Five Spring Model

Two-Parameter Kinematic Approach for the Shear
Behavior of Deep Beams Made of Fiber-Reinforced

Concrete
Author: Karolína Tvrzníková
Supervisor: Prof. Boyan Mihaylov
University: University of Liège
University: University of Liège
Date: 9.1.2017
European Erasmus Mundus Master
Sustainable Constructions under natural hazards and catastrophic events
520121-1-2011-1-CZ-ERA MUNDUS-EMMC
ABSTRACT
Deep concrete beams are characterized by small shear-span-to-depth ratios and high shear
resistance. Owing to their high strength, they are used as transfer girders in buildings, cap beams
in bridges, and pile caps in foundations. It is also characteristic of deep beams that they develop
complex deformation patterns and cannot be modelled based on the plane-sections-remain-
plane hypothesis. This thesis focuses on modelling the complex shear behavior of fiber-
reinforced concrete (FRC) deep beams. While deep beams are typically reinforced only with
steel bars in the form of flexural and shear reinforcement, experimental studies have shown that
the addition of steel fibers in the concrete can enhance their shear behavior.
The main aim of this thesis is to study a five-spring model for deep beams with conventional
reinforcement proposed by Mihaylov et al. (2015), and to extend this model to deep beams with
FRC. The five-spring model uses only two kinematic parameters to describe the deformations
in deep beams. The extended model captures the complete load-displacement response of FRC
beams by accounting for three effects associated with the steel fibers: 1) tension in the fibers
crossing the shear cracks; 2) enhanced ductility of the critical compressed zones in deep beams;
and 3) tension stiffening effect on the flexural reinforcement. To account for these three local
effects, existing models from the literature are studied, compared, and validated. Each of the
models is implemented in a Matlab code and is validated with relevant material tests. It is shown
that the most suitable models for the modelling of the three effects were proposed by Lee et al.
(2013), Ou et al. (2012) and Lee et al. (2013). Once these models were validated, they were
implemented in the global framework provided by the five-spring model for deep beams.
The extended five-spring model is validated against a database of tests of FRC deep beams
collected from the literature. It is shown that the predicted shear strengths are in good agreement
with the measured values. The validated model is then used to perform a parametric study
focused on the effects of the shear-span-to-depth (a/d) ratio, shear and longitudinal
reinforcement ratios, as well as fiber volumetric ratio on the shear behavior of deep beams.
Increased shear resistance was observed for increasing the shear and longitudinal
reinforcements, as well as increasing the fiber volumetric ratio. By increasing a/d ratio, the
shear strength decreased. At last, the effectiveness of shear reinforcement was compared with
the fiber reinforcement for different a/d ratios. It is concluded that the fiber reinforcement is
more effective only for a/d ratios lower than 0.8, while the shear reinforcement is more effective
for higher a/d ratios.
1
ACKNOWLEDGEMENTS
Firstly, I would like to thank all the members of the jury and everyone who will take time to
read this thesis.
Most of all, I would like to thank Professor Boyan Mihaylov, for his great guidance and help
provided throughout the development of this thesis. All the time he devoted to our consultations,
his patience and knowledge he shared with me throughout the semester are the main reasons
behind the completion of this thesis. Thank you for pushing me to work harder and question
myself. Working with him was a truly rewarding experience.
Special thanks to my office colleagues Jian Liu, Nikola Tatar and Renaud Franssen. They would
always answer my questions and provide help anytime I asked. It was a pleasure to work in
such a friendly environment.
I would also like to thank the whole SUSCOS team. All the coordinators and professors from
all the partner universities and in particular Professor František Wald, for all the effort and
dedication which makes SUSCOS such a unique learning experience. Special thanks also to my
classmates that have made this past year and half really great.
At last, I am thankful to my family, Abram and friends for their support no matter the distance
that separated us.
2
TABLE OF CONTENTS
Abstract ...................................................................................................................................... 1
Acknowledgements .................................................................................................................... 2
Table of Contents ....................................................................................................................... 3
List of Figures ............................................................................................................................ 5
List of Tables .............................................................................................................................. 8
List of Annotations ..................................................................................................................... 9
1 Introduction ....................................................................................................................... 14
1.1 General....................................................................................................................... 14
1.2 Scope and Objectives of the Thesis ........................................................................... 15
1.3 Thesis Outline ............................................................................................................ 16
2 Background ....................................................................................................................... 17
2.1 Fiber-Reinforced Concrete (FRC) ............................................................................. 17
2.1.1 Properties and Classification of FRC ................................................................. 17
2.1.2 Factors Affecting Behavior of FRC ................................................................... 19
2.2 Tests on FRC Deep Beams ........................................................................................ 20
2.2.1 Mansur and Ong (1991) ..................................................................................... 20
2.2.2 Mansur and Alwist (1984).................................................................................. 22
2.3 Two-Parameter Kinematic Theory and Five-Spring Model for Deep Beams ........... 24
3 Extended Five-Spring Model for FRC Deep Beams ........................................................ 33
3.1 Introduction ............................................................................................................... 33
3.2 Tension in Fibers across the Critical Diagonal crack ................................................ 33
3.2.1 Introduction ........................................................................................................ 33
3.2.2 Formulation of the SDEM .................................................................................. 35
3.2.3 Validation of the SDEM ..................................................................................... 38
3.2.4 Implementation of the SDEM into the Five-Spring Model ................................ 41
3.3 Tension-Stiffening of the Bottom Flexural Reinforcement ....................................... 45
3.3.1 Introduction ........................................................................................................ 45
3.3.2 Formulation of Tension-Stiffening Model for R/FRC ....................................... 46
3.3.3 Validation of the Tension-Stiffening Model for R/FRC .................................... 48
3.3.4 Implementation of the Tension-Stiffening Model into the Five-Spring Model . 51
3.4 Compressive Behavior of the Critical Loading Zone ................................................ 53
3
3.4.1 Introduction ........................................................................................................ 53

3.4.2 Stress-Strain Curves for FRC ............................................................................. 54
3.4.3 Implementation of the Stress-Strain Curve into the Five-Spring Model............ 61
3.5 Deflections ................................................................................................................. 64
3.5.1 Introduction ........................................................................................................ 64
3.5.2 Definition of the Additional Deflection ............................................................. 64
3.5.3 Implementation of the Additional Deflection into the Five-Spring Model ........ 65
4 Validation of the Extended Five-Spring Model ................................................................ 66
4.1 Introduction ............................................................................................................... 66
4.2 Sample Specimen Evaluation .................................................................................... 66
4.3 Effect of Fiber Volumetric Ratio ............................................................................... 70
4.4 Effect of a/d Ratio ..................................................................................................... 70
4.5 Discussion .................................................................................................................. 72
5 Parametric Study ............................................................................................................... 75
5.1 Introduction ............................................................................................................... 75
5.2 Parametric Study........................................................................................................ 75
5.2.1 a/d Ratio ............................................................................................................. 75
5.2.2 Longitudinal Reinforcement .............................................................................. 77
5.2.3 Fiber Volumetric Ratio....................................................................................... 79
5.2.4 Shear Reinforcement .......................................................................................... 80
5.2.5 Comparison of Effectiveness of Fibers vs. Shear Reinforcement ...................... 83
5.2.6 Size Effect .......................................................................................................... 85
5.3 Discussion .................................................................................................................. 87
6 Summary and Conclusions ............................................................................................... 89
References ................................................................................................................................ 91
Appendix A: Test Database ...................................................................................................... 94
Appendix B: SDEM - Matlab Code ......................................................................................... 95
Appendix C: Tension-Stiffening - Matlab Code ...................................................................... 96
Appendix D: Average Diagonal Stress in the Critical Loading Zone - Matlab Code .............. 97
Appendix E: Extended Five-Spring Model - Matlab Code ...................................................... 98
4
LIST OF FIGURES
Figure 1.1 Application of deep beam in a building (Mihaylov et al.,2013) ............................. 14
Figure 2.1 Softening (a) and hardening (b) post-cracking behavior of SFRC - adapted from fib
MC2010 (2013) ................................................................................................... 18
Figure 2.2 Typical load-CMOD curve from fib MC2010 (2013) ............................................ 18
Figure 2.3 Experimental setup by Mansur and Ong (1991) ..................................................... 21
Figure 2.4 Experimental results by Mansur and Ong (1991) ................................................... 21
Figure 2.5 Cracking pattern of the beam B4 by Mansur and Ong (1991) ................................ 22
Figure 2.6 Experimental setup by Mansur and Alwist (1984) ................................................. 23
Figure 2.7 Failure mode of control specimen by Mansur and Alwist (1984) .......................... 23
Figure 2.8 Main assumptions of kinematic model (Mihaylov et al.,2013) .............................. 24
Figure 2.9 Degrees of freedom of kinematic model (Mihaylov et al.,2013) ............................ 25
Figure 2.10 Variation of the angle α1 during transition from deep to slender beams (Mihaylov
et al.,2013)........................................................................................................... 25
Figure 2.11 Five-spring model (Mihaylov, 2015) .................................................................... 27
Figure 2.12 Free-body diagram of the rigid block (Mihaylov, 2015) ...................................... 27
Figure 2.13 Critical loading zone (Mihaylov, 2015) ................................................................ 28
Figure 2.14 Shear forces in five-spring model at 5 mm imposed distortion of the CLZ ......... 31
Figure 2.15 Equilibrium of forces in five-spring model for imposed displacement Δ ............ 32
Figure 3.1 Tensile behavior of FRC (Lee et al., 2013) ............................................................ 34
Figure 3.2 Illustration of additional force provided by fibers in the extended five-spring model
............................................................................................................................. 35
Figure 3.3 Fiber types used in the study by Lim et al. (1987) ................................................. 35
Figure 3.4 Comparison of proposed model and test results for members with straight fibers
tested by Lim et al. (1987) .................................................................................. 39
Figure 3.5 Comparison of proposed model and test results for members with end-hooked fibers
tested by Lim et al. (1987) .................................................................................. 39
Figure 3.6 Comparison of proposed model and test results for members with end-hooked fibers
tested by Susetyo (2009) ..................................................................................... 40
Figure 3.7 Definition of angle β ............................................................................................... 41
Figure 3.8 Variation of the angle β given different a/d ratios .................................................. 42
5
Figure 3.9 Variation of the angle β given different fiber volume ratios .................................. 42
Figure 3.10 Equilibrium of forces in extended five-spring model with fiber contribution at Δc=5
mm ...................................................................................................................... 43
Figure 3.11 Predicted V-Δ response with addition of fibers .................................................... 44
Figure 3.12 Tensile behavior of RC and R/FRC members by Lee et al. (2013) ...................... 45
Figure 3.13 Illustration of implemented modification due to tension-stiffening effect ........... 46
Figure 3.14 Comparison of proposed models and test results by Bischoff (2003) .................. 49
Figure 3.15 Comparison of proposed models and test results by Deluce and Vecchio (2013) 50
Figure 3.16 Separate evaluation of contribution of tensile forces in Tension-stiffening model
............................................................................................................................. 51
Figure 3.17 Equilibrium of forces in extended five-spring model with tension-stiffening model
at Δc=5 mm.......................................................................................................... 52
Figure 3.18 Predicted V-Δ response with implementation of tension-stiffening model .......... 52
Figure 3.19 Illustration of modification due to change of stress-strain curve of concrete in
compression ........................................................................................................ 53
Figure 3.20 Analytical proposal of stress-strain compression curve by Barros and Figueiras
(1999) .................................................................................................................. 56
Figure 3.21 Analytical proposal of stress-strain compression curve by Lee et al. (2015) ....... 57
Figure 3.22 Analytical proposal of stress-strain compression curve by Ou et al. (2012) ........ 59
Figure 3.23 Analytical proposal of stress-strain compression for high-strength concrete by
Mansur et al. (1999) ............................................................................................ 60
Figure 3.24 Equilibrium of forces in extended 5sm with adjusted stress-strain curve at Δc=5 mm
............................................................................................................................. 62
Figure 3.25 Predicted V-Δ response with modified stress-strain curve ................................... 63
Figure 4.1 Cracking pattern and mode of failure of beam B4 by Mansur and Ong (1991) ..... 67
Figure 4.2 Comparison of load-deflection curves for beam B4 obtained from five-spring method
and experimental results by Mansur and Ong (1991) ......................................... 68
Figure 4.3 Comparison of load-deflection curves for beam B4 obtained from finite element
model, five-spring method and experimental results by Mansur and Ong (1991)
............................................................................................................................. 69
Figure 4.4 Cracking pattern of the beam B4 obtained by the FE analysis ............................... 69
Figure 4.5 Effect of fiber volume ratio on ultimate shear strength .......................................... 70
Figure 4.6 Effect a/d ratio on ultimate shear strength .............................................................. 71
6
Figure 4.7 Comparison of results obtained by FEM and five-spring model for different fiber
volumetric ratios ................................................................................................. 73
Figure 5.1 Effect of a/d ratio for different fiber volume ratios ................................................ 76
Figure 5.2 Breakdown of effect of a/d ratio for Vf = 1% ......................................................... 76
Figure 5.3 Effect of longitudinal reinforcement for different fiber volume ratios ................... 77
Figure 5.4 Breakdown of effect of longitudinal reinforcement for Vf = 1% ........................... 78
Figure 5.5 Effect of fiber volumetric ratio for different shear reinforcement ratios ................ 79
Figure 5.6 Breakdown of effect of fiber volumetric ratio for ρv = 0.5% ................................. 80
Figure 5.7 Effect of shear reinforcement for different fiber volume ratios .............................. 81
Figure 5.8 Breakdown of effect of shear reinforcement for Vf = 1% ...................................... 81
Figure 5.9 Results for B10 obtained with the extended five-spring model .............................. 82
Figure 5.10 Average tensile strain at deflection equal to zero for different shear reinforcement
ratios .................................................................................................................... 83
Figure 5.11 Results for B10 obtained with the extended five-spring model with completed curve
............................................................................................................................. 83
Figure 5.12 Variation of Vmax for different a/d ratios with ρv = 0.2% ..................................... 84
Figure 5.13 Fiber volume ratio equivalent to 0.2% shear reinforcement ratio ........................ 84
Figure 5.14 Size effect for different fiber volume ratios .......................................................... 85
Figure 5.15 Size effect breakdown for 0% fiber volume ratio ................................................. 86
Figure 5.16 Size effect breakdown for 2% fiber volume ratio ................................................. 87
7
LIST OF TABLES
Table 2.1 Sample beam properties ........................................................................................... 30
Table 3.1 Pullout strength of single fiber by Voo and Foster (2003) ....................................... 37
Table 4.1 Standard deviation and coefficient of variation of results obtained with extended five-
spring model .......................................................................................................... 74
Table 5.1 Properties of a sample beam for parametric study ................................................... 75
8
LIST OF ANNOTATIONS
⌀h = Diameter of web reinforcement

⌀l = Diameter of bottom longitudinal reinforcement bars
⌀v = Diameter of stirrups
2PKT = Two-parameter kinematic theory
5sm = Five-spring model
a = Shear span
a/d = Shear-span-to-depth ratio
ac1 = Distance between facing edges of loading and support plate
aeff = Effective shear span
ag = Maximum size of coarse aggregate
As = Area of longitudinal bars on flexural tension side
A𝑐, 𝑒𝑓𝑓 = Area of concrete providing tension stiffening for bottom reinforcement
b = Width of cross section
c = Coefficient for tension-softening effect of concrete or mortar
C = Compression force in section with maximum moment
cf = Coefficient to consider effect of steel fibers on tension-stiffening
CLZ = Critical loading zone
CMOD = Crack mouth opening displacement
CTOD = Crack top opening displacement
d = Effective depth of section
db = Diameter of bottom longitudinal bars
dbv = Stirrups bar diameter
df = Fiber diameter
DOF = Degree of freedom
Ec = Elastic modulus of concrete
Es = Elastic modulus of steel
Ev = Elastic modulus of stirrups
f ’c = Concrete cylinder strength
9
fc = Local stress of concrete

fc, TS = Tensile stress due to tension-stiffening
fc.soft = Tensile stress in concrete due to tension-softening
fcon = Tensile stress in concrete
Fcon = Tensile force of concrete
fcr = Cracking strength of concrete of mortar matrix
fct = Tensile stress by concrete/mortar tension-softening at given crack width
= Tensile stress due to mechanical anchorage effect of end-hooked steel
feh
fibers
FEM = Finite Element Model
Ff = Force provided by fibers
ff = Tensile stress at crack due to steel fibers
ffib = Tensile stress in fibers
Ffib = Tensile force of fibers
fR1k = Stress at serviceability limit state
fR3k = Stress at ultimate limit state
FRC = Fiber-reinforced concrete
Freb = Tensile force of reinforcing bar
freb = Tensile stress in reinforcing bar
fSFRC = Total tensile stresses of SFRC member
fst = Tensile stress due to frictional bond behavior of steel fibers
fy = Yield strength of bottom longitudinal bars
fyv = Yield strength of stirrups
h = Height of member
Icr = Moment of inertia of cracked section
Ig = Moment of inertia of un-cracked section
k = Factor influencing post-peak decay in stress
l0 = Length of heavily cracked zone at bottom of critical diagonal crack
lb1 = Width of loading plate parallel to longitudinal axis of member
lb1e = Effective width of loading plate parallel to longitudinal axis of member
10
lb2 = Width of support plate parallel to longitudinal axis of member

lf = Fiber length
li = Distance between mechanical anchorages for end-hooked fibers
lk = Length of dowels provided by bottom longitudinal reinforcement
ls = Distance between loading points
lt = Length of bottom reinforcement within the cracked part of shear span
M = Bond parameter
n = Curve fitting factor
nb = Number of bottom longitudinal bars
P = Applied concentrated load
Pcr = Applied load at moment of formation of crack
RIv = Fiber reinforcing index based on fiber volumetric ratio

RIw = Fiber reinforcing index based on fiber weight ratio
s = Slip and displacement in critical diagonal crack
sb = Maximum spacing between reinforcing bars
= Distance between the radial cracks along bottom longitudinal
scr
reinforcement
SDEM = Simplified diverse embedment method
= Slip at maximum tensile force due to mechanical anchorage of an end-
seh
hooked fiber
sf = Slip at frictional bond strength for straight fiber
SFRC = Steel fiber reinforced concrete
T = Tensile force in bottom reinforcement
Tf = Tensile force in bottom reinforcement including action of fibers
V = Shear force
Vci = Shear resisted by aggregate interlock
VCLZ = Shear carried by critical loading zone
Vd = Shear resisted by dowel action
Vexp = Measured shear strength
VF = Shear resisted by fibers
Vf = Fiber volumetric ratio
11
vf = Vertical component of stress resisted by fibers

Vpred = Predicted shear strength
Vs = Shear resisted by stirrups
w = Width of critical diagonal crack halfway along the crack
wcr = Average crack width
wcr.T = Width of the crack applied in SDEM to define tensile stress in fibers
Wf = Fiber weight percentage in the mixture
= Angle of the line extending from inner edge of support plate to far edge of
α
tributary area of loading plate responsible for shear force V
α1 = Angle of critical diagonal crack
αf = Fiber orientation factor
= Angle between the force provided by the fibers Ff and the vertical
β
component of the force VF
= Coefficient to consider effect of fiber slip on longer embedment side on
βeh
frictional behavior
= Coefficient to consider effect of fiber slip on longer embedment side on
βf
mechanical anchorage effect
Δ = Midspan deflection of beam
Δ1 = Additional deflection to account for four-point loading
Δc = Transverse displacement of critical loading zone
= Deflection of shear span due to elongation of bottom longitudinal
Δt
reinforcement
δx = Displacement along x axis
δy = Displacement along z axis
ε = Diagonal compressive strains in critical loading zone
ε’c = Strain at peak strength
εc0 = Strain at peak for plain concrete
εcf = Strain at peak for FRC
εmax = Maximum diagonal compressive strain in critical loading zone
εt, avg = Average strain in the bottom longitudinal reinforcement
= Strain in transverse reinforcement in critical diagonal crack halfway along
εv
the crack
θ = Angle of diagonal cracks in uniform stress field
12
ρh = Ratio of longitudinal web reinforcement

ρl = Ratio of bottom longitudinal reinforcement
ρv = Ratio of transverse reinforcement
σ = Diagonal compressive stress in critical loading zone
σavg = Average diagonal stress in critical loading zone
σv = Stress in transverse reinforcement
τeh, max = Pullout strength due to mechanical anchorage of hooked-end fiber
τf, max = Frictional bond strength
υci = Aggregate interlock shear stress
Ω = Area under stress-strain curve
13
1 INTRODUCTION
1.1 GENERAL
Deep beams are characterized by relatively small shear-span-to-depth (a/d) ratios. Maximum
value of ratio for deep beams is around 2.5. Because of their proportions, they develop
mechanisms of load resistance that are quite different from those in slender beams, and their
strengths are likely to be governed by shear rather than flexure (Mansur and Ong, 1991). Thanks
to their high shear strengths deep beams are used for example as transfer girders in buildings,
where they carry heavy loads from discontinuous columns, as shown in Figure 1.1. Structural
safety of a building with deep beam highly depends on the resistance of the girder as its failure
could result in a partial or complete collapse of the building. When subjected to vertical ground
accelerations due to strong earthquakes, the ability of structure to redistribute forces from
damaged girders to other structural members is crucial. The extent of such force redistribution
depends on displacement capacity and post-peak behavior of the transfer beam (Mihaylov,
2015).
slender
beam
transfer
girder
deep beam
footing
Figure 1.1 Application of deep beam in a building (Mihaylov et al.,2013)
The load-deformation response of slender beams can be modelled based on the hypothesis that
“plane sections remain plane” first proposed by Robert Hook in 1678. However, for deep beams
such as transfer girders, plane sections do not remain plane and shear strains become dominant.
The deformation patterns of such beams become more complex and a different modelling
approach is required.
In deep beams, a significant portion of the shear is carried by strut action, where compressive
stresses flow directly from the load to the support (Mihaylov et al.,2013). Due to this, strut-and-
tie models are recommended when designing deep beams. However, strut-and-tie method is not
always capable of predicting the shear behavior of deep beams due to the large number of
parameters that influence the shear behavior. A model that takes into account the number of
parameters that influence the shear behavior has been proposed by Mihaylov et al. (2013). The
model uses equations with two degrees of freedom and predicts the shear behavior of deep
beams through equilibrium, compatibility and stress-strain relationships.
14
A two-parameter kinematic theory (2PKT) for deep beams has been introduced by Mihaylov et
al. (2013) and is capable of describing the deformed shape of deep beams and the ultimate shear
strength of such members. Later the model has been extended to non-linear five-spring model,
which is able to predict the complete load-displacement response of shear-critical deep beams.
Four springs of the model represent the shear resistance mechanisms of the beam, and the fifth
spring represents the flexural behavior. The five-spring model has shown to predict the post-
peak shear behavior effectively, which is important for the analysis of structures under extreme
loading.
To enhance the post-peak capacity and displacement capacity, the addition of fibers to concrete
matrix has shown to significantly improve brittle behavior of concrete. The post-peak behavior
of fiber reinforced beams improves compared to regular concrete and provides higher ductility
of the members. Some of the observed enhancements include higher tensile strength of
concrete, increased post-cracking ductility, higher toughness and reduced crack width and crack
spacing. Fibers are produced from materials such as steel, carbon, glass, plastic or cotton. For
the reinforcement of deep beams used within this study, steel fibers were considered. Steel
fibers can be found in various profiles including straight, hooked-end, crimped and flattened-
end. The effectiveness of steel fibers in improving the behavior depends on several factors like
fiber volume ratio, fiber length and fiber aspect ratio.
1.2 SCOPE AND OBJECTIVES OF THE THESIS

