation of
the transverse shear capacity of the concrete rib. According to the Eurocode 2, only the smallest width
in the tensile area bw can be used. In analogy to Stark[18] and Pereira et al.[17], it is proposed to use
the mean width of the concrete rib b0. This gives a calculated resistance of 32.15 kN, which makes a
much better prediction when being compared to the FEA results of this thesis. This is illustrated in
Figure 4.2. Furthermore, using the mean width b0 of the concrete rib also gives that the reinforcement
ratio ρl now can be calculated at 0.0167 (= 0.25π∙Ø2/(b0∙d)), which is below 0.02.

1 3 1
𝑉𝑅𝑚,𝑐 = 𝐶𝑅𝑚,𝑐 𝑘(100𝜌𝑙 𝑓𝑐𝑚 )3 (𝑏𝑤 𝑜𝑟 𝑏0 )𝑑 ≥ 0.035 𝑘 2 𝑓𝑐𝑚 2 (𝑏𝑤 𝑜𝑟 𝑏0 )𝑑

Effective area when using bw Effective area when using b0

Figure 4.2 – Calculated transverse shear capacity of the concrete rib vs. FEA result

• What is the influence of the steel deck on the transverse shear capacity and how does this relate to
the current Eurocode 4: Composite structures (EN 1994)?
After verification of the transverse shear capacity of the concrete section, the steel deck has been
added to the finite element model to determine its influence on this capacity. Two different FEA’s of
70
ComFlor 210 have been executed, of which the results are described in section 3.3 and 3.4. From both
analyses it can be concluded that the steel deck does give a significant improvement of the transverse
shear capacity. However, due to the complexity of the problem and the strong non-linear behaviour of
the composite slab, it is difficult to give an exact explanation for the improvement. Still, 3 main factors
can be distinguished which give reasons for the increase in transverse shear capacity:
▪ Steel deck acting as reinforcement to the concrete like stirrups;
▪ Steel deck acting as reinforcement to the concrete like longitudinal rebars;
▪ Steel deck being able to resist a part of the transverse shear force in its webs.
The above-mentioned factors will be discussed separately.

Steel deck’s webs prevent the inclined cracks

from opening in transverse direction

Tensile stresses in
steel deck’s webs (σy)

Steel deck’s webs being stressed (tension) in local Transfer of the load within the shear span Ls
y-direction due to crack opening: may be described with a strut-and-tie model:
- Adhesive capacity between concrete and steel at - Diagonals representing the compressive struts
the ends of the webs (yellow lines); in the concrete;
- Mechanical interlock of the embossments in local - Vertical members for tensile stresses in steel
y-direction (blue dashed circles). deck’s webs.

Figure 4.3 – Stirrup-functioning of steel deck’s webs in deep composite slabs

It has been explained in section 3.3.3 that the steel deck’s webs have the potential to reinforce the
concrete in a similar way as stirrups. This means that vertical opening of the inclined cracks (in global
z-direction) is restrained by tensile stresses σy in the steel deck’s webs in local y-direction. This is shown
in Figure 4.3. However, for the steel deck’s webs to effectively restrain the crack opening this way, the
interface behaviour becomes very important. Either the ends of the webs should be connected to the
concrete (as has been the case in the FEA of 3.3: no separation at the interface), or the embossments
in the webs should restrain the slip in this local y-direction due to the local mechanical interlock.
Assuming that these conditions are satisfied, then the load-transferring mechanism in the deep
composite slab may be approached with a strut-and-tie model like proposed in Figure 4.3. The validity
of this strut-and-tie model for the transverse shear resistance of the deep composite slab haven’t been
confirmed in this thesis. Before further elaboration of the model, it should mainly become clear whether
71
this stirrup-functioning of the steel deck’s webs is representative for the actual transverse shear
behaviour of deep composite slabs. More knowledge should be gained on how to accurately describe
the interface behaviour in the multiple directions (local x-, y- and z-direction).

As explained in section 3.3.1, due to the orientation of the embossments in the steel deck, it is assumed
that the slip at the interface is mainly restrained in the longitudinal direction. So, due to the composite
behaviour, opening of the cracks in the concrete is restrained by the steel deck in this direction (shown
in Figure 4.4). The tensile stresses that are released from the cracks are transferred to the steel deck.
The steel deck may be considered to reinforce the concrete in a similar way as the rib reinforcement.
Evaluating the Eurocode’s formula for calculation of the transverse shear resistance of the bare
concrete section (rib), it can be established that the resistance is increased for bigger reinforcement
ratios ρl. So, assuming that the reinforcement ratio ρl is increased by the steel deck, leads to an increase
in transverse shear resistance according to this formula. Therefore, the steel deck is likely to provide
additional transverse shear resistance this way.

