See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/318461635

Shear Behaviour of Reinforced Normal Weight Concrete Deep Beams Using

Finite Element Analysis

Article · July 2017

CITATIONS READS

0 90

4 authors, including:

Ezekiel Mensah Ababio Mark Adom-Asamoah

Lawrence Technological University Kwame Nkrumah University Of Science and Technology
4 PUBLICATIONS   0 CITATIONS    56 PUBLICATIONS   141 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Earthquakes View project

Structural Use of Lesser used Timber Species View project

All content following this page was uploaded by Mark Adom-Asamoah on 22 July 2017.

The user has requested enhancement of the downloaded file.

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

Shear Behaviour of Reinforced Normal Weight

Concrete Deep Beams Using Finite Element Analysis
Ezekiel Mensah Ababio, David Amegadoe, Jack Banahene Osei ,Mark Adom-Asamoah

Department of Civil Engineering,

Kwame Nkrumah University of Science and Technology
E-mail Id [email protected],

Abstract
The shear behaviour of structural elements, particularly deep beams has been an area of
huge concern in the structural engineering community. This study investigates the shear
behaviour of deep reinforced concrete beams by presenting and evaluating results of
experimentally tested beams to their corresponding finite element models. By using ANSYS,
the finite element computational platform, a validated model is presented for three deep
reinforced concrete beams that were cast and tested in the laboratory. Their load-deflection
responses were compared. The failure modes, as well as the cracking behaviour of the
models were also analyzed. It was observed that the load-deflection behaviour from the finite
element analysis correlate very well with that of the experimentally tested beams.
Furthermore, the finite element analysis was found out to mimic the cracking behaviour. It
also predicted the failure mode and ultimate failure loads of tested beam appropriately.

Keywords: deep beam, shear failure, finite element analysis; reinforced concrete

INTRODUCTION and mechanical properties of the fine and

Reinforced concrete is one of the most coarse aggregates (sand and stones) and
widely used building materials in the binding agent (cement). The major
construction. Concrete constitute coarse property of concrete which is checked at
aggregate in a continuous matrix of first hand is its mechanical behaviour
cement paste and fine aggregate. Unlike which is the ability of the hardened
structural steel, concrete is made from concrete to withstand loading measured as
natural materials whose properties may not the characteristic strength. This
be able to be regulated. Structural steel is mechanical behaviour depends on the
an artificially manufactured material development of micro cracks. Differential
whose properties can be regulated during shrinkage and thermal mismatch between
production to meet a certain quality. aggregates in the matrix leading to tensile
However, little can be done about the stresses and initial de-bonding cracks have
properties of concrete which is made from been observed by different researchers.
natural occurring sand and stones whose The behaviour of concrete is not fully
properties can be less regulated. The known and many studies are being carried
mechanical properties therefore vary from out daily on this widely used material. A
mix to mix and even within the same mix. thorough understanding of the properties
Different samples of the same concrete of reinforced concrete will provide
mix will possess different mechanical designers with full knowledge to explore
properties. This is as a result of the natural the capacities of this material.
properties of its constituents. The
behaviour of the final product therefore Many theories have been proposed by
depends majorly on the chemical, physical different authors for the analysis and