This thesis is a continuation of work developed by Mihaylov et al. (2013). His two-parameter
kinematic theory for shear behavior of deep beams is capable of describing the shear strength
and deformation patterns for deep beams in shear failure. The kinematic theory as proposed by
Mihaylov et al. (2013) is applicable to reinforced concrete deep beams with possible transverse
reinforcement. Later, an extension of 2PKT allowed a complete prediction of pre- and post-
peak shear behavior of deep beams. Both of the methods showed to predict the results
effectively when compared to experimental results.
As it has been previously researched by Narayan and Darwish (1989) and Mansur and Ong
(1991), inclusion of fibers in concrete can increase the ductility and displacement capacity of
deep beams. The effect of fibers was, however, not considered in the five-spring model
developed by Mihaylov (2015). The main objective of this thesis is to extend the five-spring
model, to be able to capture the behavior of reinforced concrete deep beams with inclusion of
steel fibers.
First, the original five-spring model is analyzed. After a thorough analysis, changes caused by
the fibers, such as additional tensile forces from fibers bridging the cracks, are evaluated.
Appropriate models are proposed based on previous research and validated against
experimental data. Such models are then implemented into the original five-spring model to
create an extended five-spring model that captures the behavior of fiber reinforced concrete
deep beams. Proposed extended five-spring model is then validated against experimental
studies done by other researchers, and the accuracy of the model is evaluated. Additionally, a
15
parametric study is performed to study the effects of vertical and longitudinal reinforcement,
fiber volume ratio, size effect and a/d ratio on the shear strength of the beam.
1.3 THESIS OUTLINE

This thesis consists of six chapters, including Chapter 1 with some general information and
scope of the thesis.
Chapter 2 provides background information about topics relevant for the thesis. It includes a
brief overview of state-of-arts of design with fibers, including an overview of code provisions.
Selected experimental studies from other researchers are described, including the properties of
the specimen, experimental setup and obtained results with observations. It also consists of
description of two-parameter kinematic theory and five-spring model by Mihaylov (2015). The
theory behind the model is briefly explained, and the original model that served as a starting
point for the thesis is introduced.
Chapter 3 describes all the modifications introduced into the original model in order to capture
the behavior with inclusion of steel fibers. First, simplified diverse embedment model to model
the tension resisted by the fibers bridging the crack is introduced. Then, the created Matlab code
is validated against experimental and analytical results, and the version of the code as it enters
the five-spring model is explained. Secondly, tension-stiffening model is introduced. A Matlab
code is then created and a code to model the tension force provided by bottom longitudinal
reinforcement of the beam is validated against experimental and analytical results. The obtained
Matlab code is then explained as it enters the five-spring model. Later, the stress-strain behavior
of fiber-reinforced concrete under uniaxial compression is evaluated. Proposals by several
researchers are introduced and the most suitable ones are selected to enter into the extended
five-spring model. At last, additional formula to account for additional deflections in the mid-
span of deep beams under four-point loading is introduced.
Chapter 4 presents a validation of the proposed extended five-spring model. A sample specimen
is described and experimental results are compared with results obtained using extended five-
spring model and finite element model results. The effect of fiber volume ratio is analyzed and
a comparison between experimental results, analytical results and numerical results is provided.
Also the effect of a/d ratio is evaluated and experimental results are compared with analytical
ones. At last, results of other load-deflection curves based on other experimental results are
provided. Discussion is carried out to summarize the effectiveness of the model, and an issue
encountered during the validation process for deep beams with high shear reinforcement is
addressed and explained.
Chapter 5 consists of a parametric study. The effect of a/d ratio is analyzed, as well as the effect
of bottom longitudinal reinforcement. Fiber volume ratio and its influence on shear strength is
analyzed along with the effect of shear reinforcement and size effect. Finally, the amount of
fibers or shear reinforcement that is needed to obtain a given shear strength is calculated and
the effectiveness of each type of reinforcement is analyzed.
Chapter 6 contains a summary with conclusions and limitations of the proposed model.
16
2 BACKGROUND
2.1 FIBER-REINFORCED CONCRETE (FRC)

Fiber-reinforced concrete is a composite material consisting of cement matrix and discrete
fibers. The material of the fibers can vary from steel, polymers and carbon to glass or natural
materials. The main difference between regular concrete and FRC is an enhanced post-cracking
residual tensile resistance due to fibers capacity to bridge the cracks. Research regarding FRC
has started in the USA almost 50 years ago, however the research focused on structural response
of FRC has mainly developed only over last 15 years as described by Di Prisco et al. (2013).
As a consequence, international building codes for structural design are lacking guidelines for
design with FRC and due to the lack of guidelines practitioners are generally unwilling to use
FRC just based on voluntary guidelines or research papers.
2.1.1 Properties and Classification of FRC

In the fib Model Code 2010, FRC is recognized as a new material for structures (Di Prisco et
al., 2013). The code is divided in two sections, one devoted to material behavior and the other
to structural behavior. It mainly addresses steel fibers, however can be considered for other
types of fibers with limitations. The code also introduces classification of FRC to standardize
performance-based production.
Unless high volume of fibers is used, the addition of fibers does not significantly affect elastic
properties of concrete or compressive strength. However, for fiber volumes higher than 1% the
ascending branch of compressive stress-strain curve changes depending on the amount of fibers
and with increasing fiber volume ratio becomes less steep. Such results suggest higher ductility
and toughness of FRC and have been previously investigated. However, fib Model Code 2010
does not describe the compressive behavior of FRC unless experimental results are available,
and so the most recent research and analytical approach to describe the stress-strain relationship
is discussed later in this thesis.
Behavior in tension is the most important aspect of FRC. The fibers are able to bridge the crack
and transmit stresses across the crack. With increasing fiber content, the post-peak behavior is
characterized by increased toughness, becoming ductile for very high fiber contents as
described by Di Prisco et al. (2013). Depending on the orientation of the member and fibers,
steel fiber reinforced concrete (SFRC) members in uniaxial tension can have softening or
hardening post-cracking behavior as shown in Figure 2.1. Susetyo (2009) explains that tension
hardening effects occur in concrete containing higher fiber volume content than the critical fiber
volume content. The critical fiber volume is the volume of fibers which after matrix cracking
will carry the load which the composite maintained before cracking (Newman and Choo, 2013).
If there are very few fibers present, the stress on the composite may be high enough to break
the fibers thus causing the tensile-softening effect as described by Shah et al. (2012). When the
volume of fibers is lower than the minimum fiber volume fraction the fibers weaken the material
rather than strengthen it, and the failure is controlled by the matrix. The reinforcing action of
17
fibers and tensile-hardening is only observed once the fiber volume fraction exceeds the critical
fiber volume fraction (Shah et al, 2012).
Figure 2.1 Softening (a) and hardening (b) post-cracking behavior of SFRC - adapted from fib MC2010
(2013)
To determine tension behavior of FRC, uniaxial tests are not advised for standard testing due
to difficulty to carry out the tests. Instead, the relationship between tensile stress and crack
width is typically derived by inverse analysis from the results from two- or three- point bending
tests. Typical results from a bending test are presented in Figure 2.2 (fib Model Code 2010).
On the vertical axis of the plot is the applied load, and on the horizontal axis is the crack mouth
opening displacement (CMOD) measured at a notch cut in the test specimen (beam). CMOD is
defined as the opening of the notch at the bottom face of the beam for three-point bending tests.
For four-point loading tests the measured parameter is the crack top opening displacement
(CTOD).
Figure 2.2 Typical load-CMOD curve from fib MC2010 (2013)
According to fib Model Code 2010, fiber-reinforced concrete can be classified to five different
classes based on the results from two- or three- point bending tests. Two important stresses
from the measured post-cracking behavior are used: the first corresponding to CMOD
of 0.5 mm (fR1k) and the second to CMOD of 2.5 mm (fR3k). Stress fR1k is significant for
serviceability limit states while fR3k for ultimate limit state design. The classification has two
parameters where one corresponds to fR1k and represents a strength interval between consecutive
18
strength values (1, 1.5, 2, 2.5, 3, 4, 5, 6, …MPa), and a letter a, b, c, d or e, where each letter
represents different residual strength ratio as follows:
𝑓𝑅3𝑘 𝑓𝑅3𝑘
𝑎 𝑖𝑓 0.5 < < 0.7 𝑑 𝑖𝑓 1.1 ≤ < 1.3
𝑏 𝑖𝑓 0.7 ≤ < 0.9 𝑒 𝑖𝑓 1.3 ≤
𝑓𝑅3𝑘
𝑐 𝑖𝑓 0.9 ≤ < 1.1
𝑓𝑅1𝑘
For example, material labeled as 2d would have fR1k equal to 2 – 3 MPa and fR3k/ fR1k ratio from
1.1 to 1.3.
2.1.2 Factors Affecting Behavior of FRC

The impact of fibers on behavior of concrete depends on several factors which include fiber
volume ratio, fiber length, fiber aspect ratio, fiber tensile strength and the strength of concrete
matrix (ACI Committee 544, 2008).
Fiber length does not play a significant role in post-cracking behavior of SFRC containing
reinforcing bars (Deluce and Vecchio, 2013). However, the longer the fiber is, the more cracks
it can bridge. Short fibers should not be smaller than crack spacing in order to be able to transfer
stresses between the cracks.
Fiber volume ratio influences the properties of hardened concrete and also the workability of
freshly mixed concrete. The higher the fiber volume ratio, the more significant the effect
becomes. The flexural strength increases, however concrete workability decreases with high
fiber concentration and special mixing and placing methods are needed (Zollo, 1997). ACI
Committee 544 (1993) suggests a range of fiber volume ratio from 0.25% to 2% for
conventional steel fiber reinforced concrete.
Fiber aspect ratio is defined as the ratio of the fiber length to the fiber diameter and indicates
the slenderness of the fiber (Johnston, 2001). Higher aspect ratio means higher surface area of
fibers and higher surface area of fiber improves bond between the fibers and the concrete which
leads to better performance of the composite. Higher aspect ratio also improves the residual
post-cracking tensile strength and toughness of the concrete as described by Johnson (2001),
however reduced workability applies in cases with high aspect ratio.
Tensile strength of fibers is the tensile stress that results in the rupture of the fiber. Tensile
strengths of fibers can be up to 2000 MPa. The more stress the fiber is able to resist the more
ductile behavior the member experiences. When fibers with high tensile strengths are used, the
fiber will be pulled out of the matrix without rupturing, and the residual tensile strength of the
FRC will depend of the bond strength between the fibers and the concrete matrix (Susetyo,
2009).
19
2.2 TESTS ON FRC DEEP BEAMS

Four studies of the effect of steel fibers on behavior of deep beams were evaluated. Mansur and
Alwist (1984) tested 12 fiber-reinforced deep beams with openings and analyzed the effect of
volume fiber ratio, opening location and a/d ratio. Narayan and Darwish (1988) tested 12
reinforced concrete deep beams including 11 with steel fiber reinforcement. The main focus of
their study was the effect of fiber volume ratio, a/d ratio and concrete strength. Mansur and
Ong (1991) analyzed 10 reinforced concrete deep beams including 9 with steel fibers, observing
mainly the influence of fiber volume ratio, a/d ratio and transversal and longitudinal
reinforcement. One of the newest research studies focused on deep beams was carried out by
Campione (2012) and consists of 4 reinforced concrete deep beams from which 2 of them
included steel fibers.
Numerous other tests were conducted, however only two of the previously mentioned ones were
considered suitable for the validation of the extended five-spring model. When selecting studies
relevant for the validation of the extended five-spring model, only members with height over
500 mm were to be considered. Such restriction is due to the behavior of deep beams, which
cannot be captured correctly if the size of the tested specimen is smaller than 500 mm. Also
tests where cylinders were used to apply load, without presence of loading plates transferring
the load, were discarded from the tests used for the validation. This was due to the unrealistic
concentration of stresses under the loading point and thus invalid results obtained from the test.
The experimental results by Campione (2012) were discarded due to the lack of information
about the specimen.
Given all of the limitations mentioned above, only two of the studies were considered relevant
for the validation of the five-spring model. The tests of Mansur and Alwist (1984) and the
experiments conducted by Mansur and Ong (1991) were considered and a database with
provided beam properties which can be seen in Appendix A was created. Both of the studies
will be described in detail in the following sections.
2.2.1 Mansur and Ong (1991)

Mansur and Ong (1991) conducted a study including 9 fiber reinforced deep beams and one
control specimen without fibers. The major parameters observed in the study were a/d ratio,
volume fraction of fibers and ratios of longitudinal and transversal reinforcement.
All of the specimens had height of 500 mm, width of 90 mm, maximum aggregate size of 10
mm, yield strength of bottom longitudinal and web reinforcement of 440 MPa and 375 MPa,
respectively. Bottom longitudinal reinforcement consisted of 4 reinforcement bars, with 16 mm
diameter, placed in two layers. The rebars were welded to a 15 mm thick steel plate at each end.
Web reinforcement had diameter of 6 mm and the number of bars varied depending on the
length and cross section of the member. Bottom longitudinal reinforcement ratio was 1.93%
while the transversal ratio varied from 0.42% to 1.26%. The layout of the reinforcement and
experimental setup is shown in Figure 2.3. Fibers used in the experiment were straight, slightly
twisted with 0.5 mm square in cross section and length of 30 mm. Concrete strength was
20
determined for each specimen individually and is specified in table in Appendix A. Further
details regarding the test setup and properties of the specimen are shown in the table in
Appendix A with reference number 1. Data highlighted in purple was not directly specified by
the authors, but obtained by an assumption or calculation based on provided data. The beams
were tested in four-point loading, where the loads and reactions were applied through rollers
and bearing blocks. Load was applied in 50 kN increments until the failure which was noted
after the collapse.
Figure 2.3 Experimental setup by Mansur and Ong (1991)
Results showed that the reduction of the shear span resulted in lower occurrence of flexural
cracking. While for beams with a/d ratio of 1.85 the flexural cracks occurred first, for a/d ratio
of 1.23 flexural and diagonal cracks propagated simultaneously. For shorter shear spans
diagonal cracks propagated first with little or no occurrence of flexural failure. It was also
observed that increase in a/d ratio results in lower stiffness of the member as shown in Figure
2.4a.
The effect of increasing fiber volume was most significant in cracking behavior of the beam.
While in the beam with no fibers the cracks formed suddenly and then propagated rapidly, in
beams with fibers the crack propagation and widening was significantly slower. Increasing fiber
volume ratio resulted in higher ultimate strength as shown in Figure 2.4b.
Figure 2.4 Experimental results by Mansur and Ong (1991)
21
The tests also included two beams designated to study the effect of vertical and longitudinal
web reinforcement. It was concluded that the longitudinal reinforcement has almost no
influence on the shear strength of the beams. However, increasing the ratio of vertical
reinforcement proved to significantly increase the shear resistance of the member.
To describe the shear failure in detail, beam B4 from the study was selected as a sample. The
beam had 1% of fiber volume ratio, and vertical reinforcement of 0.42% and a/d ratio of 1.23.
In the test the shear cracks propagated at the same time as flexural cracks. However, it was the
propagation of shear cracks that lead to the final failure of the beam. As described in the
experiment, at 80 to 90 percent of the ultimate load, one of the diagonal cracks began to grow
excessively. This critical diagonal crack was described to originate at the mid-depth of the beam
and with increasing load to extend to the support and to the loading point. A crushing of the
concrete was observed between the loading point and the tip of the inclined crack. The size of
the critical loading zone which is susceptible to compression introduced by the loading plate is
also one of the parameters influencing the overall shear resistance of the beam. The cracking
pattern of the beam can be observed in Figure 2.5.
Figure 2.5 Cracking pattern of the beam B4 by Mansur and Ong (1991)
2.2.2 Mansur and Alwist (1984)

Mansur and Alwist (1984) tested 12 fiber reinforced concrete deep beams with openings. Such
beams are not relevant for the validation of the extended five-spring model, however one of the
tested beams was a control beam for comparison, and did not include any openings. The control
specimen had height of 850 mm and width of 80 mm, maximum aggregate size of 10 mm, yield
strength of bottom longitudinal and web reinforcement of 418 MPa and 304 MPa, respectively.
Bottom longitudinal reinforcement consisted of 2 reinforcement bars, with 16 mm diameter.
The rebars were welded to a 20 mm thick steel plate at each end. Web reinforcement had
diameter of 3.3 mm with 12 bars in longitudinal direction and 31 in vertical direction. Bottom
longitudinal reinforcement ratio was 0.81% and transversal ratio 0.43%. The specimen setup is
illustrated in Figure 2.6. Fibers used in the experiment were hook-ended fibers, glued together
in bundles. Length of the fibers was 30 mm and diameter 0.4 mm. Concrete strength of the
22
specimen was 40 MPa. Further details regarding the test setup and properties of the specimen
are shown in the table in Appendix A with reference number 2. The beam was tested in four-
point loading where the loads and reactions were applied through rollers and bearing blocks.
The beam was loaded with 20 kN increments until the failure.
Figure 2.6 Experimental setup by Mansur and Alwist (1984)
The results showed that the beam behaved linearly at the beginning. Diagonal cracks appeared
first and with increasing load started propagating and widening. Flexural cracks appeared later
in the loading but did not affect the final failure. Final failure occurred by rapid propagation
and widening of one of the diagonal cracks as illustrated in Figure 2.7.
Figure 2.7 Failure mode of control specimen by Mansur and Alwist (1984)
23
2.3 TWO-PARAMETER KINEMATIC THEORY AND FIVE-SPRING MODEL FOR

DEEP BEAMS
The two-parameter kinematic theory is based on a simple kinematic description of the
deformation patterns of deep beams. The theory has been introduced by Mihaylov et al. (2013)
and is capable to predict the shear strength and deformation patterns of deep beams at shear
failure. Later the method has been expanded to a five-spring model (5sm) which is capable of
predicting the complete pre-peak and post-peak behavior of deep beams (Mihaylov, 2015). The
5sm allows engineers to evaluate the safety and assess the deformations and crack widths of
deep reinforced concrete beams such as transfer girders (Mihaylov et al., 2013). However, the
theory does not account for the action of fibers.
Based on results from tests on the behavior of deep reinforced concrete beams under monotonic
and cyclic loading by Mihaylov (2008), observations made were used as a background for the
2PKT. The main assumptions of the kinematic approach are illustrated in Figure 2.8. It is
assumed that the shear failure of deep beams occurs along a straight critical shear crack that
extends from the support to the load. The concrete above the critical crack is modeled as a rigid
block with a critical loading zone (CLZ) located under the loading plate. The concrete below
the crack is modeled with a series of rigid radial struts that connect the loading point with the
bottom longitudinal reinforcement. The two regions are connected by the stirrups and
longitudinal reinforcement.
A B
1
2.5(h-d)
w
slip
l k=l 0 l b1
l b1e = (V/P)l b1 P
c
CLZ
A
d
h
 v
1
l b2
r =0
B z


V x
a+  t,avg d cot 
 t,min
t,max
 t,avg
Figure 2.8 Main assumptions


of kinematic model (Mihaylov et al.,2013)
DOF t,avg DOF  c c
x 
The two-parameter kinematic theory assumes that with respect to the loading plate, the motion
z   c
of the concrete block above the critical crack can be described as a rotation around the loading
d
l k=l 0
  z  z
point and a vertical translation of the rigid block, see Figurex 2.9.
 t,min
The rotation is proportional to
t
the average strain in the bottom c
a+  t,avg d cot  reinforcement
longitudinal l k=l 0 εt,avg, while the translation equals
x
the vertical displacement Δc of the critical loading zone (Mihaylov et al., 2013). The two degrees
of freedom (DOFs) as described are the two parameters of the kinematic theory. The elongation
24
of the bottom reinforcement causes the rigid radial struts to rotate around the loading point and
the cracks between them to widen. The transverse displacement in the critical loading zone
causes the critical diagonal crack to widen and a slip displacement to occur along the crack. As
illustrated in Figure 2.9, both degrees of freedom contribute to the increase of tensile strains in
transverse reinforcement (stirrups).
Figure 2.9 Degrees of freedom of kinematic model (Mihaylov et al.,2013)
The angle of the critical crack depends on the shear-span-to-depth (a/d) ratio and is defined as
α1. Angle α1 is determined as the smallest value of two other angles α and θ, see Figure 2.10.
Angle α is the angle of the line connecting the inner edge of the support plate with the outer
edge of the effective width of the loading plate. The effective width of the loading plate is the
portion of the loading plate participating in the transfer of the applied load to the support. Angle
α governs for deep beams while angle θ captures the transition to slender beams when the shear
span a becomes relatively large as compared to the effective depth d. Angle θ can be obtained
from sectional shear models for slender beams, however, for simplification, its value will be
assumed equal to 35°.
d
 
1
lk l0
 t,min
t,max
Figure 2.10 Variation of the angle α1 during transition from deep to slender beams (Mihaylov et al.,2013)
Based on the assumptions for the kinematics of deep beams, and assuming that the two degrees
of freedom εt,avg and Δc are known, the two-parameter kinematic theory is able to provide the
complete displacement field of the shear span of deep beams. In other words, the horizontal and
25
vertical displacement of each point from the shear span can be expressed with DOFs εt,avg and
Δc. As shown by Mihaylov et al. (2013), these expressions are as follows:
- for points below the critical diagonal crack

𝛿𝑥 (𝑥, 𝑧) = 𝜀𝑡,𝑎𝑣𝑔 𝑥 (1)
𝜀𝑡,𝑎𝑣𝑔 𝑥 2 (2)
𝛿𝑧 (𝑥, 𝑧) =
ℎ−𝑧
- for points above the critical diagonal crack
𝛿𝑥 (𝑥, 𝑧) = 𝜀𝑡,𝑎𝑣𝑔 (ℎ − 𝑧)cot𝛼 (3)
𝛿𝑧 (𝑥, 𝑧) = 𝜀𝑡,𝑎𝑣𝑔 𝑥cot𝛼 + ∆𝑐 (4)

These equations can be used to derive important deformations in the shear span, again as a
function of εt,avg and Δc. Such important deformation for example is the width of the critical
diagonal crack halfway along the crack. This width is expressed as follows:
𝑙𝑘 = 𝑙0 + 𝑑(𝑐𝑜𝑡𝛼 − 𝑐𝑜𝑡𝛼1 ) ≤ 2𝑙0 (5)
𝑙0 = 1.5(ℎ − 𝑑)𝑐𝑜𝑡𝛼1 ≥ 𝑠𝑐𝑟 (6)
𝜀𝑡,𝑎𝑣𝑔 𝑙𝑘 (7)
𝑤 = ∆𝑐 𝑐𝑜𝑠𝛼1 +
2𝑠𝑖𝑛𝛼1
0.28𝑑𝑏 2.5(ℎ − 𝑑) (8)