Steel deck prevents the inclined cracks

from opening in longitudinal direction

Tensile stresses in steel

deck’s webs (σx)

VRm,c

Steel deck Rib

reinforcement

Considering the steel deck to reinforce the

concrete in a similar way as the rib reinforcement,
Slip at the interface is restrained in making that the reinforcement ratio ρl is increased
longitudinal direction by the embossments by the steel deck, gives an increase in transverse
when assuming that longitudinal shear shear capacity VRm,c of the concrete rib according
failure doesn’t occur. the Eurocode’s formula.

Figure 4.4 – Steel deck acting as reinforcement in longitudinal direction

In the foregoing, the steel deck is considered to act as reinforcement to the concrete, which causes the
transverse shear capacity to increase. In addition, the steel deck can be assessed as separate
structural element as well, being able to carry a part of the load. Due to shape of the profile, the steel
deck can resist a part of the transverse shear force in its webs as shown in Figure 4.5. Since the
concrete section can’t move through the steel deck at the interface, both elements are being stressed
by the external load due to the combined deflection of the concrete and the steel deck. With a part of
the load being carried by the steel deck’s webs, it makes that the transverse shear resistance is
increased by the steel deck.

72
 Transverse shear stresses
in steel deck’s webs (τxy)

Figure 4.5 – Transverse shear stresses in steel deck’s webs

According to the FEA of ComFlor 210 in section 3.3, the steel deck provides a minimal increase in
transverse shear capacity of 56.75 kN per rib (can even be higher). In the FEA, the steel deck and
concrete couldn’t separate from each other and a more or less realistic longitudinal shear behaviour
had been implemented. Therefore, it can be assumed that all of the mentioned factors (Figure 4.3 -
Figure 4.5) have played a role in this big increase in transverse shear capacity of 131.6% with respect
to the resistance of the concrete section. However, especially because of the doubt whether this stirrup-
functioning of the webs (Figure 4.3) is valid for the transverse shear behaviour of deep composite slabs,
a second FEA has been executed as described in section 3.4. Due to the bad interface properties in
this later FEA of ComFlor 210, the first two factors (Figure 4.3 and Figure 4.4) are no longer valid. Still,
an increase in transverse shear capacity of 22.16 kN has been found. So, even when the composite
behaviour between the steel deck and concrete is bad, meaning that the steel deck isn’t effectively
reinforcing the cracks, the steel deck provides an increase in transverse shear capacity of 51.4%.
Therefore, the presence of these big webs in deep composite slabs is considered to be an important
factor for the transverse shear capacity, as the webs resist a part of the transverse shear force.

So, for the transverse shear capacity of deep composite slabs, the following may be established:
▪ When the composite behaviour between the concrete and the steel deck is neglected, then the
additional transverse shear resistance provided by the steel deck solely depends on the webs being
able to resist a part of the transverse shear force (Figure 4.5). According to the FEA of ComFlor
210, the steel deck provides an additional 22.16 kN per rib in this case, which is equivalent to an
increase of 51.4% in transverse shear capacity with respect to the resistance of the bare concrete
section.
▪ When considering that the composite behaviour in longitudinal direction is guaranteed by the
embossments, it may be expected that the transverse shear capacity of the composite slab is
increased because of the beneficial effect of the steel deck reinforcing the cracks in longitudinal
direction (Figure 4.4).
▪ When the interface behaviour between the concrete and steel deck is such, that the steel deck’s
webs are not only restraining the crack opening in longitudinal direction, but also in transverse
direction, then the transverse shear capacity becomes significantly bigger. The effectiveness of the
load being transferred by compressive struts in the concrete and tensile stresses in the steel deck’s
webs seems to be the main reason for this (Figure 4.3).

• How to calculate the transverse shear capacity of deep composite slabs?

With the Eurocode 4 failing in providing an adequate calculation method for the transverse shear
capacity of deep composite slabs, it is tried to come up with an appropriate replacement. The FEA of
73
this thesis will therefore be compared to three other models as described in the literature review:
Stark[18], Pereira et al.[17] and Hartmeyer & Kurz[27]. All these models are based on the same
assumption that, for engineering’s purposes, the total transverse shear capacity of the composite slab
should be calculated by adding the partial transverse shear resistances of the concrete section and the
steel deck. Comparison of the models with the FEA result of section 3.4 is given in Figure 4.6 (for
calculations, see Appendix D).