1 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

design of structural elements. These shear effect is predominant in beams

theories which include the beam theory are having L/D ratio less than or equal to 2.0
proposed based on various assumptions which may lead to warping of the section
made by the authors. The proposed (Niranjan and Patil 2012). The shear
theories are however subjective to whether behaviour of deep beams is complex. Due
the actual behaviour of the structural to the mode of failure of deep beams, they
elements will be as proposed. Beam are generally classified as non-flexural
theories as that proposed by Euler- members in which plane sections do not
Bernoulli and Timoshenko are made from remain plane in bending as assumed for
assumptions which best work on shallow normal beams by proposed beam theories.
beams but deviate from the behaviour of The shear in reinforced concrete is carried
deep beams. Beams with greater depth by concrete in compression shear zone,
have been known to deviate from these dowelling action of longitudinal
proposed theories. Most of the reinforcement and aggregate interlock.
assumptions used in the design of normal
reinforced concrete beams do not hold for The use of Finite Element Methods in the
the design and analysis of deep beams. analysis of deep beams has become a very
Even specialized literature such as the useful tool in modern times due to the
British Standard (BS 8110, 1997.) emergence of powerful computers. The
recommends that special literature be use of Finite Element Analysis is
consulted on the design of deep beams. extremely fast and economical as
This has given the need for engineers to compared to laboratory testing (Wolanski
investigate the behaviour of deep beams 2004).
since they are of great importance in
transferring large loads. The objective of this paper is to investigate
the shear capacity, deflection behaviour
Different methods have been used to study and mode of failure of three reinforced
the response behaviour of structural concrete deep beams that have been tested
elements. Experimental based testing is the in the laboratory and to compare the finite
most and widely used method for the element model results to the experimental
analysis and study of the behaviour and results[1-11].
mechanical properties of concrete under
loading. This method though having great BACKGROUND AND LITERATURE
cost implications produces the best real Deep beams are of great importance in
life response for understanding concrete. structures for transferring heavy loads
Finite element approach has also been usually at the lower parts of structures.
used by many researchers to study the They are used as transfer girders,
flexure, shear and even thermal behaviour foundation walls, pile caps and tanks.
of concrete. The method is also used to Deep beams are generally characterized by
verify or mimic laboratory results of short spans and thin widths as compared to
experimental based tests. depth.

Deep beams are generally beams with Deep beams as defined by the American
short spans and thin widths as compared to Concrete Institute (ACI) are members
depth. They are used as transfer girders, loaded on one face and supported on the
foundation walls, pile caps and tanks. opposite face so that compression struts
Deep beams fail majorly by shear as can develop between the loads and the
compared to normal beams that fail by supports[12-15]. They have either clear
flexure. Research conducted shows that spans equal to or less than four times the
2 Page 1-17 © MAT Journals 2017. All Rights Reserved
 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

overall member depth; or regions with bending and hence the beam can be
concentrated loads within twice the idealized as dimensional. The assumption
member depth from the face of the support is valid for regions within concrete
(ACI 318 2008). The European standard structures where the internal stresses can
also defines a deep beam as one in which be derived from sectional forces before
span is less than 3 times the overall section and after cracking of concrete. These
depth (EN 1992-1-1 2004). regions are known as B-regions (B stands
Other codes such as the Indian code (IS for Bernoulli) and the design made from
456 2000) defines a deep beam as a beam the use of sectional forces known as
with an effective span to overall depth sectional design.
ratio (L/D) less than 2 for a simply
supported beam and L/D less than 2.5 for The design and behaviour of deep beams
a continuous beam where the effective must be treated separately from sectional
span is defined as the centre-to-centre designs because the assumptions
distance between the supports or 1.15 mentioned earlier are no longer valid for
times the clear span whichever is lesser. them. Whilst bending strains dominate
The New Zealand Code (NZS-3101 2006) ordinary beams and can be predicted by
defines deep beams as members loaded on the beam theory, shear strains dominate
one face and supported on the opposite deep beams. The development of diagonal
face, so that compression struts can cracks due to shear have being observed to
develop between the loads and the be predominant in deep beams as
supports, and have either; clear span, Ln compared to flexural cracks in shallower
equal to or less than 3.6 times the effective beams (Sazzad, Younus, and Nizamud-
depth for simply supported or continuous Doulah 2002). In deep beams, a significant
beams and clear span equal or less than 1.6 amount of the load is carried to the
times the effective depth for cantilever supports by a compression force generated
beams[16-21]. The Canadian code (CSA- by the combination of the load and the
A23.3 2004) defines a deep beam as a support reaction. Nonlinear strain
beam in which the ratio of the clear span distributions are usually caused by abrupt
(l0) to the overall depth (h) is less than the change in geometry or loading and this
1.25 for simple spans and 2.5 for gives rise to disturbed or discontinuous
continuous spans. The CIRIA Guide regions often referred to as D-region. As a
(CIRIA Guide 2 1977) applies to beams result, the stress distribution is no longer
having an effective span to depth ratio (l/h) considered linear even in the elastic stage
of less than 2 for single-span beams and (Mohammad et al. 2011). Experimental
less than 2.5 for continuous beams. and finite element models have confirmed
These varied definitions of deep beams that Flexural strain and stress of deep
indicate the complexity of the behaviour of beams is not linear (Niranjan and Patil
deep beams. 2012). Shear deformations are no longer
negligible as they become excessively
Failure of Deep Beams large (Sazzad, Younus, and Nizamud-
Reinforced concrete beams fail through a Doulah 2002). Laboratory experiments
combination of shear and bending; majorly have proven that predicted deflections in
through bending based on the Bernoulli deep beams are far different from actual
hypothesis (beam theory) that the values in that the actual deformations are
distribution of strain is linear at a section. very large (Sazzad, Younus, and Nizamud-
The mechanical behaviour of beams is Doulah 2002). Plane sections do not
therefore defined by the assumption that remain plane and hence the assumption
plane sections remain plane even in that plane sections remain plane cannot be
3 Page 1-17 © MAT Journals 2017. All Rights Reserved
 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