𝑠𝑐𝑟 =
𝜌𝑙 𝑑
Where lk is the length of dowels provided by the bottom longitudinal reinforcement, l0 is the
length of heavily cracked zone at bottom of critical diagonal crack, and scr is the distance
between the cracks along bottom longitudinal reinforcement (see Figure 2.8).
Using the two degrees of freedom, the kinematic model can be used to evaluate the overall
deflection of the member. The deflection is defined as the relative displacement between the
support and the section with maximum bending moment, assuming that the latter section
remains vertical (Mihaylov, 2015). The deflection is defined as follows:
𝜀𝑡,𝑎𝑣𝑔 𝑙𝑡 (9)
∆= 𝑎 + ∆𝑐
𝑑
Where the first term of the equation describes the deflection of shear span Δt due to elongation
of bottom longitudinal reinforcement. The variable lt is the length of bottom reinforcement
within the cracked part of shear span, as illustrated in the free-body diagram in Figure 2.12, and
is defined as:
𝑙𝑡 = 𝑑 𝑐𝑜𝑡(𝛼1 ) + 𝑙𝑘 − 𝑙0 (10)
26
Given the two deflections Δt and Δc, the five-spring model represents the shear span using two
sets of springs connected in series and loaded by shear force V. As shown in Figure 2.11, a set
of four parallel springs elongates by the value of deflection Δc, while the fifth spring elongates
by Δt. The four parallel springs represent the shear behavior of the beam, while the fifth spring
represents the flexural behavior.
Figure 2.11 Five-spring model (Mihaylov, 2015)
The forces of the five-spring model, and their definition within the shear span of the beam, are
illustrated in Figure 2.12. The force VCLZ is the shear carried by the critical loading zone, Vci is
the shear resisted by aggregate interlock along the critical crack, Vd is the shear resisted by the
dowel action of the bottom flexural reinforcement, and Vs is the shear resisted by the stirrups.
The sum of these forces is the shear force V obtained from the equilibrium of the vertical forces
acting on the rigid block. The force in the flexural spring is the shear force derived from the
moment equilibrium of the shear span taken about the point of application of the compression
force C in the section with maximum moment. The equilibrium equation is as follows:
𝑇(0.9)𝑑 (11)
𝑉=
𝑎
Where T is the tension force in the flexural reinforcement and 0.9d is the assumed lever arm
between the compression force C and the tension force T.
Figure 2.12 Free-body diagram of the rigid block (Mihaylov, 2015)
Therefore, given that the shear force can be expressed in two different ways, the equilibrium of
the spring forces requires that:
27
𝑇(0.9)𝑑 (12)
= 𝑉𝐶𝐿𝑍 + 𝑉𝑐𝑖 + 𝑉𝑠 + 𝑉𝑑
𝑎
Shear failure is predicted if the set of four parallel springs fails first, and failure due to yielding
of the flexural reinforcement occurs if the flexural spring fails first. The force in flexural
reinforcement is expressed as:
0.33√𝑓′𝑐 (13)
𝑇 = 𝐸𝑠 𝐴𝑠 𝜀𝑡,𝑎𝑣𝑔 + 𝐴𝑐,𝑒𝑓𝑓 ≤ 𝐴𝑠 𝑓𝑦
1 + √200𝜀𝑡,𝑎𝑣𝑔
Where the first term of the equation models the resistance of bare steel reinforcement. The
second term of the equation models the tension stiffening effect of the concrete, where the area
of concrete involved in the tension stiffening effect is taken as:
ℎ (14)
𝐴𝑐,𝑒𝑓𝑓 = min [2.5(ℎ − 𝑑), ] 𝑏
2
The last part of the equation for the force T is the yield force of the bottom longitudinal
reinforcement.
The shear carried by the critical loading zone is derived using the degree of freedom Δc. The
critical loading zone has a triangular shape that depends on the effective width of the loading
plate and the angle of the critical crack. The concrete in the critical loading zone is subjected to
diagonal compressive stresses σ and strains ε, where the strain at the bottom plane of the critical
loading zone is εmax. The strain εmax can be expressed using DOF Δc as shown in Figure 2.13.
The diagonal compressive stress is obtained through appropriate stress-strain relationships for
concrete under uniaxial compression, which is in this case Popovics (1970) curve for stress-
strain relationship. The average stress σavg is multiplied by the area of the section passing
through the edge of the loading plate and perpendicular to the bottom face of the critical loading
zone.
Figure 2.13 Critical loading zone (Mihaylov, 2015)
The shear carried by the critical loading zone is then expressed as:
28
𝑉𝐶𝐿𝑍 = 𝜎𝑎𝑣𝑔 [𝜀𝑚𝑎𝑥 (𝛥𝑐 )]𝑏𝑙𝑏1𝑒 𝑠𝑖𝑛2 𝛼 (15)
The shear force resisted by the aggregate interlock is expressed using the average shear stress
vci transferred across the critical crack by interlocking of the rough crack surfaces:
𝑉𝑐𝑖 = 0.18𝑣𝑐𝑖 𝑏𝑑 (16)

The value of vci is evaluated halfway along the critical crack as a function of crack width w and
slip s. Where the crack slip is defined as follows:
𝑠 = 𝛥𝑐 sin(𝛼1 ) (17)
In order to compute the value of vci, contact density model (CDM) developed by Li et al. in
1989 is used. The model represents the crack surface using series of planes at different angles,
where the planes with the same angle on each side of the crack are connected by contact springs.
The springs have an elastic-perfectly-plastic behavior in compression and zero resistance in
tension. The CDM also accounts for clamping stresses, which are not taken into account in the
five-spring model and thus a factor 0.18 is adopted to reduce the shear resistance vci.
The shear carried by the stirrups across the critical diagonal crack is expressed using both
degrees of freedom of the kinematic approach. Firstly, the strain in the stirrups is evaluated as:
𝛥𝑐 + 0.25𝜀𝑡,𝑎𝑣𝑔 𝑑𝑐𝑜𝑡 2 𝛼1 (18)

𝜀𝑣 = 2
0.9𝑑
The stress in the stirrups is then expressed as:
𝜎𝑣 = min( 𝜀 𝑣 𝐸𝑣 , 𝑓𝑦𝑣 ) (19)
And the shear carried by the stirrups can be then taken as the stress in the stirrups multiplied by
the effective area of stirrups that are providing shear resistance:
𝑉𝑠 = 𝜎𝑣 𝜌𝑣 𝑏(𝑑𝑐𝑜𝑡𝛼1 − 𝑙0 − 1.5𝑙𝑏1𝑒 ) (20)
Where 𝜌𝑣 is the transverse reinforcement ratio which should not be larger than 0.15𝑓′𝑐 /𝑓𝑦 . The
value in brackets should not be smaller than 0.5𝑑𝑐𝑜𝑡(𝛼1 ). For beams without stirrups, the five-
spring model becomes a four-spring model as the value of 𝑉𝑠 is 0.
The last component of the equilibrium equations represents the shear force resisted by the dowel
action of the bottom longitudinal reinforcement. The resistance depends on the number of
bottom flexural bars, diameter of the bars, and the length of the dowels lk. The equation is then
as follows:
2
12𝐸𝑠 𝜋𝑑𝑏4 𝑑𝑏3 𝑇 (21)
𝑉𝑑 = 𝑛𝑏 3 ∆ 𝑐 ≤ 𝑛 𝑓
𝑏 𝑦 [1 − ( ) ]
64𝑙𝑘 3𝑙𝑘 𝐴𝑠 𝑓𝑦
Where the first term of the equations represents the behavior of the elastic fixed-fixed dowels
subjected to relative transverse support displacement ∆𝑐 . The second part corresponds to the
29
formation of plastic hinges at the ends of the dowels, where the moment capacity of the plastic
hinges is reduced due to the tension force T acting on the bars.
The five-spring model is then solved for increasing deflection ∆𝑐 at a step of 0.1 mm. The final
load-deflection curve is then created by calculating the shear force at each step for the given
shear distortion of the CLZ and plotted for the consecutively calculated overall deformation.
Given that the elongation of the flexural spring is ∆𝑡 = ∆ − ∆𝑐 , the only unknown kinematic
parameter is the ∆𝑐 . The deformation can be obtained solving the equilibrium equation (12).
Figure 2.14 illustrates the shear forces in the five-spring model for a sample beam with
properties described in Table 2.1 for imposed shear distortion of the CLZ of 5 mm. The figure
describes how only one point of many that create the final load-deflection curve is obtained.
Table 2.1 Sample beam properties
a/d b d h lAs ⌀I
ρI No. fy ag fc' fyv ρv ⌀v No.
2
(mm) (mm) (mm) (mm) (%) (mm ) (mm) barsI (MPa) (mm) (MPa) (MPa) (%) (mm) barsv
1,55 400 1095 1200 3900 1,27 5541,8 28 9 550 20 40 490 0,1 10 10
The horizontal axis of Figure 2.14 shows the range of tensile strain in the bottom longitudinal
reinforcement. The vertical axis represents the shear forces obtained from all the springs of the
five-spring model. Thick red line represents the sum of shear forces resisted by the critical
loading zone, aggregate interlock, stirrups and dowel action. As the shear force resistance of
the CLZ is only dependent on the ∆𝑐 parameter, the shear force does not depend on the strain
in the bottom longitudinal reinforcement, the line representing VCLZ is a constant. The shear
force resisted by the stirrups is also a constant, since in this part of the loading the stirrups are
considered to have yielded. At the moment of shear distortion of the CLZ that is considered in
this case, the stress in the stirrups is dependent on the yield strength of the bottom longitudinal
reinforcement rather than on the strain and modulus of elasticity of the reinforcement. The shear
force resisted by the aggregate interlock is dependent on the strain in the dowels provided by
the reinforcement. Therefore, the curve representing the aggregate interlock decreases with
increasing strain. The last force contributing to the sum of shear forces is the dowel action
which is dependent on both of the parameters. The shear force resisted by the dowels decreases
as the strain in the bottom longitudinal reinforcement increases until the point where the
reinforcement has yielded and is no longer able to transfer any stresses.
The thick black line represents the shear obtained from the moment equilibrium. The orientation
of this bilinear curve depends on the properties of the concrete matrix and bottom longitudinal
reinforcement. First part of the curve is the contribution of tensile resistance of the concrete
matrix and second part represents the contribution of the reinforcement. The intersection of the
two thick lines highlighted with red dot, marks the solution of the equilibrium equation.
Consecutively, the value of the shear force at the intersection is the shear resistance of the beam
at 5 mm distortion of the critical loading zone. The intersection is found iteratively using the
bisection method.
30
1200
V = ΣVi
1000 equilibrium
Shear Forces (kN) 800

V = T(0.9d)/a
600
VCLZ
400
Vci
200 Vs
Vd
0
0 0,5 1 1,5 2 2,5 3
Average tensile strain εt,avg (mε)
Figure 2.14 Shear forces in five-spring model at 5 mm imposed distortion of the CLZ
The final curve of the model is then generated by gradually increasing the distortion of the CLZ
from 0 to 15 mm, and repeating the calculations illustrated in Figure 2.14. For each step, which
represents a certain value of the distortion of the CLZ, the solution (red dot) is found using
equilibrium equations and bisection method. Obtained intersections then create a load-
displacement curve as shown in red in Figure 2.15, where the overall deflection ∆ is calculated
for given ∆𝑐 . Horizontal axis of the Figure 2.15 represents the imposed displacement ∆ with
range from 0 to 15 mm in this case. Vertical axis represents the values of shear forces obtained
from the equilibrium equations, which can be described as the path of the red dot for different
imposed distortions of the CLZ. Thick red curve corresponds to the solutions of the equilibrium
equations for increasing displacement and is described as a sum of the shear forces provided by
each of the parameters described in the 5sm. Thinner curves of the graph each represent one of
the four parallel springs of the five-spring model. These curves are obtained also using the
Figure 2.14 and generating the intersections of each of the curves with the thick black curve
describing the moment equilibrium equation.
As it can be seen in the Figure 2.15 the CLZ provides the most shear force resistance of the
model. The value of the shear force provided by the critical loading zone increases until
deflection of about 6 mm where crushing of the CLZ occurs. The peak of the VCLZ occurs at the
same deflection as the maximum shear force resisted by the beam is achieved. The aggregate
interlock, which reaches its maximum right after the beam reaches its peak shear strength and
then starts decreasing, also greatly contributes to the overall shear resistance of the beam. The
shear force provided by the dowel action of the bottom longitudinal reinforcement does not
vary much throughout the loading. The amount of shear resistance is in this case comparable
with shear force provided by the stirrups. The contribution of stirrups is increasing until the
yield strength is reached, after that the curve becomes constant since it is no longer dependent
on any of the degrees of freedom considered within the model.
31
1200
V
1000
Shear Forces (kN) 800
600 VCLZ
400
Vci
200 Vs
Vd
0
0 5 10 15
Mid-span deflection (mm)
Figure 2.15 Equilibrium of forces in five-spring model for imposed displacement Δ
As it has been already mentioned previously, the 5sm proposed by Mihaylov (2015) is only
applicable to deep beams with conventional reinforcement and with or without stirrups. The
model does not account for the action of fibers. Based on described experiments on deep beams
reinforced with fibers carried out by various researchers it has been concluded that inclusion of
fibers can enhance the behavior of the deep beams and their post-peak response. The main
objective of this thesis is to introduce the fiber action into the existing five-spring model. The
aim of the next chapter is to extend the 5sm to deep beams with conventional and fiber
reinforcement.
32
3 EXTENDED FIVE-SPRING MODEL FOR FRC DEEP

BEAMS
3.1 INTRODUCTION
The extension of the five-spring model is divided into several subchapters, each of which is
focused on a different part of the model that is affected by the addition of fibers into the concrete
mixture. An extension of the model is made by adding another spring to the spring model to
represent the shear resistance of fibers along the diagonal crack of the beam. This extension
provides a fifth spring to the set of parallel springs by using the Simplified Diverse Embedment
Model by Lee et al. (2013). Following the already existing model, two adjustments are made to
account for the effect of the fibers on the properties of the concrete.
The first adjustment is implemented in the flexural spring which now also accounts for tension
stiffening behavior of the beam caused by the addition of fibers using the Tension-Stiffening
Model by Lee et al. (2013). The second adjustment is made in the stress-strain relationship for
fiber-reinforced concrete under uniaxial compression, that now includes the impact of fibers on
the resistance of the critical loading zone under uniaxial compression. The last adjustment does
not deal with the effect of fibers, but rather extends the model to be able to represent not only
three-point loading cases by giving the deflection under the loading point, but also four-point
loading cases by introducing additional equations that enable the calculation of deflection in
the middle of the beam.
In this chapter all of these extensions and modifications are described as given in the literature,
including their limitations and formulations. Each of the implemented models is firstly
developed as a separate code in Matlab and its results are validated against information and
graphs provided in the literature. Secondly a description of the modified formulation of the code
as it enters the five-spring model is provided and justified with description of its function within
the model.
3.2 TENSION IN FIBERS ACROSS THE CRITICAL DIAGONAL CRACK

3.2.1 Introduction
One of the most crucial and beneficial aspects of adding steel fibers into the concrete structures
is the non-brittle behavior after concrete cracking, that is achieved through fibers bridging the
cracks. Whilst in normal concrete the tensile stress quickly decreases as soon as the crack starts
propagating, in fiber-reinforced concrete we can observe the action of fibers that transfer the
tensile stresses and contribute to the total response. Such behavior provides considerable tensile
resistance to the structure even after the propagation of the crack.
Figure 3.1 illustrates the tensile behavior of FRC. The behavior of regular concrete with tension-
softening effect is illustrated with dash-dotted line. The maximum value of stress depends on
the tensile strength of the concrete. For FRC members the effect of fibers is included into the
overall response of the member. As it can be seen in Figure 3.1, the fibers start acting as soon
33
as the crack starts propagating and provide additional resistance to the member even after the
concrete matrix has lost its ability to transfer stresses. The solid line in the figure represents the
response of FRC member, which consists of the sum of the tension softening effect and the
action of fibers. Based on these observations it can be concluded, that in order to accurately
predict the post-cracking response of FRC, the tensile stress behavior attainable with fibers
should be evaluated.
Figure 3.1 Tensile behavior of FRC (Lee et al., 2013)
Figure 3.2 illustrates the additional force Ff that will be implemented within the extended five-
spring model to represent the actions of fibers across the critical diagonal crack. The action of
fibers while bridging the crack results in additional vertical force VF that will enter the parallel
set of springs and provide higher shear resistance to the member.
Figure 3.2a shows the additional stress ff representing the tensile stress at crack due to steel
fibers. The stress can be calculated using the SDEM by Lee et al. (2013) which describes the
stress transfer through the crack taking into account the concrete matrix and steel fibers (hooked
or straight). However, to obtain the stress ff transferred through the crack only the contribution
of fibers, without the resistance provided by the concrete matrix will be considered. The stress
vf represents the vertical component of the force ff acting along the critical crack of the member.
For a given width b of the cross section and the length of critical diagonal crack, the shear force
resistance Ff provided by the fibers bridging the crack can be calculated. The vertical
component VF of the force Ff is then calculated using the angle α1 of the critical crack.
Figure 3.2b shows the tensile stresses ff resisted by the fibers for a given crack width. The
figure represents a sample beam with fiber volume of 1%, fiber length of 30 mm, diameter of
fibers of 0.5 mm and straight fibers. Horizontal axis of the figure describes different crack
widths varying from 0 mm to 10 mm. Vertical axis shows the stresses transferred by the fibers
along the area of the critical crack. The curve then represents the stress transferred through the
steel fibers for increasing width of the crack.
34
vf
Figure 3.2 Illustration of additional force provided by fibers in the extended five-spring model
Several researchers have tried to describe the tensile stress behavior of fibers but so far the most
realistic method is the Diverse Embedment Model (DEM) by Lee et al. (2011) that predicts
tensile stress in fibers across a given crack. In this model the pullout strength of fibers is
analyzed taking into account embedment on both sides and considering both frictional bond
behavior and mechanical anchorage effects. Given such circumstances, the DEM can be applied
to end-hooked fibers as well as straight fibers. The DEM method also takes into consideration
all possible fiber orientations and embedment lengths.
Due to the complicated calculation of the tensile strength of the fibers using the DEM, which
includes double numerical integration, Lee et al. (2013) proposed a Simplified DEM (SDEM).
SDEM is derived from the DEM by eliminating the integration and by neglecting the slip on
the longer embedded side. The frictional bond behavior and mechanical anchorage effects are
incorporated in the model as coefficients preventing an overestimation of the tensile stress
attained by fibers. The tensile stress behavior of the fibers obtained using the SDEM has shown
a good agreement with the results obtained using the DEM and thus it can be concluded that
despite the simplification the accuracy of the model remained.
3.2.2 Formulation of the SDEM

In order to be able to perform the SDEM it is necessary to know the compressive strength of
concrete that will be considered in the calculation and the fiber volumetric ratio included in the
concrete. Regarding the properties of the fibers, the length of the fibers and its diameter are
necessary. In case of using hook-ended fibers, the length of the straight part between the
mechanical anchorages should be specified, see Figure 3.3.
Figure 3.3 Fiber types used in the study by Lim et al. (1987)
35
The tensile stress in steel fiber-reinforced concrete without conventional reinforcement can be
described with the following equation:
𝑓𝑆𝐹𝑅𝐶 = 𝑓𝑓 + 𝑓𝑐,𝑠𝑜𝑓𝑡 (22)
Where ff is the tensile stress at crack due to steel fibers and fc,soft stands for the tensile stress in
concrete due to tension softening effect described by Voo and Foster (2003) as follows:
𝑓𝑐,𝑠𝑜𝑓𝑡 = 𝑓𝑐𝑟 𝑒 −𝑐𝑤𝑐𝑟 (23)
Where fcr is the cracking stress of the concrete estimated as 𝑓𝑐𝑟 = 0.33√𝑓′𝑐 (Ramaswamy et al.
1994). The coefficient c is either 15 or 30 for concrete or mortar, respectively, and wcr is the
given average crack width.
Due to the simplification in the SDEM, the bond mechanism for the pullout behavior of steel
fibers and the mechanical anchorage effect due to end-hooks are evaluated separately. The total
tensile stress at a crack due to the fibers can be described as:
𝑓𝑓 = 𝑓𝑠𝑡 + 𝑓𝑒ℎ (24)
Where fst is the tensile stress due to frictional bond behavior of steel fibers and feh is the tensile
stress due to mechanical anchorage effect of end-hooked steel fibers. These tensile stresses are
expressed as follows:
2
𝑙𝑓 2𝑤𝑐𝑟 (25)
𝑓𝑠𝑡 = 𝛼𝑓 𝑉𝑓 𝐾𝑠𝑡 𝜏𝑓,𝑚𝑎𝑥 (1 − )
𝑑𝑓 𝑙𝑓
𝑙𝑖 − 2𝑤𝑐𝑟 (26)
𝑓𝑒ℎ = 𝛼𝑓 𝑉𝑓 𝐾𝑒ℎ 𝜏𝑒ℎ,𝑚𝑎𝑥 2 ( )
𝑑𝑓
Where Vf stands for fiber volumetric ratio, lf is the fiber length, df is the diameter of the fiber,
αf is the fiber orientation factor that can be assumed to be 0.5 for cases where the dimensions
of the structural member are much larger than the fiber length (Lee et al., 2013), and τ is the
pullout strength of a single fiber taken from Table 3.1 (Voo and Foster, 2003).
36
Table 3.1 Pullout strength of single fiber by Voo and Foster (2003)
The parameters Keh and Kst are defined as follows:

𝛽𝑓 𝑤𝑐𝑟 (27)
𝐾𝑠𝑡 = 𝑓𝑜𝑟 𝑤𝑐𝑟 < 𝑠𝑓
3 𝑠𝑓
𝑠𝑓 𝛽𝑓 𝑤𝑐𝑟 (28)
𝐾𝑠𝑡 = 1 − √ + √ 𝑓𝑜𝑟 𝑤𝑐𝑟 ≥ 𝑠𝑓
𝑤𝑐𝑟 3 𝑠𝑓
2 𝑤𝑐𝑟 1 𝑤𝑐𝑟 2 (29)