It has been chosen to compare the calculation models with the FEA result of section 3.4, because this
is the most conservative one. Due to the bad composite behaviour in this analysis, the steel deck could
only provide additional transverse shear resistance to the concrete by resisting a part of the transverse
shear force in its webs. Because of the non-realistic interface properties, it can be doubted whether the
executed FEA is representative for the actual transverse shear behaviour of the composite slab.
However, for accurate assessment of the transverse shear capacity of ComFlor 210 by means of FEA,
more detailed information about the interface behaviour is needed. The FEA of ComFlor 210 in section
3.3 contained a more realistic longitudinal slip behaviour, but due to the assumption that the steel deck
and concrete couldn’t detach from each other, it may have given an overestimation of the transverse
shear capacity.
Vt,Rm [kN]

Figure 4.6 – Comparison of different calculation models with the FEA result of section 3.4
(capacity in kN for one rib of ComFlor 210)

The transverse shear capacity according to the Eurocode 4 is also given in Figure 4.6, from which it is
easy to see that the standard is unnecessarily conservative. Considering the other models, it can be
established that the model of Stark[18] overestimates the transverse shear capacity of the composite
slab when being compared to FEA. The model gives a capacity of 123.55 kN, of which 91.40 kN is
based on the partial resistance of the steel deck. This seems to be the main reason for overestimation
of the capacity. The model supposes that the webs of the steel deck reach their full shear yield capacity
before total failure of the composite slab, but this haven’t been observed in any of the analyses. In
contrast to the model of Stark[18], the other two models do give a safe calculation procedure as shown
in Figure 4.6. With the model of Pereira et al.[17], the partial resistance of the steel deck is significantly
reduced (13.59 kN) when being compared to the model of Stark[18]. This is because the model of
Pereira et al.[17] takes into account the effect of shear buckling of the steel deck’s webs, leading to a
reduction in resistance. Although the model of Hartmeyer & Kurz[27] gives a slightly better prediction,
it is considered to be less suitable as calculation method, because it requires test values for the partial

74
resistance of the steel deck. A value of 44.98 N/mm (resistance per width of the steel deck) has been
used as mentioned in the report of Tuls[22]. So, it is recommended to use the model of Pereira et al.[17]
as replacement for the procedures of the Eurocode 4, for the following reasons:
▪ With the Eurocode 4 (composite structures) lacking in detailed information about the transverse
shear capacity of composite slabs, it may be established that it is simply not known how to correctly
calculate it and to take into account the additional resistance that is provided by the steel deck.
However, by combining the procedures of Eurocode 2 (concrete structures) and Eurocode 3 (steel
structures) as proposed by Pereira et al.[17], a more accurate transverse shear capacity of ComFlor
210 can be calculated, while still being safe.
▪ For calculation of the partial transverse shear resistance of the concrete section (VRm,c), the model
of Pereira et al.[17] uses the mean width of the concrete rib (b0) instead of the smallest width (bw).
This has been confirmed by the FEA to give a better prediction.
▪ It is not perfectly clear whether shear buckling is an effect that needs to be accounted for. In section
3.4, it has been shown that the steel deck’s webs may be susceptible to this kind of local buckling.
But, due to the complex interface behaviour in real-life and/or the increased resistance against local
buckling of the embossments, it may be that shear buckling will never be observed in experiments
for deep composite slabs. However, since the model of Pereira et al.[17] does consider as reduced
shear strength (fbv) for the steel deck’s webs to take into account the effect of shear buckling, it may
be assessed as a safe method to use until there is more known about this aspect.