used in the design and analysis of the 2011; Sazzad, Younus, and Nizamud-
behaviour of deep beam design. D-regions Doulah 2002). Deep beams behave more
have therefore been designed using as two-dimensional action rather than one-
empirical procedures and experience such dimensional action. Though the use of a
as the strut-and-tie model proposed by two-dimensional modelling can yield
ACI and Eurocode 1992[22-25]. This results of high accuracy and save computer
model has been used in many space and memory, the use of three-
investigations to predict the capacity of dimensional elements has proven to be
deep beams and to verify their accuracy more accurate even though computer space
(Adhikary, Li, and Fujikake 2013; Ashour and computational time increases
and Yang 2007) and found to give a better considerably (Metwally 2015).
understanding of the distribution internal
forces within deep beams although more Analysis of Deep Beams
than one admissible strut-and-tie model Methods such as empirical equations and
can be developed (Ashour and Yang the strut-and-tie do not take into account
2007). The model assumes that the deep the nonlinear material behaviour and have
beam is a truss in which compressive limitations in predicting the behaviour of
stress fields are resisted by concrete struts deep beams under special circumstances
an and tensile stress fields are resisted by such as deep beams in combined axial and
steel reinforcement ties and the two are bending force (Kim, Lee, and Shin 2011).
connected by nodes (Birrcher et al. 2009). Under such circumstances, the behaviour
This model like other mathematical of deep beams has been observed to totally
models do not take into account the differ from that suggested by the strut-and-
redistribution of forces due to nonlinear tie model. These differences have led to
materials‟ behaviour (Metwally 2015). the generation of modifications to the use
of the strut and tie model which is widely
Investigation on the stresses developed in used and presumed to be sufficiently
deep beams showed that at the ultimate accurate (Gandomi et al. 2013; Zhang and
limit state, distribution of stresses in Tan 2007). The use of finite element
concrete in compression and the shape of approach in investigating the behaviour of
concrete compressive stress block is no structural elements has grown and is
more parabolic and therefore proves that becoming useful in recent times. The
strain distributions are not linear as method has proven to be sufficiently
commonly accepted for shallow beams accurate and time saving. With a good
(Mohammad et al. 2011). The failure of model, the behaviour of structural
deep beams is highly characterized by elements can be mimicked with high level
diagonal cracks as compared to flexural of accuracy. A good model maximizes the
cracks for shallower reinforced concrete use of the appropriate finite elements,
beams. Diagonal cracks are usually the approach of modelling and computational
first cracks to be observed in the clear span time to produce the desired result.
of the deep beam (Mohammad et al. 2011). Numerous models have been proposed for
The principal mode of failure in deep the study of deep beams yet all such
beams having adequate reinforcement is models depend solely on the designer‟s
diagonal tension cracking and the presence ability to select the right elements and
of different web reinforcement parameters to achieve the desired results.
arrangements have being observed to have Researchers have used finite element
no appreciable effect on the formation of models to try to better predict the
initial diagonal cracks and the failure nonlinear behaviour of concrete deep
behaviour of deep beam (Mohammad et al. beams.
4 Page 1-17 © MAT Journals 2017. All Rights Reserved
 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