𝐾𝑒ℎ = 𝛽𝑒ℎ [ − ( ) ] 𝑓𝑜𝑟 𝑤𝑐𝑟 < 𝑠𝑒ℎ
3 𝑠𝑒ℎ 5 𝑠𝑒ℎ
2
7𝛽𝑒ℎ 𝑠𝑒ℎ 2(√𝑤𝑐𝑟 − √𝑠𝑒ℎ ) 𝑙𝑓 − 𝑙𝑖 (30)
𝐾𝑒ℎ = 1+ ( − 1) √ − 𝑓𝑜𝑟 𝑠𝑒ℎ ≤ 𝑤𝑐𝑟 <
15 𝑤𝑐𝑟 𝑙𝑓 − 𝑙𝑖 2
2
𝑙𝑖 − 2𝑤𝑐𝑟 𝑙𝑓 − 𝑙𝑖 𝑙𝑖 (31)
𝐾𝑒ℎ = ( ) 𝐾𝑒ℎ,𝑖 𝑓𝑜𝑟 ≤ 𝑤𝑐𝑟 <
2𝑙𝑖 − 𝑙𝑓 2 2
Where 𝛽𝑓 , is a coefficient reflecting the effect of fiber slip on the longer embedded side and has
been analytically determined as 0.67. In the same way as the determination of βf, the value of
coefficient βeh has been defined as 0.76 where this coefficient prevents the tensile force by
mechanical anchorage from being overestimated due to the effect of a fiber slip on the longer
embedded side as defined by Lee et al. (2013).
37
The length li is the length of the straight part of the fiber between the mechanical anchorages
for end-hooked fibers, and can be obtained by subtracting 8 mm from the overall length of the
fiber lf based on the assumption shown in the Figure 3.3.
The value of sf, which refers to a slip at maximum tensile force due to mechanical anchorage of
fiber with inclination angle of 0 degrees with respect to the crack surface, can be assumed as
0.01 mm based on the experiments provided by Naaman and Najm (1991). Equally, the value
of seh, which describes the slip at maximum tensile force due to mechanical anchorage of an
end-hooked fiber, has been derived from the same experiments and is considered to be equal to
0.1 mm.
The last unknown value in the equations of Keh is the variable Keh.i which is the value of Keh at
wcr = (lf - li)/2. An important point that should be taken into account while applying the SDEM
is that only the DEM is suitable for SFRC in which fiber rupture is expected. Fiber rupture is
not considered in the SDEM.
3.2.3 Validation of the SDEM

To perform calculations of tensile stresses in steel fibers across cracks, a code in Matlab was
developed using the formulation of the SDEM as described in the previous chapter. The code
is verified and the SDEM is validated by performing comparisons with SDEM simulations and
test results from the literature (Lee et al. 2013, Lim et al. 1987, Susetyo 2009). The Matlab code
itself as it enters the five-spring model is included in Appendix B.
3.2.3.1 Experimental results by Lim et al. (1987)

Lim et al. (1987) tested members with straight and hook-ended fibers subjected to uniaxial
tension. These specimens had a rectangular cross section of 70 mm width and 100 mm
thickness. The input value of concrete cylinder strength was not specified and thus is just
assumed to be 40 MPa. It should be taken into consideration that such value might have a big
impact on the accuracy of the results.
For the comparison, test results for straight fibers tested by Lim et al. (1987) and SDEM results
by Lee et al. (2013) were used. In the experimental study the fiber volumetric ratio used has
been considered to be 1% and 1.5% respectively. Fiber length used was 30 mm and the diameter
of the fibers was 0.565 mm. As displayed in the Figure 3.4, we can observe that the proposed
Matlab code is slightly more conservative than results by Lee and the experimental results. It
can be assumed that the main cause of this difference is the assumed cylinder strength of the
concrete. The strength was selected with the intention to match as closely as possible all the
analytical results provided by Lee et al. (2013). Several cylinder strengths were considered and
the one that showed the best agreement with the analytical curve generated by Lee et al. (2013)
was used.
38
Figure 3.4 Comparison of proposed model and test results for members with straight fibers tested by Lim
et al. (1987)
Another validation has been performed using the data from the tests by Lim et al. (1987)
performed with end-hooked fibers, see Figure 3.5. In this experiment two types of fibers were
used. The graph on the left corresponds to a fiber volume ratio 0.5%, fiber length 30 mm and
diameter of the fiber 0.5 mm. The graph on the right represents fiber volume ratio of 1%, fiber
length 50 mm and diameter of the fiber 0.5 mm. The results show again good agreement with
the SDEM by Lee considering the shape of the curve. The curve generated using the created
Matlab code proves to be more conservative but it can be assumed that the main reason for such
difference is again the assumed concrete strength.
Figure 3.5 Comparison of proposed model and test results for members with end-hooked fibers tested by
Lim et al. (1987)
3.2.3.2 Experimental results by Susetyo (2009)

Susetyo (2009) performed direct tension tests using specimens with cross section of 70x100
mm2 and concrete compressive stress varying of 80 MPa and 50 MPa. In his experiments
Susetyo tested steel FRC members with hook-ended fibers with overall fiber length 50 mm and
39
fiber diameter 0.62 mm. Two examples shown in the Figure 3.6 refer to concrete strength of 50
MPa and fiber volume ratio of 0.5%, and 1% respectively. The frictional bond strength τf, max
and τeh,max was considered to be 0.396 f ′c and 0.429 f ′c respectively, based on previous research
of Voo and Foster (2003) and Lee et al. (2011). The input value of li was not specified in this
case and thus was based on the data provided by Lim et al. (1987). It should be taken into
consideration that such assumption might slightly affect the accuracy of results.
During the data validation it was discovered that the formula for cracking strength based on the
work of Ramaswamy et al. (1994) greatly underestimates the initial value of concrete strength.
In order to match the results closer, a different formula was used for the value of cracking
strength of concrete, in this case the most suitable formulation was one provided in the
European Code (EN 1992-1-1:2004, EC2 Section 3.1) which defines cracking strength for
2
concrete cylinder strength of 50 MPa as 𝑓𝑐𝑟 = 0.3𝑓′𝑐 3 .
Figure 3.6 Comparison of proposed model and test results for members with end-hooked fibers tested by
Susetyo (2009)
As can be seen in the Figure 3.6 the results obtained using the Matlab code almost perfectly
match the results provided by Lee et al. (2013). Overall the experimental results show slightly
higher tensile stress in the beginning of crack width propagation than the tensile stress obtained
using the SDEM, after reaching crack width of 1 mm the SDEM results show good agreement
with the experimental results.
3.2.3.3 Conclusion
Based on the results obtained using the experimental results from Susetyo (2009) and Lim et
al. (1987) it can be concluded that the developed code correctly simulates the tensile stress
behavior of the fibers.
40
3.2.4 Implementation of the SDEM into the Five-Spring Model

The SDEM was included in the model as a sixth spring that represents the shear strength
provided by the action of fibers and in the model is referred to with a symbol V F. The
contribution of the tensile stress in concrete due to tension softening effect has been neglected
in order to only account for the fiber action and thus the final tensile stress in the fibers is
described as 𝑓𝑓 = 𝑓𝑠𝑡 + 𝑓𝑒ℎ .
The shear strength provided by the action of fibers across the critical diagonal cracks is
described as follows:
𝑓𝑓 𝑏𝑑 (32)
𝑉𝐹 = cos(𝛽)
sin(𝛼1 )
Where the angle β represents the angle between the force provided by fibers Ff and the vertical
component of the force VF as illustrated in Figure 3.7. The maximum possible value for the
angle is α1 and the minimum is 0. A study has been conducted in order to define the most
suitable value of the angle β, where three different values have been considered.
Figure 3.7 Definition of angle β
The first assumption was the most conservative one and considered the angle β equal to α1. This
scenario assumes that the angle of the fibers stays the same throughout the whole process of
widening of the crack and thus only the vertical component of the force is being accounted for.
This consideration does not take into account the fact that as the diagonal crack keeps on
widening the part of the beam below the diagonal crack is sliding along the crack and the angle
β of the fibers gets lower the wider the crack gets. This is the reason that another scenario has
been taken into consideration. In this case the angle β is defined as follows:
𝑠 (33)
𝛽 = atan ( )
𝑤
Where s is the slip and w is the width of the crack and are defined as follows:
𝑠 = 𝛥𝑐 sin 𝛼1 (34)
𝑙𝑘 (35)
𝑤 = 𝜀𝑡,𝑎𝑣𝑔 + 𝛥𝑐 cos 𝛼1
2 sin 𝛼1
This scenario assumes that the force Ff is parallel to the relative displacement between the crack
faces. The last option taken into account is when the angle β is equal to zero, and thus the
41
resulting force consists only of the vertical component. An analysis was carried out to define
the most suitable scenario to include in the five-spring mode. Properties of the beam B4 from
the experimental study of Mansur and Ong (1991) were used for the analysis of the impact of
the changing angle. Predicted shear strength was obtained for changing a/d ratios and a curve
was generated. Same process was repeated for three different assumptions of angle β, resulting
in three different curves. Figure 3.8 shows the results of the analysis. The results obtained using
angle β equal to α1, and definition of the angle from equation (33), are overlapping, while β
equal to 0 provides slightly higher shear resistance.
Figure 3.8 Variation of the angle β given different a/d ratios
Figure 3.9 shows the results of the analysis performed on the beam B4 for different fiber volume
ratios. The process was same as for the Figure 3.8, where three different curves were created to
analyze the impact of changing angle β. Two of the curves are again overlapping, while the one
with β equal to 0 results in slightly higher predicted shear strength. From the results it can be
concluded, that the impact on the shear strength of the beam by using different angle β is
insignificant. The scenario with the simplest assumption, where β is equal to zero has been
implemented within the final model.
Figure 3.9 Variation of the angle β given different fiber volume ratios
42
In order to illustrate the additional force introduced into the five-spring model Figure 3.10
highlights effect of the 0.5% volume of fibers introduced into the sample beam previously
chosen to describe the five-spring model in the Chapter 2.3 of this thesis. The figure represents
the equilibrium of forces for ∆𝑐 equal to 5 mm. The force VF is indicated with the bold yellow
curve and the increased shear strength is depicted in red. It can be seen, that the main difference
caused by the addition of fibers is in the sum of shear forces 𝑉 = ∑ 𝑉𝑖 whose value has increased
by the shear force provided by the fibers. The value indicated by the dotted line represents the
value of the shear force before the fibers were added. The shear strength provided by fibers
shows a decreasing tendency for increasing tensile strain. Such phenomenon is caused by the
dependency of the shear strength on the width of the crack, where the width of the crack
increases with increasing average tensile strain.
Figure 3.10 Equilibrium of forces in extended five-spring model with fiber contribution at Δc=5 mm
Figure 3.11 shows the effect of fibers on the overall V-Δ response of the member. Thick yellow
curve represents the contribution of fibers and the red curve the overall response. The response
of the member before adding fibers is represented with red dotted line to highlight the difference
caused by addition of fibers. As it can be observed from the figure, the fibers start acting as
soon as the deformation begins. The biggest contribution of the fibers occurs at the deformation
of about 3 mm from which the curve starts slowly decreasing. The addition of 0.5% of fibers
provides the beam with an increase of about 200 kN in shear resistance and also higher post-
peak resistance.
43
Figure 3.11 Predicted V-Δ response with addition of fibers
44
3.3 TENSION-STIFFENING OF THE BOTTOM FLEXURAL REINFORCEMENT

3.3.1 Introduction
As it was described before, steel fibers are used in concrete members to compensate for the low
tensile strength and brittle response of the material. Given the constitutive equations for the
springs in the five-spring model, it is necessary to consider the effect of fibers on the tensile
force in the bottom reinforcement which is described in the original five-spring model by
Mihaylov (2015) for the cracked stage of the behavior of the member as:
0.33√𝑓′𝑐 (36)
𝑇 = 𝐸𝑠 𝐴𝑠 𝜀𝑡,𝑎𝑣𝑔 + 𝐴𝑐,𝑒𝑓𝑓 ≤ 𝐴𝑠 𝑓𝑦
√1 + 200𝜀𝑡,𝑎𝑣𝑔
Where the first term of this equation describes the behavior of bare elastic reinforcement and
the second term describes the tension stiffening effect of the concrete around the reinforcement.
However, the tension behavior of FRC members with conventional reinforcement (R/FRC) is
significantly different from the behavior of members without fibers. As described in a study by
Lee et al. (2013) the contribution of fibers to the tensile stress is considerable as illustrated in
Figure 3.12.
Figure 3.12 Tensile behavior of RC and R/FRC members by Lee et al. (2013)
As we can see in the Figure 3.12, R/FRC is able to resist higher tensile stresses not only after
initial cracking, but also after yielding of the bar. Due to this additional resistance the equation
for the tensile force in the bottom reinforcement will be modified based on the model proposed
by Lee et al. (2013) on tension-stiffening for steel fiber-reinforced concrete containing
conventional reinforcement. The model proposed by Lee et al. (2013) reflects the effect of steel
fibers on the tensile behavior of R/FRC member and is based on the DEM model whose
simplified version has been already described in the previous chapter of this thesis. Other
researchers have also researched the tension-stiffening behavior of R/FRC members, but such
45
models proposed by Bischoff (2003) or Deluce and Vecchio (2013) evaluate only the total
tensile stresses. The tensile stresses due to fiber and the bond mechanism between the concrete
matrix and the reinforcing bar cannot be separately evaluated in their models, while the model
developed by Lee et al. (2013) allows us to do so. In the model proposed by Lee et al. (2013)
the tensile resistance of the R/FRC member is divided into three different components which
are summed up in the end. The effect of concrete matrix, fibers and tension-stiffening or
softening are all evaluated separately, while in the other models only the total tensile resistance
is evaluated.
Figure 3.13a highlights tensile force T which is going to be affected by the addition of fibers.
Figure 3.13b illustrates the difference between the tensile force of the original five-spring
model, that does not include the effect of fibers, and the resulting tensile force with fibers
included in the concrete mixture causing tension-stiffening effect.
Figure 3.13 Illustration of implemented modification due to tension-stiffening effect
3.3.2 Formulation of Tension-Stiffening Model for R/FRC

The tensile stress resistance of R/FRC consists of three terms: tensile stress resistance of the
reinforcing bar, tensile stress resistance of the steel fibers, and tensile stress from tension-
stiffening or tension-softening effect, whichever is larger. The final formulation of the tensile
force in the bottom reinforcement including fibers can be described as follows:
𝑇𝑓 = 𝐹𝑟𝑒𝑏 + 𝐹𝑓𝑖𝑏 + 𝐹𝑐𝑜𝑛 (37)
Where Freb is the tensile force of the reinforcing bar, Ffib is the tensile force of the fibers, and
Fcon the tensile force from tension-stiffening or tension-softening effect. In this chapter all of
these expressions will be described and analyzed in detail, and then Tf will be used to calculate
the shear force derived from moment equilibrium in the five-spring model.
The first term of the equation describes the tensile force resisted by the bare bar, this force is
evaluated the same way it was described in the original five-spring model and thus can be
defined as:
𝐹𝑟𝑒𝑏 = 𝐸𝑠 𝐴𝑠 𝜀𝑡,𝑎𝑣𝑔 ≤ 𝐴𝑠 𝑓𝑦 (38)
46
The second term refers to the tensile force resisted by steel fibers and is expressed as follows:
𝐹𝑓𝑖𝑏 = 𝑓𝑓 𝐴𝑐,𝑒𝑓𝑓 (39)
Where Ac, eff is the area of concrete around the bottom reinforcement that contributes to the
tension-stiffening effect and can be calculated as:
𝐴𝑐,𝑒𝑓𝑓 = 2.5𝑏(ℎ − 𝑑) ≤ 𝑏ℎ/2 (40)
Because the tensile stresses sustained by the steel fibers are calculated for a given crack width
while the tensile stress due to tension stiffening is calculated for given average tensile strain, it
is necessary to define the relationship between the crack width and the average tensile strain.
Average crack spacing model has been defined based on the model proposed by Deluce (2011)
where average crack spacing is defined as:
𝑠𝑏 𝑘1 𝑘2 (41)
𝑠𝑐𝑟 = 2 (𝑐 + ) 𝑘3 +
10 𝑠𝑚𝑖
Where sb is the maximum spacing between reinforcing bars and the rest of the variables are
expressed as follows: 𝑐 = 1.5𝑎𝑔 𝑘1 = 0.4 𝑘2 = 0.25
Where ag is the maximum aggregate size.

𝑙𝑓
(42)
𝜌𝑠 𝛼𝑓 𝑉𝑓 𝑑𝑓
𝑠𝑚𝑖 = 𝑑 + 𝑚𝑎𝑥 ( 50 , 1)
𝑏 𝑑𝑓
(43)
min(𝑉𝑓 , 0.015) 50
𝑘3 = 1 − 1 − min ,1
0.015 𝑙𝑓
[ ( 𝑑𝑓 )]
Where db is the diameter of the reinforcing bar and ρs is the ratio of the reinforcement within
the concrete area Ac,eff. Having estimated the crack spacing, the width of the crack that will be
applied within the SDEM to define the tensile stress ff is defined as:
𝑤𝑐𝑟,𝑇 = 𝑠𝑐𝑟 𝜀𝑡,𝑎𝑣𝑔
The last term of the equation (37), to define the tensile force in the bottom reinforcement,
describes the tension-stiffening/tension-softening provided by the concrete and is expressed as
follows:
𝐹𝑐𝑜𝑛 = 𝑓𝑐𝑡 𝐴𝑐,𝑒𝑓𝑓 (44)
47
Where fct is the tension response due to bond mechanism of the reinforcing bar and is described
as:
𝑓𝑐𝑡 = max(𝑓𝑐,𝑠𝑜𝑓𝑡 , 𝑓𝑐,𝑇𝑆 ) ≤ 𝐸𝑐 𝜀𝑡,𝑎𝑣𝑔 (45)
Where the elastic modulus of concrete matrix Ec for fiber reinforced concrete is defined as 𝐸𝑐 =
3300√𝑓′𝑐 + 6900, as presented in the CSA A23.3-04 (2004) and fc, TS is the tension-stiffening
effect that reflects the effect of steel fibers and was derived by Lee et al. (2013). The model is
based on the conventional tension-stiffening model for reinforced concrete members by Bentz
(2005) and is modified to take into account the effect of steel fibers as follows:
𝑓𝑐𝑟 (46)
𝑓𝑐,𝑇𝑆 =
1 + √3.6𝑐𝑓 𝑀𝜀𝑡,𝑎𝑣𝑔
𝐴𝑐,𝑒𝑓𝑓
Where M is the bond parameter defined as 𝑀 = ∑ 𝑑𝑏 𝜋
and cf is the coefficient to consider the
effect of steel fibers which was defined based on a parametric study performed by Lee et al
(2013). Only the variables with significant influence on the tension-stiffening response have
been taken into account such as percentage of conventional reinforcement, fiber volumetric
ratio and fiber type. Based on the parametric study the following values for cf were proposed:
- for straight fibers

0.9
1 𝑙𝑓 100𝑉𝑓 (47)
𝑐𝑓 = 0.6 + ( )
0.058 𝑑𝑓 𝑀0.8
- for hook-ended fibers
1.5
1 𝑙 (100𝑉𝑓 ) (48)
𝑐𝑓 = 0.6 + 0.034 (𝑑𝑓 )
𝑓 𝑀0.8
3.3.3 Validation of the Tension-Stiffening Model for R/FRC

Based on the Tension-Stiffening model a code in Matlab was developed using the formulation
described in the previous section. The code to account for tension-stiffening within the five-
spring model can be found in Appendix C. The developed model will be validated against data
provided by Lee et al. (2013) and experimental results. The validation includes results from
application of the Tension-Stiffening model by Lee et al. (2013) and provided experimental
data by Bischoff (2003) and Deluce and Vecchio (2013).
3.3.3.1 Experimental results by Bischoff (2003)

In the test eight axially loaded tension specimens were tested. Each specimen had a cross-
section size of 100x100 mm. Two ratios of reinforcement were tested using either single 15 M
bar or 20 M bar with final reinforcement ratios of 2% and 3.1%, respectively. Members were
divided into two groups of four where for each reinforcement ratio were produced four
48
members. Three out of those four members were additionally reinforced with 0.78% fiber
volumetric ratio and one of the three specimens contained an un-bonded reinforcing bar.
Conclusion of the experiment by Bischoff (2003) is that after cracking, fiber reinforced
members experience more tension-stiffening than plain concrete. The difference becomes more
obvious once a sufficient number of cracks have developed. It has been also observed that while
plain concrete does not exhibit any tension-stiffening once the reinforcement yields, fiber
reinforced concrete continues to exhibit tension-stiffening after yielding since the fibers are
able to transmit additional tensile forces across the crack where the steel rebar has yielded.
For the validation only two fiber reinforced specimens were considered, each with different
amount of reinforcement ratio and both with bonded reinforcing bar. The fibers used in this
experiment were end-hooked fibers, 50 mm long with diameter of 0.5 mm. Shrinkage of the
member has been measured within the experiment and implemented into the code to reflect the
real behavior of the member.
Figure 3.14 Comparison of proposed models and test results by Bischoff (2003)
Figure 3.14 compares results of the experiment carried out by Bischoff (2003), results obtained
by Lee et al. using proposed tension-stiffening model and results obtained with the created
Matlab code. It can be observed that both models copy the behavior of the experiment quite
well. Slight difference in test BS15M between the analytical results by Lee et al. (2013) and
created code is most likely due to the simplification of the DEM used in the code. While Lee et
al. (2013) have applied the DEM within their method, the Matlab code defines stress attainable
by fibers using the SDEM, which does not account for fiber rupture as it has been explained
before. Thus it is possible that in the experiment with 2% reinforcement ratio some fiber rupture
already appears that the Matlab code cannot capture, while in the test BS20M thanks to the
3.1% reinforcement ratio no fiber rupture appears.
3.3.3.2 Experimental results by Deluce and Vecchio (2013)

Experimental investigation carried out by Deluce and Vecchio consisted of tests on total of 12
uniaxial tension reinforced concrete specimens and 48 uniaxial tension R/FRC specimens. In
49
this case, parameters studied within the experiment were the fiber volumetric ratio, fiber length,
fiber aspect ratio, conventional reinforcement ratio and reinforcing bar diameter. In his thesis
Deluce (2011) describes the tests in detail including numerous material tests. However, for the
purpose of validation only two tests were selected. Selected specimen had cross section of
150x150 mm, and concrete strength of 45 MPa and 80 MPa. Two different fiber volumetric
ratios were considered, 0.5% and 1% of fibers in the concrete matrix. Fibers used in this
experiment were end-hooked with length of 30 mm, diameter of fibers 0.38 mm and
conventional reinforcement with diameter 19.5 mm and reinforcement ratio of 1.33%.
Based on the experimental results Deluce and Vecchio (2013) concluded that steel fiber added
to the concrete mixture with conventional reinforcing bar improves the cracking characteristics
and tension-stiffening behavior and that they can increase the post-yield load-carrying capacity
to levels significantly higher than the bare-bar yield load. They also came to a conclusion that
fiber length does not appear to play a significant role in the post-cracking behavior of SFRC
containing conventional reinforcing bars, provided that the crack spacing is not so short that a
fiber bridges multiple cracks.
Due to the unclear definition of cracking strength of concrete, an assumption has been made in
order to match the curve of the proposed model. In this experiment the effect of shrinkage has
not been included within the final graph. As presented in Figure 3.15 the proposed model by
Lee et al. (2013) agrees very well with the results of Deluce and Vecchio and also with the
proposed Matlab code.
Figure 3.15 Comparison of proposed models and test results by Deluce and Vecchio (2013)
3.3.3.3 Conclusion
Given the experimental results of Bischoff (2003) and Deluce and Vecchio (2013) it can be
concluded that developed model simulates correctly the tension-stiffening behavior of R/FRC
members. Thanks to the Tension-Stiffening model derived by Lee et al. (2013) it is possible to
evaluate the tension force resistance of each of the considered parts of the member and thus the
impact of fibers, concrete matrix and conventional reinforcement can be evaluated separately
as shown in the Figure 3.16. A limitation that should be considered while implementing the
50
code is the same as for the SDEM and thus that fiber rupture should not occur and is not being
accounted for in the model.
Figure 3.16 Separate evaluation of contribution of tensile forces in Tension-stiffening model
3.3.4 Implementation of the Tension-Stiffening Model into the Five-Spring Model