At last, it can be established that for both hogging and sagging bending moment conditions, the
transverse shear capacity of ComFlor 210 is underestimated. Based on real experiments, Abspoel et
al.[21] found a minimal value of 73.59 kN (one rib) for hogging bending moment conditions. In this thesis,
a value of 65.27 kN is found for sagging bending moment conditions. The FEA that resulted in this
capacity included very bad interface properties (section 3.4), from which it can be expected that in real
life the transverse shear capacity of ComFlor 210 is even bigger. The dependency of the transverse
shear capacity on this interface behaviour is clearly shown when comparing both FEA’s of ComFlor
210 in section 3.3 and 3.4. As explained in the beginning of section 3.1, it was intended to analyse the
exact same cross-section as in the research of Abspoel et al.[21]. However, in the FEA of the concrete
section (section 3.2), the rib reinforcement diameter had to be increased to make sure that bending
failure didn’t become decisive. This adjustment has been continued in the rest of the thesis, making
that considered cross-section in FEA of ComFlor 210 isn’t completely in accordance with the research
of Abspoel et al.[21]. However, due to this initial assumption, additional rebars were included in the top
layer of the finite element model (in accordance with Abspoel et al.[21]). It may be established that this
is quite unnecessary to do for sagging bending moment conditions and it is assumed that it hardly
affects the transverse shear capacity in this case.

75
76
5. CONCLUSION
The current Eurocode 4 (composite structures) suggests that the transverse shear capacity of a
composite slab solely depends on the resistance of its concrete section (i.e. concrete ribs). The
standard simply refers to the calculation procedures of the Eurocode 2 (concrete structures), so that
the transverse shear capacity of a composite slab is calculated according to this empirical formula which
was originally derived for regular reinforced concrete members. No information is provided about the
contribution of the steel deck to the transverse shear capacity of the composite slab, while the
Eurocode 3 (steel structures) implies that this steel deck can resist a certain transverse shear force in
its webs. Therefore, it is expected that the Eurocode 4 provides an unnecessarily conservative design
principle for the transverse shear capacity of composite slabs. This expectation has been investigated
for deep composite slabs in this thesis by means of non-linear finite element modelling. The finite
element study was started with merely analysing the concrete section of ComFlor 210, after which the
steel deck had been added to the finite element model to investigate its influence on the transverse
shear capacity. When evaluating the FEA results, the following can be concluded:
➢ The empirical formula provided by the Eurocode 2 gives an inaccurate prediction of the transverse
shear capacity of the concrete section (i.e. concrete ribs). Using the mean width of the concrete rib
(b0), instead of the minimum width in the tensile area (bw), has proved to give a better prediction of
the transverse shear capacity of this concrete section.
➢ The steel deck provides a minimum increase of 51.4% in transverse shear capacity for ComFlor
210, with respect to the resistance of its concrete section, because of the steel deck’s webs being
able to resist a part of the transverse shear force. When considering that the bonding at the interface
will activate the reinforcing properties of the steel deck as well, this increase in transverse shear
capacity is expected to become even bigger. This expectation is based on the observation in the
FEA that when the steel deck acts as reinforcement to the concrete in both longitudinal direction
(like rebars) and transverse direction (like stirrups), an increase of 131.6% in transverse shear
capacity is obtained.
➢ The Eurocode 4 provides a too conservative calculation method for transverse shear capacity of
ComFlor 210: a capacity of 16.54 kN can be calculated for the considered cross-section (one rib),
while a capacity of 65.27 kN is found in the FEA as a lower bound value.
➢ The model of Pereira et al.[17], which simply combines the procedures of the Eurocode 2 and 3,
gives a better prediction of the transverse shear capacity of ComFlor 210 (45.74 kN per rib), while
still being safe compared to the FEA results of this thesis. Please notice that with respect to the
Eurocode 2, the model does use the mean width of the concrete rib (b0) instead of the minimum
width (bw), as mentioned in the first concluding remark.

Recommendations
From the FEA results, it is concluded that the Eurocode 2 gives inaccurate predictions for the transverse
shear capacity of the concrete ribs. In trying to find a good calculation method for the total transverse
shear capacity of the composite slab, this inaccuracy of the Eurocode 2 already seems a critical aspect.
Therefore, it is recommended to start a second FEA program, as an extension to this thesis, which
solely focusses on the transverse shear capacity of the concrete ribs. Varying the model parameters
(like: compressive strength, rib reinforcement diameter, effective depth, etc.) should verify whether the
77
Eurocode 2 is inaccurate for every case and if the calculation procedure can be improved with the
proposal of using b0 instead of bw (see chapter 4). The FEA program may be supported by real
experiments.

Furthermore, more detailed information about the interface behaviour is necessary for the verification
of the transverse shear capacity of deep composite slabs by means of FEA. The new constitutive
relationships for the interface should include the following:
▪ The total slip restraining properties of the embossments: not only for the longitudinal direction (local
x-direction), but also for the transverse direction (local y-direction);
▪ The effect of opening of the interface (in local z-direction) on the slip restraining properties of
embossments (in local x- and y-directions).