fluid flow, structural, magnetism,

Finite Element Analysis electricity etc. They possess various
Several finite element modelling packages deformational behaviours and different
have been used by various researchers to elements can be used for different
model the behaviour of structural purposes based on the inherent behaviour
components of which deep beams is of no of the material of interest.
exception. The behaviour of deep beams
investigated using finite element methods METHODOLOGY
has proven to be very accurate in Member Properties and Laboratory
predicting the behaviour of deep beams Testing
(Niranjan and Patil 2012). ANSYS Three deep beams labelled
(ANSYS, Inc. 2015) has proven to be one N3/15/T10/LH4, N3/15/T10/LH8 and
of the most accurate and effective research N3/15/T10/LV4H4 which have been tested
tools in engineering. Results from finite and investigated by (Adom-Asamoah and
element models using ANSYS has shown Afrifa 2012) in the laboratory under three-
good agreement with experimental point loading were modeled using
data(Kachlakev and Yim 2001). ANSYS ANSYS. The beam dimensions and
possesses quite a number of elements with percentages of steel reinforcements are
different properties. These elements range shown in Table 1 with Figures 1 to 3
through varieties of engineering showing the details of the beams.
applications involving soil mechanics,

Table 1: Beam Details

Beam ID Span Widt Heig Eff. a/d Eff. Conc. Conc. fy Flex Vert.
L(mm h ht Dept Rati Span Cylind Cube (MPa tensio Shear
) b(mm h(mm h o / er Strengt ) n Reinf.
) ) d(mm dept Strengt h reinf. R.v(%
) h h fcu(MP R.w )
fc(MPa a) (%)
)
N3/15/T10/L 940.0 150.0 350.0 312.0 1.5 2.7 22.5 30.0 413.8 1.15 0.4
H4
N3/15/T10/L 940.0 150.0 350.0 312.0 1.5 2.7 22.5 30.0 413.8 1.15 0.8
H4
N3/15/T10/L 940.0 150.0 350.0 312.0 1.5 2.7 22.5 30.0 413.8 1.15 0.4
H4

Fig 1; Reinforcement details of Beam N3/15/T10/LH4

5 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

Fig 2; Reinforcement details of Beam N3/15/T10/LH8

Fig 3; Reinforcement details of Beam N3/15/T10/LV4H4

Confined and Unconfined Concrete D modelling of confined and unconfined

Concrete properties were modelled concrete. The solid is capable of cracking
separately for confined and unconfined in tension and crushing in compression.
concrete. The material model for concrete The element is defined by eight nodes
used in the model is that suggested by having three degrees of freedom at each
Mander et al. (1988b). The model node: in nodal x, y, and z directions. The
considers different stress-strain behaviour most important aspect of this element is
for concrete in zones confined by stirrup the treatment of nonlinear material
and that outside the stirrup zone as in properties. The concrete is capable of
Figure 4. The concrete as shown in Figure cracking (in three orthogonal directions),
4 in the confined stirrup zone possesses crushing, plastic deformation, and creep.
better strength capacities than that in the The rebar is capable of tension and
unconfined zone. compression, but not shear. They are also
capable of plastic deformation and creep.

Fig 4; Confined and unconfined

integration point Fig 5; Geometry of Solid 65

Element Types BEAM 188

Two separate elements were used in Beam 188 as shown in figure 6 is used for
modelling the reinforced concrete beams. reinforcements in the model. The element
is based on Timoshenko‟s beam theory
SOLID 65 which includes shear-deformation effects.
Solid 65 as shown in figure 5 is used for 3- The element provides options for