Within the five-spring model, the Tension-Stiffening model has been included just as it has
been described in the formulation of the model. The only change concerns the definition of
tensile force contribution from the rebar, where the yielding of the bar was neglected in order
to focus on shear failures of deep beams.
After obtaining the tension force in the bottom flexural reinforcement using tension-stiffening
model, moment equilibrium equation of the shear span is used to define the shear force. The
equilibrium is calculated about the point of application of the compression force in the section
with maximum moment as presented by Mihaylov (2015), and the final equation to determine
the shear force can be described as follows:
𝑇𝑓 (0.9𝑑) (49)
𝑉=
𝑎
Where d is the effective depth and a is the shear span of the member, and V in this case
represents the force in the flexural spring of the five-spring model.
In order to illustrate the modification introduced into the five-spring model, the tension-
stiffening model was implemented into the sample beam example previously used in sections
2.3 and 3.2.4. In the Figure 3.17 the thick black curve represents the shear force derived from
the moment equilibrium, which is influenced by the tension stiffening model. The dotted curve
represents the original shear force derived from moment equilibrium. As it can be observed
from the figure, the tension-stiffening does not influence the result significantly. However, the
transition of the curve results in a change of equilibrium points obtained through intersection
of the thick black curve with the rest of the curves.
51
Figure 3.17 Equilibrium of forces in extended five-spring model with tension-stiffening model at Δc=5 mm
Figure 3.18 illustrates the effect that tension-stiffening has on the load displacement curve.
Slightly steeper orientation of the curve causes that forces influenced by strain in bottom
reinforcement change their values from the original model due to the transition of the
equilibrium points. The original values are depicted with dotted lines within the graph, while
the values corresponding to the tension-stiffening model and addition of fibers are illustrated
with solid lines. The tension-stiffening effect influences very slightly the values of most of the
shear forces. The only forces that are not affected are the critical loading zone, whose value is
consistent and does not depend on the strain in bottom flexural reinforcement, and the effect of
fibers. Overall, the effect of tension stiffening does not significantly influence the resistance of
the member unless a very high volume of fibers is included in the mixture.
1400
V
1200
1000
Shear Forces (kN)
800
600
VCLZ
400 Vci
VF Vs
200
Vd
0
0 5 10 15
Figure 3.18 Predicted V-Δ response with implementation of tension-stiffening model
52
3.4 COMPRESSIVE BEHAVIOR OF THE CRITICAL LOADING ZONE

3.4.1 Introduction
Apart from enhanced tensile strength that the addition of fibers provides significant increase in
toughness, in compression has been observed in fiber reinforced concrete members.
Traditionally, transverse reinforcement is adopted to improve the toughness of concrete in
compression. However, the use of steel fibers has advantages over transverse reinforcement,
namely less labor cost and shorter construction time. Steel fibers bridge longitudinal cracks
caused by the lateral expansion of concrete in compression. As the cracks widen, the pull-out
strength of fibers increases the toughness of the member as described by Ou et al. (2012).
Figure 3.19a illustrates the effect of fibers on compression behavior of the member within the
five-spring model. The resistance of the critical loading zone is increased due to the inclusion
of fibers and the ultimate shear strength increases. Figure 3.19b describes shear resisted by
critical loading based on transverse displacement of the critical loading zone, the solid line
describes behavior with fibers and the dotted line without fibers.
Figure 3.19 Illustration of modification due to change of stress-strain curve of concrete in compression
Many researchers have performed experiments in order to describe the compressive stress-
strain response of FRC. The most recognized formulation is by Di Prisco et al. (2013) where it
is concluded that when using FRC, compressive strength is not particularly influenced by the
presence of fibers up to 1% of content by volume. However, in the extended five-spring model,
fiber volumetric ratios higher than 1% are taken into account as well. It is necessary to adjust
the stress-strain curve to reflect the effect of fibers on the concrete under uniaxial compression
using a different model suitable for fiber volumes higher than 1%.
In the original model, the compressive behavior of concrete is described using Popovics (1970)
definition of stress-strain relationships for concrete as follows:
𝜀𝑐𝑓 (50)
𝑛 (𝜀 ′ )
𝑐
𝑓𝑐 = 𝑓′𝑐
𝜀𝑐𝑓 𝑛𝑘
𝑛 − 1 + (𝜀 ′ )
𝑐
53
Where f’c is the peak stress obtained from a cylinder test, ε’c, is the strain when the stress fc
reaches f’c and is described as:
𝑓′𝑐 𝑛 (51)
𝜀 ′𝑐 =
𝐸𝑐 𝑛 − 1
𝜀 𝜀
Factor k increases the post-peak decay in stress, for 𝜀𝑐𝑓 𝑐𝑓
′ < 1 k is taken as 1, and for 𝜀 ′ > 1 as:
𝑐 𝑐
𝑓′𝑐 (52)
𝑘 = 0.67 +
62
However, k must not be taken less than unity.
Curve fitting factor n is defined as:
𝑓′𝑐 (53)
𝑛 = 0.8 +
17
Tangent stiffness Ec, when εcf equals zero and described as:
- for normal strength concrete
𝐸𝑐 = 4730√𝑓′𝑐 (54)
- for concrete strength higher than 40 MPa
𝐸𝑐 = 3320√𝑓′𝑐 + 6900 (55)
3.4.2 Stress-Strain Curves for FRC

The next section evaluated the effect of steel fibers with fiber volumes higher than 1% on the
compressive behavior of FRC. Several proposals from various researchers are presented and
analyzed in order to choose the most adequate model that is to be used within the five-spring
model. As a comparison Popovics (1970) stress-strain curve was used to ensure correct behavior
of concrete in the pre-peak branch of the curve.
3.4.2.1 Barros and Figueiras (1999)

In the experimental study of Barros and Figueiras (1999) five series were manufactured where
fiber percentage, fiber aspect ratio and water-to-cement ratio were the parameters whose
influence was of the main focus. Cylinder specimens, 150 mm in diameter and 300 mm in
height with concrete strength varying from 30 – 60 MPa, were tested under displacement
control. Only two of the series were used to determine the compression behavior from which
each series consisted of four specimens. In those two series end-hooked fibers were used with
length of 60 mm, diameter equal to 0.8 mm and strength of 1100 MPa.
54
The analytical approach suggested by Barros and Figueiras is similar to Ezeldin and Balaguru
(1992) where only one parameter is proposed and the whole expression is based on the
following stress-strain relationship:
𝜀𝑐𝑓 (56)
𝜀′𝑐
𝜀 𝜀𝑐𝑓 (1−𝑞)/𝑝
(1 − 𝑝 − 𝑞) + 𝑞 ( 𝑐𝑓 ) + 𝑝 (
𝜀′𝑐 𝜀′𝑐 )
𝐸 1−𝑞
𝑞 = 1 − 𝑝 − 𝐸𝑐 𝑝 + 𝑞 ∈ ]0,1[ >0 (57)
𝑐𝑓 𝑝
𝑓 ′𝑐 (58)
𝐸𝑐𝑓 =
𝜀𝑐𝑓
1/3
𝑓 ′𝑐 (59)
𝐸𝑐 = 21500 ( )
10
Where for fibers with length of 30 mm and diameter 0.5 mm:
𝜀 ′ 𝑐 = 𝜀𝑐0 + 0.0002𝑊𝑓 (60)
𝑝 = 1 − 0.919𝑒 −0.394𝑊𝑓 (61)
𝜀𝑐0 = 0.0022 (62)
𝑊𝑓 = 325𝑉𝑓 (63)
And for fibers with length of 60 mm and diameter 0.8 mm:
𝜀 ′ 𝑐 = 𝜀𝑐0 + 0.00026𝑊𝑓 (64)
𝑝 = 1 − 0.722𝑒 −0.144𝑊𝑓 (65)
Where 𝜀𝑐0 is the strain at peak for plain concrete and 𝑊𝑓 is the fiber weight percentage in the
mixture. Proposed analytical approach has been validated against experimental results and
showed a good agreement. In order to evaluate the model, the model was applied on an example
specimen with concrete strength of 40 MPa, fiber volumetric ratio 1%, fiber length of 30 mm
and fiber diameter 0.5 mm and compared to results with curve proposed by Popovics (1970).
As it can be seen in the Figure 3.20 the strain at maximum concrete strength is higher which is
one of the effects that prove that fibers have influence on the compression behavior of concrete.
55
The effect of fibers on the post peak behavior is significant and instead of brittle failure which
appears in case of regular concrete, FRC exhibits more ductile behavior.
It can be concluded that model by Barros and Figueiras (1999) represents the stress-strain
compression behavior quite well, however it should be taken into account that only one fiber
aspect ratio was considered in the study, so it is questionable whether proposed model
reasonably represents the effect of aspect ratio of steel fibers on compressive behavior.
Figure 3.20 Analytical proposal of stress-strain compression curve by Barros and Figueiras (1999)
3.4.2.2 Lee, Oh and Cho (2015)

Experimental study of Lee et al. (2015) consisted of tests on 48 cylinder specimens subjected
to uniaxial compression. Specimens had a size of 150 mm in diameter and 300 mm in height.
Several variables were investigated including concrete compressive strength, fiber volumetric
ratio, and fiber aspect ratio. Fiber volumetric ratio was ranging from 0.5% to 2% and end-
hooked steel fibers of three different fiber aspect ratios (47%, 65%, 79%) were used. Concrete
strength varied from 40 MPa to 90 MPa.
Results of the experiments showed that the compressive strength increased slightly with
increasing fiber volumetric ratio up until 1.5% of fiber volume while with 2% of fiber volume
the compressive strength slightly decreased. This phenomenon has been previously described
by Ezeldin and Balaguru (1992) and is assumed to be caused by the transverse confinement
effect of the steel fibers which restrain the lateral expansion of SFRC specimens. Due to
observed changes in slump effect while changing the aspect ratio and fiber volumetric ratio it
was concluded that the compressive strength of SFRC can be affected by the fiber aspect ratio
as well as the fiber volumetric ratio.
In the analytical model the effect of steel fibers was represented by employing the fiber
reinforcing index 𝑅𝐼𝑣 = 𝑉𝑓 𝑙𝑓 /𝑑𝑓 which has been previously introduced by several researchers
56
such as Ezeldin and Balaguru (1992) and Mansur et al. (1999) and is implemented in the model
as follows:
𝜀𝑐𝑓 (66)
𝐴 𝜀′
𝑐
𝜀𝑐𝑓 𝐵
𝐴 − 1 + (𝜀 ′ )
𝑐
0.12
𝜀 ′ 𝑐 = (0.0003𝑅𝐼𝑤 + 0.0018)𝑓 ′ 𝑐 (67)
0.41
𝐸𝑐 = (−367𝑅𝐼𝑤 + 5520)𝑓′𝑐 (68)
𝑅𝐼𝑤 = 3.25𝑅𝐼𝑣 (69)

Where A and B are parameters to reflect the effect of the steel fibers on compressive behavior
and are defined as:
1 𝜀𝑐𝑓
𝐴=𝐵= 𝑓′ for ≤1 (70)
1− ′ 𝑐 𝜀′ 𝑐
𝜀 𝑐 𝐸𝑐
𝐴 = 1 + 0.723(𝑅𝐼𝑤 )−0.957 (71)
𝑓′ 0.064 𝜀𝑐𝑓 (72)

𝐵 = ( 50𝑐) [1 + 0.882(𝑅𝐼𝑤 )−0.882 ] ≥ 𝐴 for >1
𝜀′ 𝑐
Analytical model by Lee et al. (2015) has been implemented on hypothetical specimen
previously considered for Barros and Figueiras (1999) example and compared with Popovics
(1970) curve. As it can be seen on the Figure 3.21 the pre-peak behavior agrees well with all
the models but there is a considerable difference between the post-peak behavior of the curve
obtained by Barros and Figueiras (1999) which assumes much higher stresses in concrete than
Lee et al. (2015).
Figure 3.21 Analytical proposal of stress-strain compression curve by Lee et al. (2015)
57
3.4.2.3 Ou, Tsai, Liu and Chang (2012)

In the experimental study 40 specimens were tested with variation of diameter of the fiber, fiber
length and fiber volumetric ratio. Compressive strength of concrete was 40 MPa. Steel fibers
used in the study were end-hooked and had tensile strength of 1000 MPa. Maximum reinforcing
index taken under consideration was 1.7 due to reduced workability of the concrete mixture that
has been observed with higher reinforcing index. All the specimens had diameter of 150 mm
and height of 300 mm.
Based on the experiment it was concluded that since the modulus of elasticity fluctuated as the
reinforcing index increased, the reinforcing index has low impact on the value of modulus of
elasticity. For the compressive strength, little correlation with the reinforcing index was
observed in the pre-peak branch but for the post-peak it was confirmed that fibers significantly
contribute to the ductility of the material. Strain at the peak stress tends to increase with
reinforcing index based on the obtained results. It was also observed that the toughness limit of
the material was reached at a fiber volume fraction of approximately 2% for all steel fibers and
that long fibers (50 – 60 mm) outperformed short fibers (30 mm).
An analytical model developed by Ou et al. (2012) is based on a model proposed by Carreira

and Chu (1985) and predicts a slightly steeper descending branch at a reinforcing index of 1.7
than at a reinforcing index of 1.3. The proposal, given that 𝜀𝑐0 = 0.002 is following:
𝜀 ′ 𝑐 = 𝜀𝑐0 + 0.0007(𝑅𝐼𝑣 ) (73)
𝛽 = 0.71(𝑅𝐼𝑣 )2 − 2(𝑅𝐼𝑣 ) + 3.05 (74)
𝑓′𝑐𝑓 = 𝑓′𝑐 + 2.35(𝑅𝐼𝑣 ) (75)
𝜀𝑐𝑓 (76)
𝛽 (𝜀 ′ )
𝑐
𝑓𝑐 = 𝑓′𝑐𝑓
𝜀𝑐𝑓 𝛽
𝛽 − 1 + (𝜀 ′ )
𝑐
Proposed model has been also implemented on the same hypothetical specimen as previous
models and Figure 3.22 shows obtained results. It can be observed that model proposed by Ou
et al. (2012) shows very good agreement with model by Lee et al. (2015) however the concrete
compressive peak strength is assumed to be slightly higher and strain at peak a bit lower.
58
Figure 3.22 Analytical proposal of stress-strain compression curve by Ou et al. (2012)
3.4.2.4 Mansur, Chin and Wee (1999)

Experimental study of Mansur et al. (1999) investigated high-strength concrete members with
end-hooked fibers with 0.5 mm diameter and length of 30 mm. Fiber volumetric ratios that were
tested were 0.5%, 1% and 1.5%. In order to define the shape effect, three different shapes were
casted for the experiment. Tested specimens consisted of 100x200 mm cylinders, 100 mm cubes
and 100x100x200 prisms.
Based on results it was concluded, that for vertically casted prisms and for cylinders, inclusion
of fibers caused smaller initial tangent modulus while for horizontally casted prisms the fibers
have practically no influence on the initial tangent modulus. Regarding the shape effect it was
concluded that fiber volumetric ratio effect was found to be smaller for cylinders. Regarding
compressive strength of the material it was concluded that vertically cast specimens are likely
to exhibit higher strengths than those casted horizontally.
Proposed analytical model is following:

𝜀𝑐𝑓
𝛽( ′ ) 𝜀𝑐𝑓
(77)
𝜀 𝑐
𝑓𝑐 = 𝑓′𝑐 𝜀𝑐𝑓 𝛽
for 0≤ ≤1
𝜀′ 𝑐
𝛽−1+( ′ )
𝜀 𝑐
𝜀𝑐𝑓
𝑘1 𝛽( ′ ) 𝜀𝑐𝑓
(78)
𝜀 𝑐
𝑓𝑐 = 𝑓′𝑐 𝜀𝑐𝑓 𝑘2 𝛽
for 1≤ 𝜀′ 𝑐
𝑘1 𝛽−1+( ′ )
𝜀 𝑐
50 3 (79)
𝑘1 = ( ) [1 + 2.5(𝑅𝐼𝑣 )2.5 ]
𝑓′𝑐
59
50 1.3 (80)
𝑘2 = ( ) [1 − 0.11(𝑅𝐼𝑣 )−1.1 ]
𝑓′𝑐
1 (81)
𝐸𝑐 = (10300 − 400𝑉𝑓 )𝑓′𝑐 3
1
(82)
𝜀𝑐0 = (0.0005 + 00000072𝑅𝐼𝑣 )𝑓′𝑐 3
1 (83)
𝛽=
[1 − (𝑓′𝑐 /𝜀𝑐0 𝐸𝑐 )]
Figure 3.23 shows that for given hypothetical specimen the analytical prediction of Mansur et
al. (1999) agrees well with the model proposed by Popovics (1970) for the ascending branch.
For the post-peak behavior, we can see that there is significant increase in stresses as it is
expected due to presence of fibers.
Figure 3.23 Analytical proposal of stress-strain compression for high-strength concrete by Mansur et al.
(1999)
3.4.2.5 Conclusions
All of the previously mentioned models were compared and their limitations taken into account
while selecting the most suitable stress-strain curve for the five-spring model. In the proposal
by Barros and Figueiras (1999) the main problem was the limitation of the model to two specific
fiber types and only one fiber aspect ratio, which does not provide accurate simulation of
behavior of FRC for other fiber aspect ratios. It is also questionable whether the model estimates
the behavior of the FRC after it reaches the peak stress correctly. The proposal by Lee et al.
(2015) takes into account different fiber aspect ratios and is more conservative in the post-peak
branch but if low amounts of fiber volumetric ratio are being considered, the resulting curve
60
does not capture the behavior correctly. Also, the curve by Lee et al. (2015) cannot be applied
on the regular concrete without fiber inclusion.
Most suitable analytical method and method considered within the five-spring model is the
proposal by Ou et al. (2012). This method takes into account different fiber aspect ratios and
thus is not limited to one type of fibers, gives reasonable results for lower fiber volume ratios,
can also be used for regular concrete without fiber inclusion, has a more conservative
assumption for the post-peak behavior of the concrete than Barros and Figueiras (1999) and its
pre-peak behavior agrees well with behavior described by Popovics (1970). However, one
significant limitation has to be taken into account and that is the range of concrete strength
considered within the method that only covers concrete strengths of 30 – 60 MPa and thus
cannot be implemented for high-strength concretes.
In order to include high-strength concrete in the five-spring model, for concretes with strength
higher than 50 MPa the method proposed by Mansur et al. (1999) was implemented. This
method is suitable for different fiber aspect ratios and the assumption of post-peak behavior is
as conservative as the model by Ou et al. (2012). The limitation in this case is that it cannot be
used for regular concrete without fiber inclusion. In order to provide a solution for cases without
fibers, Popovics (1970) curve was implemented in the five-spring model to describe the
behavior of concrete under compression if no fibers are included.
3.4.3 Implementation of the Stress-Strain Curve into the Five-Spring Model

Adjusted stress-strain curve for FRC under uniaxial compression has been implemented in the
five-spring model based on the given conclusions. For deep beams without fiber reinforcement
and any concrete strength Popovics (1970) curve is used for the calculation of stress in the
critical loading zone. For regular strength concretes with inclusion of fibers, the proposal by
Ou et al. (2012) has been implemented and the definition of elastic modulus is taken according
to Lee et al. (2015). For high-strength concrete the five-spring model uses the definition of
stress-strain relationship by Mansur et al. (1999). The formulation of the methods used in the
model is exactly as mentioned in the description of each analytical method.
After obtaining the stress-strain curve the area under the curve is computed as follows:
𝜀
𝑓𝑐 (𝜀𝑖 ) + 𝑓𝑐 (𝜀𝑖−1 ) (84)
Ω(𝜀) = ∫ 𝑓𝑐 𝑑𝜀 ≈ ∑
0 2(𝜀𝑖 − 𝜀𝑖−1 )
𝑖
Where the right-hand-side of the equation is the numerical integration of the integral using the
trapezium rule. The strain axis is divided into small constant intervals ∆𝜀 = 𝜀𝑖 − 𝜀𝑖−1, and the
area under the fc curve within each interval is approximated as a trapezium.
Average diagonal compressive stress in the critical loading zone is computed:

𝛺 (85)
𝜎𝑎𝑣𝑔 = −
𝜀
61
Where the ε represents the diagonal compressive strains in the CLZ which increase with
increasing imposed shear distortion of the CLZ Δc. The code describing the calculation of the
𝜎𝑎𝑣𝑔 within the five-spring model can be found in Appendix D. Obtaining 𝜎𝑎𝑣𝑔 allows us to
compute the shear carried in the critical loading zone:
𝑉𝐶𝐿𝑍 = 𝜎𝑎𝑣𝑔 𝑏𝑙𝑏1𝑒 sin2 (𝛼) (86)
Figure 3.24 and 3.25 highlight the modification introduced into the five-spring model using a
stress-strain curve with consideration of fibers. Figure 2.24 illustrates the effect for given
imposed displacement. The dotted lines represent values obtained using the original five-spring
model and solid lines are values obtained from modified five-spring model. Thick blue line
represents increased value of shear force provided by the critical loading zone. The only curve
that is affected by the change of stress-strain response of FRC is the VCLZ. The value remains
constant, as it is not dependent on the strain in the bottom flexural reinforcement, however its
value increases significantly due to the action of fibers.
Figure 3.24 Equilibrium of forces in extended 5sm with adjusted stress-strain curve at Δc=5 mm
Figure 3.25 illustrates the influence of modification on the load-displacement curve. The only
curve affected by the modification is the VCLZ due to the shift of the curve representing shear
force provided by VCLZ showed in Figure 2.24. The original VCLZ curve is depicted with blue
dotted line while the modification with blue solid curve. As it is shown in the figure, adjusted
stress-strain curve has mainly impact on the post-peak behavior of the beam. Overall, the
inclusion of fibers results in less brittle behavior of the member in the post-peak branch of the
curve.
62
Figure 3.25 Predicted V-Δ response with modified stress-strain curve
63
3.5 DEFLECTIONS
3.5.1 Introduction
The original five-spring model only accounts for deflection in three-point loading where the
deflection of the beam is equal to deflection under the loading point. In the extension of the
model due to limited experimental results, it became one of the crucial points to take into
account also deflection in four-point loading. Experimental results by Mansur and Ong (1991),
which served for the validation of the extended five-spring model, tested all the members in
four-point loading. In order to be able to compare the maximum deflection obtained from the
five-spring model and the experiment, it was necessary to include the additional deflection in
the original five-spring model. While in the three-point loading the calculated deflection under
the loading point is equal to the midspan deflection in four-point loading an additional value
that accounts for the difference in deflection under the loading point and the midspan needs to
be introduced.
3.5.2 Definition of the Additional Deflection