At last, it may be established that the execution of real experiments is a must for verification of the
transverse shear capacity of ComFlor 210. In the FEA, a lot of assumptions have been made to come
to a finite element model, like the specification of the interface properties; the application of reduced
material properties for the embossments and neglecting their actual geometry; not considering the local
effects of cold forming, etc. These assumptions can only be validated with experiments.

78
Bibliography
[1] Şamhâl, E. (2005). Lecture 1.1: Composite Construction. Retrieved from
https://web.itu.edu.tr/~haluk/COMPOSITE%201.pdf
[2] BCSA, Steel for Life, & SCI. (n.d.). Composite construction. Retrieved from
https://www.steelconstruction.info/Composite_construction
[3] CEN. (2004). EN 1994-1-1, Eurocode 4: Design of composite steel and concrete structures - Part 1-1:
General rules and rules for buildings.
[4] Dutch Engineering Raadgevend Ingenieurs Bureau B.V. (2013). Staal-beton Vloersystemen. Retrieved
from
https://www.dutchengineering.nl/media/vk_1198/files/67579_DutchEngineering_Brochure_Online.pdf
[5] Toniolo, G. (2009). Composite Slabs with Lightweight Concrete. Retrieved from
https://www.politesi.polimi.it/bitstream/10589/2347/1/2010_07_Penza.pdf
[6] Hedaoo, N. A. and Gupta, L. M. (2008). State of the Art Report on Thin-walled Cold-formed Profiled
Steel Decking. In ORG. Missouri University of Science and Technology (Ed.), Nineteenth International
Specialty Conference on Cold-Formed Steel Structures (pp. 307–323).
[7] SFIA. (n.d.). How Cold-Formed Steel is Made. Retrieved from
https://sfia.memberclicks.net/how-cold-formed-steel-is-made
[8] Stark, J. W. B., & Abspoel, R. (2010). STEEL STRUCTURES 3, CT4121, Part: Cold Formed Sections.
Delft University of Technology.
[9] Rackham, J. W., Couchman, G. H., & Hicks, S. J. (2009). Composite Slabs and Beams using Steel
Decking: Best Practice for Design and Construction (Revised Edition). The Metal Cladding & Roofing
Manufacturers Association.
[10] Simms, W. I., & Hughes, A. F. (2011). Composite design of steel framed buildings, in accordance with
Eurocodes and the UK National Annexes. SCI.
[11] Leskelä, M. V. (2017). BEHAVIOUR AND DESIGN OF STEEL-CONCRETE COMPOSITE
STRUCTURES. Lahti: Peikko Group Corporation.
[12] Schuurman, R. (2001). Physical Behaviour of Shear Connections in Composite Slabs (doctoral thesis,
Delft University of Technology, The Netherlands). Retrieved from
http://resolver.tudelft.nl/uuid:75b035c6-bc9a-4d61-bbd4-8c82c8543572
[13] Abbas, H. S., Bakar, S. A., Ahmadi, M., & Haron, Z. (2015). Experimental studies on corrugated steel-
concrete composite slab. GRAĐEVINAR, 67(3), 225-233. https://doi.org/10.14256/JCE.1112.2014
[14] Diana FEA BV. (2019). DIANA Finite Element Analysis (10.3) [Software]. Retrieved from
https://dianafea.com/diana-downloads
[15] Johnson, R. P. (2004). Composite Structures of Steel and Concrete: Beams, Slabs, Columns, and
Frames for Buildings (4e ed.). Oxford, United Kingdom: Wiley.
[16] Van Erp, P. (2017). Horizontal Shear Resistance of ComFlor 210 (Master's thesis, Delft University of
Technology, The Netherlands). Retrieved from
http://resolver.tudelft.nl/uuid:613a4ebf-7d48-48f6-9c07-4a84fca028bc
[17] Pereira, M., Simões, R., & Duarte, J. (2017). ANALYSIS OF THE VERTICAL SHEAR DESIGN MODEL
FOR STEEL‐CONCRETE COMPOSITE SLABS. ce/papers, 1(4), 405-414.
https://doi.org/10.1002/cepa.540
[18] Stark Partners. (2012). VERTICAL SHEAR RESISTANCE OF COMPOSITE FLOORS WITH ComFlor
60 (EVALUATION OF TESTS REPORTED IN SPO/ICON-RT-11-10-VO1 (IMPERIAL COLLEGE)).
[19] Popo-Ola, S. O. (2012). Test Report on Shear Line Load Tests on Composite Slabs with CF60 Decking
(Report nr. SPO/ICON-RT-11-10-V01). London, United Kingdom: Imperial College.