Given the assumption that plane sections remain plane in bending, the value of the additional
deflection can be taken as a sum of two deflections:
𝛥1 = 𝛥11 + 𝛥12 (87)

Where:
𝜙𝑙𝑠 2 (88)
𝛥11 =
8
𝜙𝑙𝑠 (89)
𝛥12 = 𝑎
2
Where ls is the distance between the loading points, which in case of three-point loading is equal
to zero and the curvature φ is described as:
𝑉𝑎 (90)
𝜙=
𝐸𝑐 𝐼𝑐𝑟
So the final formulation of the additional deflection is:
𝑉𝑎 𝑙𝑠 2 𝑙𝑠 (91)
𝛥1 = ( + 𝑎)
𝐸𝑐 𝐼𝑐𝑟 8 2
Where the Icr is the moment of inertia of cracked section and for cases without compression
steel is calculated as follows:
𝑏𝑘 3 𝑑 3 (92)
𝐼𝑐𝑟 = + 𝑛𝐴𝑠 (𝑑 − 𝑘𝑑)2
3
64
And the coefficients are defined as follows:
𝐸𝑠 (93)
𝑛=
𝐸𝑐
𝑏 (94)
𝐵=
𝑛𝐴𝑠
𝑏ℎ3 (95)
𝐼𝑔 =
12
√2𝑑𝐵 + 1 − 1 (96)
𝑘𝑑 =
𝐵
Where Ec is the tangent modulus of fiber-reinforced concrete and Ig is the moment of inertia of
un-cracked section.
3.5.3 Implementation of the Additional Deflection into the Five-Spring Model

Within the extended five-spring model the additional deflection to account for four-point
loading is added to the sum of the deflection of the shear span due to elongation of the bottom
longitudinal reinforcement Δt and the shear distortion of critical loading zone Δc. Final
formulation of the overall deflection at the midspan is defined as follows:
𝛥 = 𝛥𝑐 + 𝛥𝑡 + 𝛥1 (97)
Where Δ is the overall deflection when defining the shear strength and displacement capacity
of deep beams under single curvature.
65
4 VALIDATION OF THE EXTENDED FIVE-SPRING MODEL
4.1 INTRODUCTION
Developed extended five-spring model as shown in Appendix E is validated against existing
experimental results. Two experimental studies were considered for the validation, one being
the study of FRC deep beams in shear by Mansur and Ong (1991) and another experimental
study by Mansur and Alwist (1984) on FRC deep beams with web openings where one of the
control specimens was considered without openings and thus relevant for the validation of the
extended five-spring model.
Validation consists of comparisons of the experimental results with extended five-spring model
and results obtained from a finite element model. Analyses of the effect of fiber volumetric
ratio and effect of shear span over effective height ratio were carried out, where four different
fiber volumetric ratios and five different a/d ratios were tested and compared within the
experimental study.
In the end, results of the comparisons are discussed and a validation of the model is provided
with corresponding figures. A verification for FRC beams with high shear reinforcement is
performed and relevant issues that arise with high shear reinforcement ratios are discussed and
explained.
4.2 SAMPLE SPECIMEN EVALUATION

In order to validate the extended five-spring model, one representative specimen was selected
from the experimental study of Mansur and Ong (1991) and analyzed in detail. Out of ten
specimen tested in the experimental study five had variety of shear span to depth ratios from
0.31 to 1.85. Four of the specimens were tested to analyze the effect of fiber volume ratio
varying from 0% to 2%. From these specimens the most representative one was selected being
the specimen B4 with 1% fiber volume ratio and a/d of 1.23.
Selected specimen is 500 mm high and has a width of 90 mm. Effective depth has been
calculated based on provided drawings of the setup to 463 mm. Length of the specimen is 1590
mm which gives a shear length of 570 mm. Concrete strength of the specimen is 31.1 MPa with
maximum aggregate size of 10 mm. Flexural reinforcement of the beam is provided by four
reinforcing bars of 16 mm in diameter, yield strength 440 MPa and reinforcement area of 804
mm2 which gives a reinforcement ratio of 1.93%. The beam has additional web reinforcement
provided by bars with diameter 6 mm and yield strength 375 MPa with spacing in transversal
direction of 110 mm and in longitudinal direction of 125 mm which gives a reinforcement ratio
of 0.27% and 0.47%, respectively. Fibers used in this experimental study were 30 mm long
straight fibers, slightly twisted and with square cross section and sides length of 0.5 mm. In
order to get equivalent diameter of fiber that is used in the extended five-spring model,
perimeter of the square was calculated from which the corresponding diameter was expressed
as 0.564 mm. Longitudinal length of loading plates is 80 mm and is equal to the support plates.
66
The beam is loaded in four-point loading where distance between the loading plates is equal to
200 mm.
Mansur and Ong (1991) provide a detailed description of behavior of the beam throughout the
experiment. According to Mansur and Ong (1991) in the early stages of loading the beam
behaved in a truly elastic manner giving a linear load-deflection curve. Diagonal crack within
the shear span was the first one to form and is described to appear approximately at mid-depth
of the beam with extension towards both the support and the loading points. Further increase in
load resulted in the propagation and widening of the existing cracks while simultaneously new
diagonal cracks developed more or less parallel to the existing ones. Some of the cracks
originated vertically, but later became inclined in a diagonal direction. Flexural cracks
propagated within the constant moment region of the beam but only after formation of diagonal
cracks and, according to Mansur and Ong (1991), they hardly reached the mid-depth of the
beam. At load ranging from 80% - 90% of the ultimate load, one of the diagonal cracks began
to grow excessively wide, finally leading to failure which for this particular beam was at load
of 228 kN. At impending failure, some crushing of the concrete was observed between the
loading points and the tip of the major inclined crack. The failure load of the beam and its
cracking patterns can be observed in Figure 4.1.
Figure 4.1 Cracking pattern and mode of failure of beam B4 by Mansur and Ong (1991)
All of the previously described properties of the beam were implemented into the extended five-
spring model and a load-deformation curve was generated, the curve can be seen in Figure 4.2.
Based on the detailed description of behavior of the beam throughout the experiment some
general assumptions of the five-spring model theory can be validated against real life
observations. One of the main assumptions of the five-spring theory is a formulation of a critical
diagonal crack. In this case the crack propagated as expected and also was the reason of the
failure of the beam as assumed in the model. Furthermore, the experiment description also
mentions the presence of crushing of the concrete between the loading points and the tip of the
major inclined crack which is in the five-spring model described as crushing of the critical
loading zone.
67
Figure 4.2 Comparison of load-deflection curves for beam B4 obtained from five-spring method and
experimental results by Mansur and Ong (1991)
As it can be seen in Figure 4.2, ultimate shear strength obtained from the five-spring method is
more conservative than the shear strength obtained in the test. Shear capacity obtained by the
five-spring method is 214 kN, which gives ratio between experimental result and prediction of
1.06. As it can be observed from Figure 4.2, deformations of the experimental results are
slightly higher than the five-spring predictions, which results in the deformation at a peak
strength of 4.5 mm for the test and only 2.8 mm for the five-spring model. Such difference can
be caused by the settlement of supports due to high compression forces applied to the beam
during the loading.
In order to evaluate the effect of support settlement and to define its impact on the overall
deflections of the beam, a finite element model (FEM) was created using a VecTor2 program
for 2D static and dynamic analysis of reinforced concrete structures. The analysis was carried
out by Jian Liu (2016) who provided the results for the comparisons. The sample specimen was
modeled in the software using same properties as the five-spring model and specimen B4 in the
experimental study. Force method using a load step of 2 kN was applied in order to define the
shear strength of the member. Resulting curve can be seen in Figure 4.3 which also includes
experimental and five-spring model results.
From Figure 4.3 we can observe that the shear resistance is slightly higher than the experimental
result. For the finite element model the failure occurs at shear force of 238 kN which gives ratio
of experimental shear resistance to FEM prediction of 0.96. For the FEM the shear failure
occurs at 3 mm deflection which supports the theory of overestimated deformation in the
experimental study and suggests that some settlement of supports, that might have not been
accounted for, takes place in the experimental study.
68
250
Maximum documented
200 shear strength
Shear force V (kN)
150
Finite element model
100 Extended five-spring model

Experimental data
50
0
0 1 2 3 4 5 6 7
Figure 4.3 Comparison of load-deflection curves for beam B4 obtained from finite element model, five-
spring method and experimental results by Mansur and Ong (1991)
Figure 4.4 pictures the cracking pattern of one half of the beam B4 right before the failure as
predicted by the FE model. The formation of diagonal cracks corresponds to the formation of
cracks as described in the experimental study, as well as the flexural cracks in the constant
moment region which only propagated to the mid-depth of the beam at the moment of the
failure. Additional cracking can be observed on the edge of the beam where horizontal cracks
started forming, as well as cracking on top of the beam with some vertical cracks, these cracks
do not copy the real cracking behavior of the beam as observed in the experimental study and
are most likely caused by incorrect simulation of cracking by the FE model.
Figure 4.4 Cracking pattern of the beam B4 obtained by the FE analysis
69
4.3 EFFECT OF FIBER VOLUMETRIC RATIO

In the experimental study, Mansur and Ong (1991) have focused on two main parameters one
of which was the effect of fiber volumetric ratio. Based on their study they concluded that
adding fibers to the concrete mixture significantly influenced cracking behavior of the beams.
The experiment showed that increase of fiber volume ratio decreased maximum crack width
and also average crack width at all load levels. Higher fiber volume ratio also resulted in smaller
concrete strains and smaller deflections at a particular load level. Increasing fiber content
increased both the cracking and ultimate strengths of the beams.
The effect of fiber volume ratio on the ultimate shear strength is shown in Figure 4.5. Apart
from the experimental results and VecTor2 results, Figure 4.5 shows also the effect of fiber
volume ratio according to the extended five-spring model. Separate curves represent the shear
strength provided by each of the springs of the five-spring model. The biggest shear resistance
is provided by the critical loading zone and the stirrups. For the extended five-spring model an
average compressive strength of 33.1 MPa was implemented to generate the average curve of
the behavior of the specimen, and different fiber volume ratios with step of 0.1% were applied
to generate resulting curve. Results show that the model created in VecTor2 overestimates the
ultimate shear strength at lower fiber volumes, however the extended five-spring model copies
the behavior of the beams quite accurately with comparison to the experimental results.
300
250 Vd + Vs + Vci + VCLZ + VF
200 Vd + Vs + Vci + VCLZ

V max (kN)
150
Vd + Vs + Vci
100
Vd + Vs
50 VecTor2
Vd
Experimental results
0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6
Vf (%)
Figure 4.5 Effect of fiber volume ratio on ultimate shear strength
4.4 EFFECT OF A/D RATIO

Second parameter analyzed in the experimental study of Mansur and Ong (1991) was shear-
span-to-depth (a/d) ratio. Five different a/d ratios were considered in the study of Mansur and
Ong (1991), varying from 0.31 to 1.85. Shear length of the members varied from 740 mm to
2140 mm and the concrete cylinder strength varied from 31.1 MPa to 35.7 MPa. All of the
70
tested specimen had the same longitudinal and transversal reinforcement ratios of 1.93% and
0.27%, respectively and fiber volume ratio of 1%.
Conclusions of the experimental study state that shortening of shear span reduced the
occurrence and extent of the flexural cracking. In the member with the highest a/d ratio of 1.85,
flexural cracks appeared first followed by diagonal cracking. In the beam with a/d ratio of 1.23,
both flexural and diagonal cracking occurred almost simultaneously and for the beam with a/d
ratio of 0.93, diagonal cracks were the first ones to form. In tested beams with ratios 0.31 and
0.62 no flexural cracking was observed up to failure. It was also concluded that an increase in
a/d ratio decreases the stiffness of the beam.
Within the extended five-spring model an average concrete cylinder strength of 33.9 MPa was
implemented to obtain the corresponding behavior of the members and gradually increasing a/d
value with step of 0.1 was implemented to generate the a/d ratio curve. Due to limited time,
only one member with a/d ratio of 1.23 was analyzed in VecTor2 software and included in the
comparison of the effect of a/d ratio, properties of the member were the same as the
corresponding specimen used in the experimental study.
400
VecTor2
350
300 Experimental results
250 Vs + VF + Vd + Vci + VCLZ Angle change

Vmax (kN)
200
Vs + VF + Vd + Vci
150
Vs + VF + Vd
100
50 Vs + VF Vs
0
0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
a/d
Figure 4.6 Effect a/d ratio on ultimate shear strength
Figure 4.6 shows results from the experimental study compared with one result from VecTor2
and curve generated using the extended five-spring model. Separate curves show the influence
of each of the springs of the five-spring model. From the graph we can observe that for a/d ratio
up to 1.6 the biggest contribution is caused by critical loading zone, while after a/d ratio of 1.6
the stirrups provide the most shear strength. Proposed analytical model copies the behavior of
the member well, however a sudden change of the slope of the curve occurs around a/d of 1.6.
The cause of such change is most likely the change of the angle of critical crack as defined in
the model, which changes when the beam is transitioning from deep beam to slender beam.
Highlighted on the curve is a point, where the angle of critical crack α1 changes its value from
71
α to θ, determined using the extended five-spring method. α is the angle of line connecting
inner end of support plate lb2 and outer end of the loading plate lb1 and is described as follows:
ℎ (98)
𝛼 = atan
𝑎𝑒𝑓𝑓
Where h is the height of the member and effective shear span aeff is defined as:
𝑎𝑒𝑓𝑓 = 𝑎𝑐1 + 𝑙𝑏1𝑒 (99)
𝑙𝑏1 𝑙𝑏2 (100)

𝑎𝑐1 = 𝑎 − +
2 2
𝑉 (101)
𝑙𝑏1𝑒 = max( 𝑙𝑏1 , 3𝑎𝑔 )
𝑃
Where ac1 is the distance between the facing edges of the loading plate and support plate and
lb1e is the effective width of loading plate parallel to longitudinal axis of member which is
defined as the maximum of either three times the size of aggregate or contribution of length of
the loading plate defined by the ratio of shear force to applied point load. The value of angle θ,
which is the angle of the shear cracks obtained from sectional shear models for slender beams,
is in this case simplified to the value of 35°, however can be defined using level three
approximation described in fib MC2010 (2013).
4.5 DISCUSSION
Finally, conclusions based on the results obtained from the performed analyses are made and a
discussion about the results is carried out. In order to verify the results obtained using the
extended five-spring theory, a load-displacement curve was plotted for four specimens from
experimental study of Mansur and Ong (1991) along with the curve obtained in the experiment.
Curve obtained using the extended five-spring model captures well the ultimate shear strength
of the model, however the deflections are smaller at a particular load level and the stiffness of
the model is overestimated. An explanation for such behavior could be the settlement of
supports, caused by the high compression force applied on the beam. Since the material of the
support plates and its stiffness was not defined in the article by Mansur and Ong (1991) and the
settlement of the supports was not measured to provide another explanation, finite element
analysis of the beams was performed.
Finite element analysis was performed by Jian Liu (2016) with the same properties as described
in the experiment and the obtained load-displacement curves were compared with the
experimental and analytical ones. Results obtained by using the finite element analysis captured
well the ultimate shear strength of the specimen according to the experiments, however the
deflections were again smaller at a particular load level and the stiffness of the model was
overestimated. The finite element analysis showed good agreement with the analytical results
and thus it can be concluded that the main reason for the difference in deflections and stiffness
is the settlement of supports.
72
Figure 4.7 Comparison of results obtained by FEM and five-spring model for different fiber volumetric
ratios
Figure 4.7 shows comparisons of results obtained from the experimental study performed by
Mansur and Ong (1991), extended five spring model and VecTor2 analysis for four different
specimens with varying fiber volume ratios, B4, B6, B7 and B8. The ocumented shear strength
depicted in the graphs with dotted line is the ultimate shear strength of the member as defined
by Mansur and Ong (1991).
Apart from the displayed comparisons, all the members from specimens from Mansur’s
experimental studies relevant for the extended five-spring model were analyzed and the ultimate
shear strength for each member was predicted. Table 4.1 shows most important properties of
each beam and ratio of experimental over predicted ultimate shear strength.
73
Table 4.1 Standard deviation and coefficient of variation of results obtained with extended five-spring
model
Beam a/d d Vf lf df ρI f c' ρv V exp V pred V exp

Name (mm) (%) (mm) (mm) (%) (MPa) (%) (kN) (kN) V pred
B1 0,36 463 1 30 0,56 1,93 35,7 0,42 375 350 1,07
B2 0,62 463 1 30 0,56 1,93 35,7 0,49 360 307 1,17
B3 0,93 463 1 30 0,56 1,93 35,5 0,48 291 262 1,11
B4 1,23 463 1 30 0,56 1,93 31,1 0,47 228 214 1,06
B5 1,85 463 1 30 0,56 1,93 31,5 0,49 183 177 1,04
B6 1,23 463 - - - 1,93 34,4 0,47 205 190 1,08
B7 1,23 463 0,5 30 0,56 1,93 33,8 0,47 220 205 1,07
B8 1,23 463 1,5 30 0,56 1,93 33,2 0,47 260 235 1,11
B9 1,51 463 1 30 0,56 1,93 29,5 0,48 224 195 1,15
B10 1,51 463 1 30 0,56 1,93 30,1 1,26 290 299 0,97
WO-1/1 0,58 624 1 30 0,40 0,81 40 0,43 345 355 0,97
Average: 1,07
Standard deviation: 0,064
Coefficient of variation: 0,060
Given results from Table 4.1 it can be concluded that the shear strength obtained using the
extended five-spring model shows good accuracy when compared to the experimental results.
With coefficient of variation of 6% it can be concluded that the effect of fibers has been well
represented within the extended five-spring model, and the five-spring model is now able to
capture not only behavior of regular deep beams, but also fiber reinforced concrete deep beams.
74
5 PARAMETRIC STUDY
5.1 INTRODUCTION
To study the effects of several variables, a parametric study was carried out using the extended
five-spring model. First a sample beam was created, properties of the beam were based on an
experimental beam from tests by Mihaylov (2008). Properties of the sample beam as used in
the study are described in Table 5.1. The beam is considered to be loaded in three-point loading
and reinforced with straight fibers.
Table 5.1 Properties of a sample beam for parametric study
a/d b d h l
As ⌀I ρI
No. fy
2
(mm) (mm) (mm) (mm) (%) (mm ) (mm) barsI (MPa)
1,55 400 1095 1200 3900 0,70 3060 25 6 650
Vf lf df l b1 l b2 ag fc' fyv ρv ⌀v
(%) (mm) (mm) (mm) (mm) (mm) (MPa) (MPa) (%) (mm)
1 30 0,5 300 150 20 40 490 0,1 10
Variables investigated in the parametric study were the effect of a/d ratio, effect of longitudinal
reinforcement, shear reinforcement, fiber volume ratio and size effect. Each of the investigated
variables was adjusted within each parametric study, with the rest of the properties remaining
as described in the Table 5.1. Some of the parameters were evaluated for beams with no shear
reinforcement, most of the studies were carried out for three different fiber volume ratios of
0%, 1% and 2%. The last study investigated the effectiveness of fiber and shear reinforcement.
The impact of each parameter included within the study is evaluated and described in this
chapter.
5.2 PARAMETRIC STUDY

5.2.1 a/d Ratio
First, the effect of the shear-span-to-depth ratio was evaluated. The range of the ratio taken into
account varied from 0.5 to 2.5 a/d. The effect was investigated for fiber volume ratios of 0%,
1% and 2%. For each of the fiber volume ratios a curve was created by generating the shear
capacity for given a/d ratio where the step of 0.1 was used. Sample beam with no shear
reinforcement was considered in this part of study to focus on the effect of fibers.
The results of the study are shown in Figure 5.1. Horizontal axis describes the values of a/d
ratio, and the vertical axis represents the shear strength of the beam for given a/d ratio. Three
different lines represent three different fiber volume ratios. The black dots mark the point of
change of the angle α1 from α to θ as explained previously in Chapter 2.2. From the graph we
75
can observe gradual decrease of shear strength for increasing a/d ratios, where for the shortest
considered ratio the shear resistance has the highest value. For the highest a/d ratio the shear
strength of the beam shows to be the lowest. The shear strength resisted by the beam between
the ratios 0.5 and 2.5 dropped by 78% for the beam with no fiber reinforcement, by 73% for
1% of fibers, and 69% for 2% of fiber volume ratio.
Figure 5.1 Effect of a/d ratio for different fiber volume ratios
The effect of fiber volume ratio on the shear strength was evaluated next. Comparison between
the curves with different fiber volume ratios in Figure 5.1 shows a significant increase in shear
strength for fiber volume ratio of 1% and 2% compared to the beam with no fibers included.
The shear strength at 0.5 a/d ratio has increased by 21% for 1% fiber volume ratio compared to
the beam with no fibers, and by almost 30% for 2% fiber volume ratio. However the difference
between the shear resistance for a/d ratio of 2.5 is slightly more significant. The beam with 1%
fiber volume ratio shows increase of 32% in the shear strength compared to deep beam with no
fiber inclusion, and 2% fiber volume ratio shows and increase of 48% in the shear strength.
Figure 5.2 Breakdown of effect of a/d ratio for Vf = 1%
76
Figure 5.2 shows the breakdown of the forces contributing to the shear resistance for 1% fiber
reinforcement ratio. The decreasing tendency of the curve is mainly dependent on the
contribution from the critical loading zone, whose size decreases as the a/d ratio gets higher.
The aggregate interlock also decreases with increasing a/d ratio, however not as significantly
as the critical loading zone. The effect of fibers slightly reduces the decreasing tendency of the
curve, where for higher a/d ratios the contribution of fibers is more significant than for lower
ones.
In conclusion, the shear force resisted by the beam has desceding tendencies for increasing a/d
ratios. The inclusion of fibers provides additional shear strength resistance, and the shear
strength provided by fibers increases with higher fiber volume ratios.
5.2.2 Longitudinal Reinforcement
Next parameter investigated within the parametric study was the bottom longitudinal
reinforcement and its effect on the shear strength of the beam. Reinforcement ratios taken into
account in this case varied from 0.5% to 3%. The effect was investigated for three different
fiber volume ratios of 0%, 1% and 2%. For each fiber volume ratio a curve was created by
generating the shear strength resisted by the beam at a step of 0.1%. To focus on the effect of
fibers no shear reinforcement was considered in this study.
Figure 5.3 Effect of longitudinal reinforcement for different fiber volume ratios
Figure 5.3 shows results of the study. Horizontal axis describes different ratios of shear
reinforcement taken into account. Vertical axis shows the shear strength achieved for each
reinforcement ratio. Three different lines represent different fiber volume ratios. The tendency
of the graph shows linear increase in shear strength resistance for increasing ratio of
longitudinal reinforcement. The increase of shear strength between the ratios 0.5% and 3% for
beam without fibers was by 107%, for 1% fiber volume ratio by 77% and for 2% fiber volume
77
ratio by 64%. Thus the results show that bottom longitudinal reinforcement contributes greatly
to the resistance of the deep beams.
Inclusion of fibers shows again a significant increase in shear capacity. From Figure 5.3 we can
observe that the contribution of fibers is a constant value added to the overall shear strength of
the member. The increase in shear strength for beam with 1% of fiber volume ratio and
longitudinal reinforcement ratio of 0.5% is by 42% compared to beam with no fiber
reinforcement. The beam with fiber reinforcement of 2% has shear strength higher by 78% than
beam with no fiber reinforcement. For higher longitudinal reinforcement the percentage of
increase in overall shear strength decreases and for 1% fiber volume ratio is only 22% higher.
For fiber volume ratio of 2% the shear strength increases by 42%.
Figure 5.4 Breakdown of effect of longitudinal reinforcement for Vf = 1%
Figure 5.4 shows the breakdown of the forces contributing to the shear resistance of the beam
for 1% fiber reinforcement ratio. As it can be seen in the figure the contribution of fibers and
the critical loading zone is constant. Biggest contribution to the shear resistance comes from
the critical loading zone, and depending on the reinforcement ratio also from the dowel action
for higher ratios. The main reason for the increase in the shear strength is the contribution from
the dowel action which increases with increasing reinforcement ratio.
It can be concluded, that higher longitudinal reinforcement ratios provide higher shear
resistance of the beam. The inclusion of fibers increases the shear resistance of the beam, where
the contribution of fibers highly depends on the fiber volume ratio.
78
5.2.3 Fiber Volumetric Ratio
The effect of fiber volumetric ratio was investigated in each of the parametric studies. However,
a separate analysis focused on the effect of fibers was also carried out. The shear resistance of
the beam was evaluated for fiber volumes varying from 0% to 2.5%. Three curves were
generated for three different shear reinforcement ratios of 0%, 0.5% and 1% and curves were
created by generating the shear strength at a step of fiber volume ratio of 0.1%. The rest of the
properties were as proposed for the sample beam in the introduction of this chapter.
Figure 5.5 shows the results of the study, where the horizontal axis represents different fiber
volume ratios, and vertical axis the variation in the shear strength of the beam. Each of the three
different lines represents one shear reinforcement ratio. As it was previously observed in the
parametric studies for a/d ratio and longitudinal reinforcement, the inclusion of fibers increases
the capacity. The resistance of the beam linearly increases with increasing fiber volume ratio.
For beams with no shear or fiber reinforcement the increase of shear resistance compared to the
beam with 2.5% volume of fibers is by 91%. For 1% volume of fibers the shear resistance
increases by 38% compared to the beam with no fibers. When shear reinforcement is 0.5% the
resistance of the beam with 1% fiber volume ratio increases by 46%, and for 1% shear and fiber
reinforcement by 29% compared to no fiber reinforcement.
Figure 5.5 Effect of fiber volumetric ratio for different shear reinforcement ratios
The impact of shear reinforcement was also evaluated within the study. Increase in shear
reinforcement caused significant increase in shear resistance of the beam. Increase in shear
resistance for 0.5% shear reinforcement and 1% fiber volume is almost by 70%. For 1% shear
reinforcement and 1% fiber volume ratio the increase in shear strength is by 140%.
79
Figure 5.6 Breakdown of effect of fiber volumetric ratio for ρv = 0.5%
Figure 5.6 shows the breakdown of forces contributing to the shear strength for 0.5% shear
reinforcement. As it can be seen in the figure the main contribution is caused the shear
reinforcement. The effect of fibers increases with increasing fiber volumetric ratio.
In conclusion, the inclusion of fibers provides significant increase in shear resistance for deep
beams. However, shear reinforcement such as stirrups seems to contribute to the overall
strength more than the fibers.
5.2.4 Shear Reinforcement
Next study focused on the effect of shear reinforcement on the shear strength of the beam. Shear
reinforcement ratio considered in the study varied from 0% to 0.7%. As in the previous studies,
the effect was evaluated for three different fiber volume ratios of 0%, 1% and 2%. The curves
were created by generating the shear strength at a step of 0.1%. The rest of the properties of the
beam remained as mentioned in the introduction of this chapter.
Figure 5.7 represents the outcome of the study. Horizontal axis consists of different shear
reinforcement ratios, and vertical axis describes the shear strength achieved. Three different
curves represent three different fiber volume ratios. The study shows that for increasing shear
reinforcement ratio the shear strength linearly increases. The increase in the shear strength is
most significant for case with no fiber reinforcement where the difference between the shear
strength with 0% shear reinforcement and 0.7% is by 134%. The increase in shear strength for
1% fiber reinforcement is 98% and for 2% fiber reinforcement ratio is 76%.
The inclusion of fibers provides additional shear strength to the beam. A constant value of shear
force is added to the original shear strength depending on the amount of fibers added. Beams
with no shear reinforcement to 1% fiber reinforcement increased in the shear strength by 39%
and with 2% of fiber reinforcement ratio by 75%. For members with 0.7% shear reinforcement
80
ratio the shear strength increased less significantly by 11% for 1% fiber volume ratio and by
20% for 2% fiber volume ratio.
Figure 5.7 Effect of shear reinforcement for different fiber volume ratios
Figure 5.8 shows the breakdown of the effect of shear reinforcement. As shown in the figure,
the biggest contribution to the shear resistance comes from the critical loading zone or shear
reinforcement depending on the reinforcement ratio. The force due to the shear reinforcement
linearly increases as the ratio of reinforcement increases.
Figure 5.8 Breakdown of effect of shear reinforcement for Vf = 1%
5.2.4.1 FRC with High Shear Reinforcement
While generating results to obtain the ultimate shear strength using the extended five-spring
model, it was observed that in some cases the resulting curve started at a nonzero deformation
value as shown on Figure 5.9. Such phenomenon was observed for example in specimen B10
from the experimental study of Mansur and Ong (1991) where the beam B10 was included in
81
the program to study effects of shear reinforcement on the ultimate shear strength. Unlike other
beams that contained shear reinforcement around 0.47%, beam B10 had percentage of shear
reinforcement of 1.26% along with fiber reinforcement ratio of 1%. Such high reinforcement
ratio resulted in the phenomenon depicted in Figure 5.9 where the resistance provided by the
stirrups is so high that the model is not able to capture the behavior of the curve from the
beginning but rather from a certain value of deflection.
When provided with large amount of flexural reinforcement, the beams start to propagate
flexural deformations before the shear deformations. In this particular case, the transverse
displacement of critical loading zone does not start propagating only after the deflection of the
shear span due to elongation of bottom reinforcement has reached 2.9 mm. Due to the high
shear resistance of the beam the deflection due to transverse displacement is restrained by action
of stirrups and fibers in the beam and thus only flexural deflections of the shear span propagate
at the beginning of the loading of the beam.
Figure 5.9 Results for B10 obtained with the extended five-spring model
Figure 5.10 shows the average tensile strength to shear force at deflection of the beam equal to
zero for two different shear force reinforcement ratios of 0.47% and 1.26%, respectively. As
shown in the figures, resulting shear force curve intersects the flexural spring force curve only
once for 0.47% shear reinforcement ratio. However, due to significant increase in shear force
provided by stirrups the curve representing the sum of forces, consisting of 1% fiber
reinforcement ratio and 1.26% shear reinforcement ratio, intersects the flexural spring curve
three times.
Five-spring model defines the equilibrium of forces using a bisectional method where it
searches for a value of the intersection iteratively. Only one correct solution is assumed to exist
by the model and thus causing the phenomenon showed in Figure 5.9 for beams with high shear
reinforcement. In highly reinforced members, for fixed Δ equal to zero, the multiple intersection
of the flexural spring force is causing the model to define the equilibrium at an intersection
82
point with the highest shear force instead of zero where the equilibrium is found for members
with regular shear reinforcement as shown in Figure 5.10.
Figure 5.10 Average tensile strain at deflection equal to zero for different shear reinforcement ratios
With the provided explanation an assumption can be made regarding the missing part of the
plot for a given load-deflection curve. A linear growth of the curves can be assumed to model
the behavior of the beam before the propagation of transverse displacement of the critical
loading zone as shown in Figure 5.11. However, the assumption is only a rough prediction of
the behavior of the curve and does not necessarily represent the real behavior of the specimen.
Figure 5.11 Results for B10 obtained with the extended five-spring model with completed curve
5.2.5 Comparison of Effectiveness of Fibers vs. Shear Reinforcement
Previously, the effects of fiber volume ratio and shear reinforcement were separately evaluated.
It was concluded that both of the parameters contribute significantly to the shear strength of the
83
beam. However, another parametric study was carried out to determine which of the two
reinforcements is more efficient. Firstly, the shear strength was obtained considering shear
reinforcement ratio of 0.2% and no fiber reinforcement for different a/d ratios. Then, for each
step of a/d ratio a fiber volume ratio was specified to match the shear strength obtained using
0.2% of shear reinforcement. For this case a beam with no shear reinforcement was considered.
The results then show the amount of fibers necessary to match the shear strength achieved with
0.2% shear reinforcement for different a/d ratios.
Figure 5.12 Variation of Vmax for different a/d ratios with ρv = 0.2%
Figure 5.12 illustrates the shear strength for different a/d ratios for beam with 0.2% of shear
reinforcement. The curve was created by generating the shear strength for different a/d ratios
at a step of 0.1. The black dot represents the angle change as previously explained in Chapter
5.2.1. The same curve was then generated for beam with certain ratio of fiber reinforcement
and 0% shear reinforcement. However, to generate such curve, fiber volume ratio that provides
the beam with equal shear strength as 0.2% shear reinforcement had to be found.
Figure 5.13 Fiber volume ratio equivalent to 0.2% shear reinforcement ratio
84
Figure 5.13 shows the fiber volume ratio equivalent to the shear strength obtained by using
0.2% shear reinforcement for different a/d ratios. Horizontal axis shows the range of a/d ratios
taken into consideration, vertical axis describes the fiber volume ratio equivalent to 0.2% of
shear reinforcement.
The outcome of the study shows, that the fiber reinforcement is more effective than shear
reinforcement only for a/d ratios under 0.8. This is due to the effective length of shear span
considered within the equation that defines the contribution of shear strength provided by
stirrups. To obtain the effective shear length, the lengths l0 and 1.5lb1e are subtracted from
expression 𝑑𝑐𝑜𝑡𝛼1 to neglect the stirrups too close to the edges of the shear span. Unfortunately,
for smaller shear spans such subtraction greatly reduces the amount of stirrups contributing to
the shear strength. Such reduction does not apply to the contribution of fibers and thus the fibers
are able to provide more shear resistance than the shear reinforcement in these cases. For higher
a/d ratios than 0.8 however, the shear reinforcement shows to be more efficient than fibers.
5.2.6 Size Effect
Last parameter investigated within the parametric study was the size effect. The size effect was
examined for effective depth ranging from 700 mm to 2700 mm. For the study a member with
only fiber reinforcement and without additional shear reinforcement was considered. Three
different fiber volume ratios of 0%, 1% and 2% were considered. Figure 5.14 shows the results
of the study. Horizontal axis shows different effective depths and vertical axis describes the
shear strength divided by the effective cross section of the beam. Three different lines represent
three different fiber volume ratios.
Figure 5.14 Size effect for different fiber volume ratios
85
From the Figure 5.14 it can be observed that for beams with no fiber reinforcement the size
effect is very small. The decrease in stress resisted by the cross section is only 10% between
the and minimum effective depth considered. For 1% fiber reinforcement ratio the effect
becomes more significant, the size effect decreases by 24%, and for 2% fiber volumetric ratio
by 31%. Thus it can be concluded, that the size effect gets bigger for members with higher fiber
reinforcement ratio.
Figure 5.15 Size effect breakdown for 0% fiber volume ratio
Figure 5.15 represents a breakdown of the size effect for members with 0% fiber reinforcement
ratio. The figure shows a breakdown of the size effect to contribution of separate shear forces.
Because in this case no additional shear reinforcement was considered, the graph only consists
of three shear forces contributing to the stress resisted by the cross section. As it is shown in
the figure, the dowel action and the effect of the critical loading zone have almost no impact on
the size effect. However, the aggregate interlock contribution is influenced by the size of the
member the most.
Figure 5.16 shows the breakdown of size effect for member with 2% fiber volume ratio. Just as
for the breakdown of member with 0% fiber volume ratio, this member does not include any
shear reinforcement. Figure 5.16 shows four different forces contributing to the stress resisted
by the cross section. The dowel action in this case does not contribute much to the size effect.
The effect of aggregate interlock decreases with increasing effective height. However, the effect
of the critical loading zone is more significant than the one of aggregate interlock. Biggest
contribution to the size effect is provided by the fibers, which is depicted in the figure with
yellow line. The difference between the contribution to the size effect by the critical loading
zone for beam with 0% fibers and one with 2% is caused by the redefined stress-strain curve.
The stress-strain curve definition used in the Figure 5.16 is affected by inclusion fibers, while
the one in Figure 5.15 is not.
86
Figure 5.16 Size effect breakdown for 2% fiber volume ratio
It can be concluded, that the size effect mainly depends on the fiber volume ratio. While for
beams with no fiber reinforcement the size effect is insignificant, for beams with increasing
fiber volume ratios the effect becomes bigger and more significant.
5.3 DISCUSSION
Based on the results of the parametric study several conclusions can be made. Regarding the
effect of a/d ratio, the study has shown that for increasing a/d the shear strength becomes
smaller. This is caused mainly by the reduction of the size of the critical loading zone which
has shown to have significant impact on the shear resistance of the beam. The effect of bottom
longitudinal reinforcement has shown to improve the shear strength of the beam with increasing
reinforcement ratio. Higher ratio of flexural reinforcement increases the resistance of the beam,
and with addition of fibers the tension stiffening effect increases the resistance even more.
Shear reinforcement such as stirrups and its impact on the shear strength was investigated in
three of the studies. Firstly, shear reinforcement and its effect on the shear strength was
evaluated as a separate parameter. In this study, a linear increase of shear resistance was
observed for increasing shear reinforcement ratios. The effect of shear reinforcement was also
included in the study of fiber reinforcement ratio as well as its impact on the maxim shear
strength of the beam. Again, a significant contribution of the shear reinforcement to the shear
strength was observed. Lastly, the efficiency of shear reinforcement was compared with the
efficiency of the fiber reinforcement. It was concluded, that for beams with a/d ratio higher than
0.8 the shear reinforcement is more effective than fiber reinforcement. For lower a/d ratios, the
effect of stirrups was lower due to the limitations of the definition of effective area for shear
reinforcement considered to contribute to the shear resistance.
87
Fiber volume ratio and its effect on the shear resistance was evaluated separately and then also
in each of the studies. It can be concluded, that fibers provide additional shear strength to the
beam and with increasing amount of fibers the contribution becomes more significant.
However, based on the study of comparing the effectiveness of fiber and shear reinforcement,
fibers have shown to provide less shear strength to the beam.
Last parameter investigated was the size effect. It was concluded that while for beams with no
fiber or shear reinforcement the size in effect is not significant, the addition of fibers notably
affects the size effect. It was observed that for increasing fiber reinforcement ratios, the size
effect becomes more significant than for beams with no fiber reinforcement.
88
6 SUMMARY AND CONCLUSIONS

First, all data available on tests performed on fiber reinforced concrete deep beams was
collected. Based on a literature review, several experimental studies were found. However, only
the study by Mansur and Ong (1991) and Mansur and Alwist (1984) were summarized and
placed in a test database due to limitations such as missing data or unsuitable experimental
setup in other studies. All the tests showed that the addition of fibers increases the shear
resistance of deep beams. The effects of fiber volumetric ratio, a/d ratio, shear and additional
longitudinal reinforcement ratios were evaluated in the experimental study by Mansur and Ong
(1991), where for increasing fiber volume ratio a higher shear resistance was observed. For
increasing a/d ratio, the shear strength showed a significant decrease. The shear reinforcement
enhances the shear strength significantly, while the additional longitudinal reinforcement shows
to have a negligible effect on the shear resistance.
The main goal of the thesis was to propose an extension to an existing five-spring model (5sm)
for the shear behavior of deep beams (Mihaylov, 2015) to include the effects of steel fibers in
such members. A two-parameter kinematic theory proposed by Mihaylov et al. (2013), on
which the five-spring model is based, was analyzed in detail and explained before an extension
of the model was implemented.
The extension of the five-spring model accounts for three effects caused by steel fibers in the
concrete. To represent the shear resistance provided by the fibers across the critical shear crack,
the SDEM model proposed by Lee et al. (2013) was selected. The SDEM model is a simplified
version of DEM model introduced by Lee et al. (2011), and can be used for straight or hooked-
end fibers. To model the tension-stiffening effect of the fibers around the bottom longitudinal
reinforcement, a tension-stiffening model by Lee et al. (2013) was used. The model takes into
account the tension resistance of the bottom longitudinal reinforcing bar, the effect of fibers
and tension-stiffening or softening effect whichever is higher. The last adjustment introduced
into the model regarding the effect of fibers was the stress-strain curve to represent the
compressive behavior of the critical loading zones in deep beams. The original code considered
Popovics (1970) stress-strain curve for the behavior of concrete under uniaxial compression.
However, extensive research on stress-strain behavior of FRC has shown that, while the
inclusion of fibers does not significantly affect the pre-peak part of the stress-strain response, it
greatly enhances the post-peak response compared to regular concrete. A model proposed by
Ou et al (2012) was used to describe the stress-strain response of regular strength concrete,
while for high-strength FRC the approach by Mansur et al. (1999) was used. Finally, an
extension of the Matlab code of the five-spring model was introduced to account for deflections
in members under four-point loading.
The proposed extended five-spring model was validated against the collected experimental data.
The model showed good agreement regarding the shear strength of the test specimens. Yet the
deflections at the maximum predicted shear force (displacement capacity) were only about 50%
of the deflections measured in the tests. A possible explanation of such difference was
89
considered to be the settlement of the supports that was possibly unaccounted for in the
experimental study. To confirm this assumption, four of the beams from the experimental study
of Mansur and Ong (1991) were analyzed by Liu (2016) using non-linear finite element models,
and the load-deflection curves obtained from the models were compared with the experimental
and analytical results. The results of the analyses supported the assumption of support
settlements, and it was concluded that they were the main cause of the difference in deflections.
Finally, the measured shear strength of the FRC deep beams from the database was compared
with the predictions of the extended 5sm. With an average experimental-to-predicted ratio of
1.07 and a coefficient of variation of 6%, the proposed model was shown to predict the shear
strength of FRC deep beams with adequate accuracy.
With the validated model, a parametric study was carried out to study the effect of a/d ratio,
shear and longitudinal reinforcement ratio, fiber volumetric ratio, and the size effect in shear.
As it was previously concluded by Mansur and Ong (1991), the 5sm shows that the shear
strength decreases with increasing a/d ratio. For the fiber volumetric ratio and shear
reinforcement, the same conclusion was reached as by Mansur and Ong (1991) that for
increasing fiber volumetric ratio and shear reinforcement ratio the shear resistance of the beam
increases. A comparison of both types of shear reinforcement was carried out for different a/d
ratios where the stirrups showed to be more efficient than equal amounts of fiber reinforcement
for a/d ratios higher than 0.8. The effect of bottom longitudinal reinforcement was also
evaluated where it was concluded that higher reinforcement ratios provide higher shear strength
to the member. The last parameter evaluated was the size of the member (size effect in shear).
Three different fiber volumetric ratios were considered in the size effect series. For beams
without fibers, the size effect was negligible, while for increasing fiber volume ratios the effect
became more significant.
Finally, there are several limitations of the proposed extended 5sm model for FRC deep beams
that should be considered for future research. It should be taken into account that the SDEM
does not account for fiber rupture, and therefore when using hooked-ended fibers with low
strength, the model can overestimate the shear strength. Second, the lack of experimental data
did not allow an extensive validation of the model. The database created for the validation
should be extended as new tests become available to validate the model against more
experimental results. Also, the impact of fibers on the post-peak behavior was not evaluated
due to the lack of tests in which the post-peak response was measured.
90
REFERENCES
ACI Committee 544, [1993]. “Guide for Specifying, Proportioning, Mixing, Placing, and
Finishing Steel Fiber Reinforced Concrete,” ACI Materials Journal, Vol. 90, No. 1, pp. 94-
101.
ACI Committee 544, [2008]. “State-of-the-Art Report on Fiber Reinforced Concrete - ACI
544.1R-96 (Reapproved 2002),” ACI Manual of Concrete Practice, Part 6, pp. ACI544.1R-
7 - ACI544.1R-24.
Barros, J. A. O., Figueiras, J. A., [1999] “Flexural behavior of SFRC: Testing and modeling,”
Journal of Materials in Civil Engineering, ASCE, Vol. 11, No.4, pp. 331–339.
Bentz, E. C., [2005] “Explaining the Riddle of Tension Stiffening Models for Shear Panel
Experiments,” Journal of Structural Engineering, ASCE, Vol. 131, No. 9, pp. 1422-1425.
Bischoff, P. H., [2003] “Tension Stiffening and Cracking of Steel Fiber-Reinforced Concrete,”
Journal of Materials in Civil Engineering, ASCE, Vol. 15, No. 2, pp. 174-182.
Campione, G., [2012] “Flexural Behavior of Steel Fibrous Reinforced Concrete Deep Beams,”
Journal of Materials in Civil Engineering, ASCE, Vol. 138, No. 2, pp. 235 - 246.
Carreira, D. J., and Chu, K. H., [1985] “Stress-Strain Relationship for Plain Concrete in
Compression,” ACI Structural Journal, Vol. 82, No. 6, pp. 797-804.
CSA A23.3-04, [2004] “Design of Concrete Structures,” Canadian Standards Association,
Mississauga, ON, Canada, 214 p.
Deluce, J. R., [2011] “Cracking Behavior of Steel Fiber Reinforced Concrete Containing
Conventional Steel Reinforcement,” Master Thesis, Department of Civil Engineering,
University of Toronto, Toronto, ON, Canada, 308 p.
Deluce, J. R., Vecchio, F. J., [2013] “Cracking of SFRC Members Containing Conventional
Reinforcement,” ACI Structural Journal, Vol. 110, No. 3, pp. 481-490.
Di Prisco, M., Colombo, M., Dozio., D., [2013] “Fiber-reinforced concrete in fib Model Code
2010: principles, models and test validation,” Structural Concrete, Vol. 14, No. 4, pp. 342-
361.
European Committee for Standardization, CEN, EN 1992-1-1 [2004] Eurocode 2: Design of
Concrete Structures- Part 1-1: General Rules and Rules for Buildings, Brussels, Belgium,
225 pp.
Ezeldin, A. S., and Balaguru, P. N., [1992] “Normal- and High-Strength Fiber-Reinforced
Concrete under Compression,” Journal of Materials in Civil Engineering, ASCE, Vol. 4,
No. 4, pp. 415–429.
fib – International Federation for Structural Concrete, [2013] “fib Model Code for Concrete
structures 2010,” Berlin: Verlag Ernst & Sohn.
91
Hooke, R., [1678] “Lectures de Potentia Restitutiva (Spring Explaining the Power of Springing
Bodies),” printed for John Martyn Printer to The Royal Society, at the Bell in St. Paul’s
Church-Yard, 24 p.
Hsu, L.S., Hsu, C.T.T., [1994] “Stress-strain Behavior of Steel-Fiber High-Strength Concrete
under Compression,” ACI Structural Journal, Vol. 91, No. 4, pp. 448-457.
Johnston, C. D., [2001]. “Fiber-Reinforced Cements and Concretes”, Gordon and Breach
Science Publishers, Ottawa, Canada, 372 p.
Lee, S.-C., Cho, J.-Y., Vecchio, F. J., [2011] “Diverse Embedment Model for Steel Fiber-
Reinforced Concrete in Tension: Model Verification,” ACI Materials Journal, Vol. 108,
No. 5, pp. 526-535.
Lee, S.-C., Cho, J.-Y., Vecchio, F. J., [2011] “Diverse Embedment Model for Steel Fiber-
Reinforced Concrete in Tension: Model Development,” ACI Materials Journal, Vol. 108,
No. 5, pp. 516-525.
Lee, S.-C., Cho, J.-Y., Vecchio, F. J., [2013] “Simplified Diverse Embedment Model for Steel
Fiber-Reinforced Concrete Elements in Tension,” ACI Structural Journal, Vol. 110, No.
4, pp. 403-412.
Lee, S.-C., Cho, J.-Y., Vecchio, F. J., [2013] “Tension-Stiffening Model for Steel Fiber-
Reinforced Concrete Containing Conventional Reinforcement,” ACI Structural Journal,
Vol. 110, No. 4, pp. 639-648.
Lee, S.-C., Cho, J.-Y., Vecchio, F. J., [2016] “Analysis of Steel Fiber-Reinforced Concrete
Elements Subjected to Shear,” ACI Structural Journal, Vol. 113, No. 2, pp. 275-285.
Lee, S.-C., Oh, J.-H., and Cho, J.-Y., [2015] “Compressive Behavior of Fiber-Reinforced
Concrete with End-Hooked Steel Fibers,” Materials, Vol. 8, pp. 1442-1458.
Li, B., Maekawa, K., Okamura, H., [1989] “Contact Density Model for Stress Transfer Across
Cracks in Concrete,” Journal of the faculty of engineering, University of Tokyo (B), Vol.
40, No. 1, pp. 9–52.
Lim, T. Y., Paramasivam, P., Lee, S. L., [1987] “Analytical Model for Tensile Behavior of
Steel-Fiber Concrete,” ACI Materials Journal, Vol. 84, No. 4, pp. 286-298.
Liu, J., [2016] “Personal communication,” November 21st 2016.
Mansur, M. A., Alwist, W. A. M., [1984] “Reinforced Fiber Concrete Deep Beams with Web
Openings,” The International Journal of Cement Composites and Lightweight Concrete,
Vol. 6, No. 4, pp. 263-271.
Mansur, M. A., Chin, M. S., Wee, T. H., [1999] “Stress-Strain Relationship of High-Strength
Fiber Concrete in Compression,” Journal of Materials in Civil Engineering, ASCE, Vol.
11, No. 1, pp. 21–29.
92
Mansur, M. A., Ong, K. C. G., [1991] “Behavior of Reinforced Fiber Concrete Deep Beams in
Shear,” ACI Structural Journal, Vol. 88, No. 1, pp. 98-105.
Mihaylov, B. I., [2008] “Behavior of Deep Reinforced Concrete Beams under Monotonic and
Reversed Cyclic Load,” doctoral thesis, European School for Advanced Studies in
Reduction of Seismic Risk, Pavia, Italy, 2008, 379 pp.
Mihaylov, B. I., [2015] “Five-Spring Model for Complete Shear Behavior of Deep Beams,
“Structural Concrete, Vol. 16, No.1, pp. 71-83.
Mihaylov, B. I., Bentz, E. C., Collins, M. P., [2013] “Two-Parameter Kinematic Theory for
Shear Behavior of Deep Beams,” ACI Structural Journal, Vol. 110, No. 3, pp. 447-455.
Naaman, A. E., and Najm, H., [1991] “Bond-Slip Mechanisms of Steel Fibers in Concrete,”
ACI Materials Journal, Vol. 88, No. 2, pp. 135-145.
Narayan, R., Darwish., I. Y. S., [1988] “Fiber Concrete Deep Beams in Shear,” ACI Structural
Journal, Vol. 85, No. 2, pp. 141 - 149.
Newman, J., Choo, B. S., [2013] “Advanced Concrete Technology 3: Processes,” Elsevier Ltd.,
Oxford, Great Britain, 704 p.
Ou, Y.-C., Tsai, M.-S., Liu, K.-Y., Chang, K.-C., [2012] “Compressive Behavior of Steel-Fiber-
Reinforced Concrete with a High Reinforcing Index,” Journal of Materials in Civil
Engineering, ASCE, Vol. 24, No. 2, pp. 207-215.
Popovics, S., [1970] “A Review of Stress-Strain Relationships for Concrete,” ACI Journal, Vol.
67, No. 3, pp. 243-248.
Ramaswamy, A., Barzegar, F., Voyiadjis, G. Z., [1994] “Post-cracking Formulation for
Analysis of RC Structures Based on Secant Stiffness,” Journal of Engineering
Mechanics, ASCE, Vol. 120, No. 12, pp. 2621-2640.
Shah, D. U., Schubel, P. J., Licence, P., Clifford, M. J., [2012] “Determining the Minimum,
Critical and Maximum Fiber Content for Twisted Yarn Reinforced Plant Fiber
Composites,” Composites Science and Technology, Vol. 72, No. 15, pp. 1909-1917.
Susetyo, J., [2009] “Fiber Reinforcement for Shrinkage Crack Control in Pre-stressed, Precast
Segmental Bridges,” Doctoral Thesis, Department of Civil Engineering, University of
Toronto, Toronto, ON, Canada, 307 p.
Voo, J. Y. L., Foster, S. J., [2003] “Variable Engagement Model for Fiber-Reinforced Concrete
in Tension,” Uniciv Report No. R-420, University of New South Wales, School of Civil
and Environmental Engineering, 86 p.
Zollo, R. F., [1997] “Fiber-reinforced Concrete: An Overview after 30 Years of Development,”
Cement and Concrete Composites, Vol. 19, pp. 107-122.
93
No. Ref. No. Beam a/d b d h l a:M/V Vf lf li df Fiber l b1
Name (mm) (mm) (mm) (mm) (mm) (%) (mm) (mm) (mm) type (mm)
1 1 B1 0,31 90 463 500 740 145 1 30 - 0,564 S 80
2 1 B2 0,62 90 463 500 1020 285 1 30 - 0,564 S 80
3 1 B3 0,93 90 463 500 1310 430 1 30 - 0,564 S 80
4 1 B4 1,23 90 463 500 1590 570 1 30 - 0,564 S 80