79
[20] ABNT. (2008). ABNT NBR 8800, Projeto de estruturas de aço e de estruturas mistas de aço e
concreto de edifícios. Rio de Janeiro: Associação Brasileira de Normas Técnicas.
[21] Abspoel, R., Stark, J., & Prins, H-J. (2018). The influence of vertical shear on the hogging bending
moment resistance of ComFlor composite slabs. In V. Albero, A. Espinos, C. Ibabez, A. Lapuebla, & M.
L. Romero (Eds.), Proceedings 12th international conference on Advances in Steel-Concrete
Composite Structures - ASCCS 2018 (pp. 183-190). Valencia, Spain: Universitat Politecnica de
Valencia. https://doi.org/10.4995/ASCCS2018.2018.7056
[22] Tuls, J. J. (2017). Influence of a Vertical shear Force on the Hogging Bending Moment Resistance in
Composite Slabs (Master's thesis, Delft University of Technology, The Netherlands). Retrieved from
http://resolver.tudelft.nl/uuid:21663b9c-2840-485a-bc36-b039ad71ef5a
[23] Tata Steel. (2017). ComFlor® manual: composite floor decking design and technical information.
Retrieved from
https://www.tatasteelconstruction.com/static_files/Tata%20Steel/content/products/Building%20Systems/
Comflor/ComFlor%20manual.pdf
[24] CEN. (2004). EN 1992-1-1, Eurocode 2: Design of concrete structures - Part 1-1: General rules and
rules for buildings.
[25] CEN. (2006). EN 1993-1-3, Eurocode 3 - Design of steel structures - Part 1-3: General rules -
Supplementary rules for cold-formed members and sheeting
[26] CEN. (1992). ENV 1994-1-1, Eurocode 4: Ontwerp en berekening van staal-betonconstructies - Deel 1-
1: Algemene regels en regels voor gebouwen
[27] Hartmeyer, S., & Kurz, W. (2013). Evaluation of the Shear Force Behaviour of Composite Slabs. In
Bradford, M. & Uy, B. (Eds.), Composite Construction in Steel and Concrete VII - Proceedings of the
2013 International Conference on Composite Construction in Steel and Concrete (pp. 731-742).
Reston, Virginia: American Society of Civil Engineers. Retrieved from
https://app.knovel.com/hotlink/pdf/id:kt010XWGGC/composite-construction/evaluation-shear-force
[28] Yang, Y. (2014). Shear behaviour of reinforced concrete members without shear reinforcement: A new
look at an old problem (doctoral thesis, Delft University of Technology, The Netherlands).
https://doi.org/10.4233/uuid:ac776cf0-4412-4079-968f-9eacb67e8846
[29] Rombach, G., & Henze, L. (2017). Shear Capacity of Concrete Slabs Without Shear Reinforcement
Under Concentrated Loads Close to Support. In D. A. Hordijk, & M. Luković (Reds.), High Tech
Concrete: Where Technology and Engineering Meet (pp. 676–683).
https://doi.org/10.1007/978-3-319-59471-2_80
[30] CEN. (2005). EN 1993-1-1, Eurocode 3: Design of steel structures - Part 1-1: General rules and rules
for buildings.
[31] Stark, J. & Stark, R. (2009). Staal-Beton: toepassing en berekening van staal-beton constructies voor
gebouwen volgens Eurocode 4 bij normale temperatuur en brand. Delft: Bouwen Met Staal.
[32] Tata Steel. (n.d.). ComFlor Steel Composite Metal Deck Design Software (v9.0.34) [Software].
Retrieved from https://www.tatasteelconstruction.com/en_GB/Products/structural-buildings-and-
bridges/Composite-floor-deck/Comflor%C2%AE-Active
[33] Tuladhar, R. (2014). Shear failure of RC beam [photo]. Retrieved from https://i.ytimg.com/vi/zn3-
VM9Eurw/maxresdefault.jpg
[34] Veljković, M. (1996). BEHAVIOUR AND RESISTANCE OF COMPOSITE SLABS: Experiments and
Finite Element Analysis (doctoral thesis, Luleå University of Technology, Sweden).
[35] Ríos, J. D., Cifuentes, H., Martínez-De La Concha, A., & Medina-Reguera, F. (2017). Numerical
modelling of the shear-bond behaviour of composite slabs in four and six-point bending
tests. Engineering Structures, 133, 91–104. https://doi.org/10.1016/j.engstruct.2016.12.025
[36] Braam, C. R., Lagendijk, P., & Dees, W. C. (2011). Constructieleer Gewapend Beton. Boxtel, the
Netherlands: Æneas.
[37] Composite Profiles UK. (2018). comflor-decking [photo]. Retrieved from
https://www.compositeuk.com/services-solutions/comflor-metal-decking/
[38] CEN. (2015). EN 10346, Continuously hot-dip coated steel flat products for cold forming - Technical
delivery conditions.

80
Appendix A
Shear buckling strength of steel decks used in composite slabs

Due to shearing of a steel deck’s web, compressive and tensile stresses arise in the diagonal directions
(principal directions). This is illustrated in Figure A.1. These compressive stresses can lead to local
buckling, because of the thinness of the web. This phenomenon is known as shear buckling.

Figure A.1 – Shear buckling of a plate

For considering the effect of shear buckling on the transverse shear resistance of a steel deck, the
Eurocode 3 introduces a shear buckling strength fbv. The determination of this strength parameter is
prescribed in clause 6.1.5 of EN 1993-1-3[25]. All the following information is directly copied from the
mentioned standard.

81
 Definition of Is in clause 5.5.3.4.3(7):

More information on how to determine seff,1 can be found in EN 1993-1-3[25].

With the procedure as described above, it is possible to determine the shear buckling strength of
ComFlor 210. The required dimensions of the webs are shown in Figure A.2 and Figure A.3. These
figures correspond to the finite element model of this thesis.

82
 Figure A.2 - Dimensions of ComFlor 210 as applied in the FEA (unit: mm)

Figure A.3 - Stiffener (web, unit: mm)

The calculation of fbv for ComFlor 210 is shown below, with fy = 409 MPa and E = 210000 MPa.
210000
𝑠𝑒𝑓𝑓,1 = 0.76𝑡 √𝐸/(𝛾𝑀0 𝜎𝑐𝑜𝑚,𝐸𝑑 ) = 0.76 ∙ 0.96 ∙ √ = 16.5 𝑚𝑚
409
14.2∙0.96
𝐼𝑠 = 12
(14.22 cos2 66˚ + 0.962 sin2 66˚) + 2 ∙ 16.5 ∙ 0.96 ∙ 2.92 = 305.19 𝑚𝑚 4
1
2.10 305.19 3
𝑘𝜏 = 5.34 + 0.96 (31.7+14.2+173.6) = 7.78
31.7+14.2+173.6 5.34 409 173.6 409
𝜆𝑤 = 0.346 ∙ 0.96
∙ √7.78 ∙ 210000 = 2.89 > 𝜆𝑤 = 0.346 ∙ 0.96
∙ √210000 = 2.76
409
𝑓𝑏𝑣 = 0.67 ∙ = 32.8 𝑀𝑃𝑎
2.892

83
Appendix B
Calculation of the overhang Lo as used in the FEA

Figure B.1 – Test setup of the FEA (side view)

It has been chosen to include an overhang length L0 which is bigger than lbd + d. The value for lbd can
be determined by using clause 8.4.3 and 8.4.4 of EN 1992-1-1[24]:

𝑙𝑏𝑑 = 𝛼1 𝛼2 𝛼3 𝛼4 𝛼5 𝑙𝑏,𝑟𝑞𝑑 ≥ 𝑙𝑏,𝑚𝑖𝑛 B.1

Ø 𝜎𝑠𝑑
𝑙𝑏,𝑟𝑞𝑑 = B.2
4 𝑓𝑏𝑑
𝑙𝑏,𝑚𝑖𝑛 = 𝑚𝑎𝑥{0.3𝑙𝑏,𝑟𝑞𝑑 ; 10Ø; 100𝑚𝑚} B.3

From equation B.2 it can be seen that lb,rqd is dependent on the stress level in the reinforcement σsd at
the section from where the anchorage length is considered. If we consider the anchorage length from
the support, it may be established that σsd is equal to zero, since the bending moment and normal force
are also equal to zero at this location. This makes that lb,rqd is also equal to zero, meaning that the value
for lbd (equation B.1) is equal to lb,min (equation B.3). Using a diameter of 12 mm gives lb,min = 120 mm
and lbd ≥ 120 mm. Consequently, the summation of lbd and d is equal to 120 + (280 – 40 – 6) = 354 mm.
Therefore, a value of 355.0 mm has been applied for L0. This value is kept constant throughout the
whole thesis.