5 1 B5 1,85 90 463 500 2160 855 1 30 - 0,564 S 80
6 1 B6 1,85 90 463 500 1590 570 0 0 - - - 80

7 1 B7 1,23 90 463 500 1590 570 0,5 30 - 0,564 S 80
8 1 B8 1,23 90 463 500 1590 570 1,5 30 - 0,564 S 80
Ref. No 1 – Mansur and Ong (1991)
Ref. No 2 – Mansur and Alwist (1984)
9 1 B9 1,51 90 463 500 1850 700 1 30 - 0,564 S 80

10 1 B10 1,51 90 463 500 1850 700 1 30 - 0,564 S 80
11 2 WO-1/1 0,58 80 624 650 1550 360 1 30 22 0,4 EH 100
94
No. V/P ⌀I No. ⌀h No.
l b2 ls ρI As fy ag fc' fyh ρh fyv ρv ⌀bv No. V u

2
(mm) (mm) (%) (mm ) (mm) barsI (MPa) (mm) (MPa) (MPa) (%) (mm) barsh (MPa) (%) (mm) barsv (kN)
1 80 1 200 1,93 804,25 16 4 440 10 35,7 375 0,54 6 4 375 0,42 6 5 375
2 80 1 200 1,93 804,25 16 4 440 10 35,7 375 0,54 6 4 375 0,49 6 8 360
3 80 1 200 1,93 804,25 16 4 440 10 35,5 375 0,54 6 4 375 0,48 6 10 291
4 80 1 200 1,93 804,25 16 4 440 10 31,1 375 0,54 6 4 375 0,47 6 12 228
5 80 1 200 1,93 804,25 16 4 440 10 31,5 375 0,54 6 4 375 0,49 6 17 183
APPENDIX A: TEST DATABASE
6 80 1 200 1,93 804,25 16 4 440 10 34,4 375 0,54 6 4 375 0,47 6 12 205
7 80 1 200 1,93 804,25 16 4 440 10 33,8 375 0,54 6 4 375 0,47 6 12 220
8 80 1 200 1,93 804,25 16 4 440 10 33,2 375 0,54 6 4 375 0,47 6 12 260
9 80 1 200 1,93 804,25 16 4 440 10 29,5 375 1,22 6 9 375 0,48 6 14 224
10 80 1 200 1,93 804,25 16 4 440 10 30,1 375 0,54 6 4 375 1,26 6 37 290
11 100 1 580 0,81 402,12 16 2 418 10 40 304 0,206 3,3 12 304 0,43 3,3 31 345
APPENDIX B: SDEM - MATLAB CODE

function [ff]=SDEM(wcr,fc,Vf,lf,lh,li,df)
taufmax = 0.396*sqrt(fc);
tauehmax = 0.429*sqrt(fc);
betaf = 0.67;
betaeh = 0.76;
sf = 0.01;
seh = 0.1;
Kehi = 1+(1+(7*betaeh/15-1)*sqrt(seh/(lf-li/2))-(2*(sqrt(lf-li/2))-
sqrt(seh))^2)/(lf-li);
if wcr < sf
Kst = betaf*wcr/3/sf;
else
Kst = 1-sqrt(sf/wcr)+betaf/3*sqrt(sf/wcr);
end
if wcr < seh

Keh = betaeh*(2*wcr/3/seh-1/5*(wcr/seh)^2);
elseif seh <= wcr < lf-li/2
Keh = 1+(7*betaeh/15-1)*sqrt(seh/wcr)-(2*(sqrt(wcr)-
sqrt(seh))^2)/(lf-li);
elseif lf-li/2 <= wcr < li/2
Keh = (li-2*wcr/(2*li-lf))^2*Kehi;
else
Keh = 0;
end
alphaf = 0.5;
fst = alphaf*Vf.*Kst.*taufmax.*lf./df.*(1-2.*wcr./lf).^2;
if lh >= 1
feh = alphaf*Vf.*Keh.*tauehmax.*2.*(li-2.*wcr)/df;
else
feh = 0;
end
ff = feh+fst;
return
95
APPENDIX C: TENSION-STIFFENING - MATLAB CODE

function [N]=TS_R_FRC_S(Aceff,Ec,e,ag,Vf,lf,df,db,sb,fc,nb,lh,li,Er,As)
rs=As/Aceff;
c=1.5*ag;
k1=0.4;
k2=0.25;
k3=1-(min(Vf,0.015)/0.015)*(1-min(50/(lf/df),1));
smi=rs/db+0.5*Vf/df*(max((lf/df)/50,1));
scr=2*(c+sb/10)*k3+k1*k2/smi;
fcr=0.33*sqrt(fc);
M=Aceff/(nb*db*pi);
if lh<=0
cf=0.6+1/0.058*(lf/df)^0.9*((100*Vf)/M^0.8);
else
cf=0.6+1/0.034*(lf/df)*((100*Vf)^1.5/M^0.8);
end
wcri=scr*e;
[ff]=SDEM(wcri,fc,Vf,lf,lh,li,df);
fcsofti=fcr*exp(-15*wcri);
fcTSi=fcr/(1+(3.6*cf*M*e)^0.5);
fctj=min(Ec*e,max(fcsofti,fcTSi));
freb=min(Er*As*e);
ffib=min(ff*Aceff);
fcon=fctj*Aceff;
N=(freb+ffib+fcon);
return
96
APPENDIX D: AVERAGE DIAGONAL STRESS IN THE

CRITICAL LOADING ZONE - MATLAB CODE
function [favg,Ec,Dcenv]=FAVG(Vf,lf,df,fc,alfa,lb1e)
e=-(0:0.0001:1.0001)'*0.3;
RIv=Vf*lf/df;
if Vf<=0;
n=0.8+fc/17;
kk=0.67+fc/62;
if fc<=41
Ec=4730*fc^0.5;
else
Ec=(3320*fc^0.5+6900);
end
ec=-fc/Ec*n/(n-1);
for i=1:length(e)
f(i,1)=fc*n*e(i)/ec./(n-1+(e(i)/ec).^(n*kk));
end
else
if fc<=60
Ec=(-367*Vf*lf/df+5520)*fc^(0.41);
fcf=fc+2.35*RIv;
eco=0.002;
ecf=-(eco+0.0007*RIv);
beta=0.71*RIv^2-2*RIv+3.05;
for i=1:length(e)
f(i,1)=fcf*beta*(e(i)/ecf)/(beta-1+(e(i)/ecf)^beta);
end
else
eo=-((0.0005+0.00000072*RIv)*fc^0.35);
Ec=(10300-400*Vf)*fc^(1/3);
beta=1/(1-(fc/(-eo*Ec)));
k1=(50/fc)^3*(1+2.5*RIv^2.5);
k2=(50/fc)^1.3*(1-0.11*RIv^(-1.1));
for i=1:length(e)
if e(i)/eo <= 1
f(i,1)=fc*beta*(e(i)/eo)/(beta-1+(e(i)/eo)^beta);
else
f(i,1)=fc*k1*beta*(e(i)/eo)/(k1*beta-1+(e(i)/eo)^(k2*beta));
end
end
end
end
Om(1,1)=0;
for i=2:10002
Om(i,1)=Om(i-1)-(f(i)+f(i-1))/2*(e(i)-e(i-1));
end
Dcenv=-3*lb1e*e/tan(alfa);
favg=-Om./e;
end
97
APPENDIX E: EXTENDED FIVE-SPRING MODEL - MATLAB

CODE
%% Input
kci=0.18;
[fc, h, d, a, b, As, rs, sb, fy, Er, ls, Vf, lf, lh, li, df, fyf, nb, ag,
rv, fyv, Ev, lb1, lb2, VoP, Id, Vexp]=BeamData('B1');
PDexp1=B1;
Dc=(0:0.1:15)';
%% Geometry
lb1e=max(VoP*lb1,3*ag);
acl=a-lb1/2-lb2/2;
acalc=a-lb1/2+lb1e/2;
aeff=acl+lb1e;
alfa=atan(h/aeff);
alfa1=max(alfa,35/180*pi);
db=(As/nb*4/pi)^0.5;
rl=100*As/b/d;
scr=0.28*db*2.5*(h-d)/(rl/100)/d;
l0=max(scr,1.5*(h-d)/tan(alfa1));
l0=min(l0,d/tan(alfa1)/2);
lk=l0+min(l0,d*(1/tan(alfa)-1/tan(alfa1)));
lt=d/tan(alfa1)+(lk-l0);
Av=max(min(rv/100,0.15*fc/fyv)*b*(d/tan(alfa1)-l0-1.5*lb1e),0);
Aceff=b*min(2.5*(h-d),h/2);
z=0.9*d;
%% CLZ
[favg,Ec,Dcenv]=FAVG(Vf,lf,df,fc,alfa,lb1e);
favg(1)=0;
Vclzenv=favg*b*lb1e*sin(alfa)^2;
[Vclzmax, row]=max(Vclzenv);
Dcmax=Dcenv(row);
Kclz0=Vclzenv(2)/Dcenv(2);
%% Bottom reinforcement
kt=1;
%% Solution
LS=size(Dc,1);
for i=1:LS
etl=0;
etr=20/100;
for j=1:100
etj=(etl+etr)/2;
etminj=kt*etj;
etmaxj=(etj*lt-etminj*lk)/(lt-lk);
98
Dtj=etj*lt/d*a;
% CLZ
Vclzj=interp1(Dcenv,Vclzenv,Dc(i),'linear');
% Stirrups
evj=(Dc(i)+0.25*etj*d/(tan(alfa1))^2)/(0.5*z);
fvj=min(Ev*evj,fyv);
Vsj=Av*fvj;
% Shear derived from moment equilibrium
[N]=TS_R_FRC_S(Aceff,Ec,etmaxj,ag,Vf,lf,df,db,sb,fc,nb,lh,li,Er,As);
Vtj=N*z/acalc;
etminj=N/Er/As;
% Dowel action
fyej=fy*max((1-(Er*etminj/fy)^2),0);
Vdj=min(12*Er*Id/lk^3*Dc(i),nb*fyej*db^3/3/lk);
% Aggregate interlock
sj=Dc(i)*sin(alfa1);
wj=Dc(i)*cos(alfa1)+0.5*etminj*lk/sin(alfa1);
[ncij,vcij]=CDMfunc(fc,ag,sj,wj);
Vcij=kci*vcij*b*d;
% Fiber action
[ff]=SDEM(wj,fc,Vf,lf,lh,li,df);
beta=0;
VFj=ff*b*d/sin(alfa1);
% Shear resistance across diagonal crack

Vccj=Vclzj+Vcij+Vsj+Vdj+VFj;
% Error in equilibrium condition Vtj=Vccj

erj=100*abs((Vtj-Vccj)/Vccj);
if erj<0.01
break
end
% Additional deformation to account for four-point loading

n=Er/Ec;
B=b/(n*As);
Ig=b*h^3/12;
kd=(sqrt(2*d*B+1)-1)/B;
Icr=b*kd^3/3+n*As*(d-kd)^2;
D11ij=(Vtj*a)/(Ec*Icr)*((ls^2/8)+(ls/2*a));
% Adjustment of boundaries for bisection method

if Vtj<Vccj
etl=etj;
else
etr=etj;
end
end
er(i,1)=erj;
99
V(i,1)=Vtj/1000;
Vclz(i,1)=Vclzj/1000;
VF(i,1)=VFj/1000;
Vci(i,1)=Vcij/1000;
Vs(i,1)=Vsj/1000;
Vd(i,1)=Vdj/1000;
Dt(i,1)=Dtj;
Vcc(i,1)=Vccj;
if Vtj<0;
LS=i;
break
end
end
etmax=et*(lt-kt*lk)/(lt-lk);
Dc=Dc(1:i);
D=Dt+Dc+D11;
[Vmax, rvmax]=max(V);
VFmax=VF(rvmax);
Vcimax=Vci(rvmax);
Vclzmax=Vclz(rvmax);
Vsmax=Vs(rvmax);
Vdmax=Vd(rvmax);
Dcu=Dc(rvmax);
% max error
max(abs(er))
100

Five Spring Model

Uploaded by

Copyright:

Available Formats

Five Spring Model

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Five Spring Model

Uploaded by

Copyright:

Available Formats

Two-Parameter Kinematic Approach for the Shear

Behavior of Deep Beams Made of Fiber-Reinforced

Author: Karolína Tvrzníková

Supervisor: Prof. Boyan Mihaylov

University: University of Liège

University: University of Liège

3.4.1 Introduction ........................................................................................................ 53

Table 5.1 Properties of a sample beam for parametric study ................................................... 75

⌀h = Diameter of web reinforcement

fc = Local stress of concrete

lb2 = Width of support plate parallel to longitudinal axis of member

RIv = Fiber reinforcing index based on fiber volumetric ratio

vf = Vertical component of stress resisted by fibers

ρh = Ratio of longitudinal web reinforcement

Figure 1.1 Application of deep beam in a building (Mihaylov et al.,2013)

1.2 SCOPE AND OBJECTIVES OF THE THESIS

1.3 THESIS OUTLINE

2.1 FIBER-REINFORCED CONCRETE (FRC)

2.1.1 Properties and Classification of FRC

Figure 2.2 Typical load-CMOD curve from fib MC2010 (2013)

2.1.2 Factors Affecting Behavior of FRC

2.2 TESTS ON FRC DEEP BEAMS

2.2.1 Mansur and Ong (1991)

Figure 2.3 Experimental setup by Mansur and Ong (1991)

Figure 2.4 Experimental results by Mansur and Ong (1991)

2.2.2 Mansur and Alwist (1984)

Figure 2.6 Experimental setup by Mansur and Alwist (1984)

2.3 TWO-PARAMETER KINEMATIC THEORY AND FIVE-SPRING MODEL FOR

Figure 2.8 Main assumptions

Figure 2.9 Degrees of freedom of kinematic model (Mihaylov et al.,2013)

- for points below the critical diagonal crack

𝛿𝑧 (𝑥, 𝑧) = 𝜀𝑡,𝑎𝑣𝑔 𝑥cot𝛼 + ∆𝑐 (4)

𝑙𝑘 = 𝑙0 + 𝑑(𝑐𝑜𝑡𝛼 − 𝑐𝑜𝑡𝛼1 ) ≤ 2𝑙0 (5)

𝑙0 = 1.5(ℎ − 𝑑)𝑐𝑜𝑡𝛼1 ≥ 𝑠𝑐𝑟 (6)

0.28𝑑𝑏 2.5(ℎ − 𝑑) (8)

Figure 2.11 Five-spring model (Mihaylov, 2015)

Figure 2.12 Free-body diagram of the rigid block (Mihaylov, 2015)

Figure 2.13 Critical loading zone (Mihaylov, 2015)

𝑉𝐶𝐿𝑍 = 𝜎𝑎𝑣𝑔 [𝜀𝑚𝑎𝑥 (𝛥𝑐 )]𝑏𝑙𝑏1𝑒 𝑠𝑖𝑛2 𝛼 (15)

𝑉𝑐𝑖 = 0.18𝑣𝑐𝑖 𝑏𝑑 (16)

𝛥𝑐 + 0.25𝜀𝑡,𝑎𝑣𝑔 𝑑𝑐𝑜𝑡 2 𝛼1 (18)

𝜎𝑣 = min( 𝜀 𝑣 𝐸𝑣 , 𝑓𝑦𝑣 ) (19)

𝑉𝑠 = 𝜎𝑣 𝜌𝑣 𝑏(𝑑𝑐𝑜𝑡𝛼1 − 𝑙0 − 1.5𝑙𝑏1𝑒 ) (20)

Shear Forces (kN) 800

Shear Forces (kN) 800

Figure 2.15 Equilibrium of forces in five-spring model for imposed displacement Δ

3 EXTENDED FIVE-SPRING MODEL FOR FRC DEEP

3.2 TENSION IN FIBERS ACROSS THE CRITICAL DIAGONAL CRACK

Figure 3.1 Tensile behavior of FRC (Lee et al., 2013)

3.2.2 Formulation of the SDEM

𝑓𝑆𝐹𝑅𝐶 = 𝑓𝑓 + 𝑓𝑐,𝑠𝑜𝑓𝑡 (22)

𝑓𝑐,𝑠𝑜𝑓𝑡 = 𝑓𝑐𝑟 𝑒 −𝑐𝑤𝑐𝑟 (23)

𝑓𝑓 = 𝑓𝑠𝑡 + 𝑓𝑒ℎ (24)

The parameters Keh and Kst are defined as follows:

2 𝑤𝑐𝑟 1 𝑤𝑐𝑟 2 (29)

3.2.3 Validation of the SDEM

3.2.3.1 Experimental results by Lim et al. (1987)

3.2.3.2 Experimental results by Susetyo (2009)

3.2.4 Implementation of the SDEM into the Five-Spring Model