84
Appendix C
Determination of the bond-slip behaviour for the FEA of ComFlor 210

As explained at the end of section 3.3.1, the bond-slip behaviour of the interface in longitudinal direction
contains 3 different linear branches. The diagram as applied in the FEA is given in Figure C.1.
Derivation of the corresponding values of the diagram is given below.

Figure C.1 - Bond-slip behaviour for local x-direction of interface elements

In section 2.4, it has been mentioned that Van Erp [16] found an average value of 0.103 MPa for the
bond strength of the interface based on his experiments on ComFlor 210. Although this thesis doesn’t
follow his recommendation on using an effectiveness of 12.5% for the embossments (thickness
reduction), this ultimate strength value of 0.103 MPa for the interface has still been used. Figure C.2
shows 3 typical load-displacement curves of experiments by Van Erp[16], with corresponding slip
diagrams in Figure C.3. From these slip diagrams, it can be estimated that failure occurs at a slip value
of approx. 4 mm. This value has been determined by averaging the slip that occurs from starting point
to failure (most right vertical red lines in Figure C.3), for all measure points in every specimen. From
the load-displacement curves in Figure C.2, it can be established that large non-linear behaviour starts
at approx. 75% of the ultimate failure load. The final value of the first linear branch in Figure C.1 is
therefore estimated at 0.75∙0.103 = 0.077 MPa. The corresponding slip value is again determined from
the diagrams in Figure C.3: the average value of slip, from starting point up till the most left red vertical
lines, is estimated at 0.5 mm. The descending part of Figure C.1 is the least important for the FEA,
since it is not demanded to simulate longitudinal shear failure. The recommendations by Ríos et al.[35]
have been followed to describe this part of the bond-slip diagram.

85
 Figure C.2 – Typical load-displacement curves of experiments by Van Erp[16]

SPCM 1

SPCM2

SPCM 3
Figure C.3 – Slip diagrams for the 3 different specimens (6 measure points per specimen)[16]

86
Appendix D
Transverse shear capacity of composite slabs according Eurocode 4 and literature

In this appendix, the calculations for the transverse shear capacity according to 4 different models will
be given: Eurocode 4, Stark[18], Pereira et al.[17] and Hartmeyer & Kurz[27] (see chapter 2). Most
calculations have been done in an Excel spreadsheet.

Eurocode 4

The transverse shear resistance according to the Eurocode 4 is equal to 16.54 kN per concrete rib.

Stark

87
 Figure D.1 – Parameters hL and α for ComFlor 210

The transverse shear resistance according to Stark [18] is equal to 123.55 kN per concrete rib of the
composite slab. The values for hL and α are shown in Figure D.1. The original yield strength of the steel
plate (409 MPa) has been used in calculation.

Pereira et al.

The transverse shear resistance according to Pereira et al.[17] is equal to 45.74 kN for one rib.

88
Hartmeyer & Kurz
In the method of Hartmeyer & Kurz[27], it is mentioned that eq. 2.14 may be used to calculate the height
of the compressive zone in the concrete. Instead of this, the height of this compressive zone has been
estimated from the FEA. In Figure D.2, the normal stress in the concrete is plotted over the height of
the composite slab. The compressive zone at the top is relatively small and its height is approx. equal
to 20 mm. The given distribution of normal stresses is taken from the analysis of ComFlor 210 as
discussed in section 3.4. The calculation of the transverse shear capacity of the composite slab is
shown below. For the partial resistance of the steel deck Vp,Rm, a value of 44.98 N/mm has been used
as mentioned in the report of Tuls[22]. This value is based on real tests.
2 2
𝑉𝑡,0,𝑅𝑚 = 𝑉𝑝,𝑅𝑚 + ∙ 𝑥𝑚 ∙ 𝑏𝑐 ∙ 𝑓𝑐𝑡𝑚 = 44.98 ∙ 600 + ∙ 20 ∙ 600 ∙ 3.1 = 26988 + 24800 = 51788 𝑁
3 3
𝑉𝑡,0,𝑅𝑚 = 51.79 𝑘𝑁

Figure D.2 – Distribution of normal stress in the concrete over the height of the slab at maximum load
(in the middle of the rib at the location of the applied load Vt)

89