6 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

unrestrained warping and restrained

warping of cross sections. The element is a Real Constants
linear, quadratic, or cubic two-node beam Solid 65 requires real constants for rebar
element in 3-D. Beam 188 has six or seven assuming a smeared model. Values can be
degrees of freedom at each node. These entered for Material Number, Volume
include translations in the x, y, and z Ratio, and Orientation Angles. The
directions and rotations about the x, y, and material number refers to the number code
z directions. A seventh degree of freedom for the type of material for the
(warping magnitude) is optional. reinforcement. The volume ratio refers to
the ratio of steel to concrete in the element.
The element is suited for linear, large In the modelling, Beam 188 was modeled
rotation, and/or large strain nonlinear using discrete reinforcements that
applications. Beam 188 includes stress connected nodes. Therefore, a value of
stiffness terms that enable the elements to zero was used for all real constants. No
analyze flexural, lateral, and torsional real constants were defined for Solid 65
stability problems (using eigenvalue and Solid 185 as they do not require any.
buckling, or collapse studies with arc
length methods or nonlinear stabilization). Material Properties
The material properties used for the
modelling and analysis is shown in Table
2. Material model 1 and 2 refer to the
confined and unconfined concrete models
respectively. Concrete properties were
modelled separately for confined and
unconfined concrete. The ultimate tensile
and compressive strengths for each beam
model were calculated using equations 1
Fig 6; Geometry of Beam 188 and 2 respectively.

Table 2; Material Properties

Material Model Number Element Type Material Properties
Linear Isotropic
EX 22.04 GPa
PRXY 0.2
Multi linear Isotropic
Point Strain Stress (N/mm2)
1 0.00040 8.800
2 0.00080 15.624
3 0.00120 19.640
4 0.00160 21.830
5 0.00205 22.500
1 Solid 65
6 0.00250 21.830
Concrete
Open Shear Transfer Coefficient 0.4
Closed Shear Transfer Coefficient 1
Uniaxial Cracking Stress 2.9551
Uniaxial Crushing Stress -1
Biaxial Crushing Stress 0
Hydrostatic Pressure 0
Hydro Biax Crush Stress 0
Hydro Uniax Crush Stress 0

7 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

Tensile Cracking Factor 0

Linear Isotropic
EX 18.5 GPa
PRXY 0.2
Point Strain Stress (N/mm2)
1 0.00040 7.200
2 0.00080 13.320
3 0.00120 17.760
4 0.00160 20.490
5 0.00240 22.500
6 0.00320 21.312
2 Solid 65
Concrete
Open Shear Transfer Coefficient 0.3
Closed Shear Transfer Coefficient 1
Uniaxial Cracking Stress 2.4
Uniaxial Crushing Stress -1
Biaxial Crushing Stress 0
Hydrostatic Pressure 0
Hydro Biax Crush Stress 0
Hydro Uniax Crush Stress 0
Tensile Cracking Factor 0
Linear Isotropic
EX 200 GPa
PRXY 0.3
3 Beam 188
Bilinear Isotropic
Yield Stress 413.8 N/mm2
Tangent Modulus 4000 N/mm2
Linear Isotropic
EX 200 GPa
PRXY 0.3
4 Beam 188
Bilinear Isotropic
Yield Stress 250 N/mm2
Tangent Modulus 4000 N/mm2

2004; Kachlakev and Yim 2001).

( ) (1)
(3)
( ) (2) ( )
Where Ec is Elastic modulus of the
(4)
concrete, fc‟ is concrete cylinder strength
and fr is the ultimate tensile strength (5)
(modulus of rupture), all in MPa. Where f is the stress at any strain ε in
MPa, Ec is the concrete elastic modulus in
Material model 1 was used for the MPa, ε is the strain at any stress f and εo is
confined concrete. The Elastic Modulus the strain at the ultimate compressive fc.
and tensile strengths for each beam model
are calculated from the adopted empirical The stress-strain relationship for concrete
equations respectively in (ACI 318 2008). in tension is assumed to be linearly elastic
The uniaxial compressive stress-strain up to the ultimate tensile strength (MOR)
curve for concrete in this study is obtained after which the concrete cracks and the
from equations 3 and 4 alongside equation strength reduces to zero. Figure 7 shows
5 to compute the multilinear isotropic the simplified uniaxial stress-strain
stress-strain curve. This is obtained from relationship used for concrete in most
(Wight and MacGregor 2012; Wolanski studies. However, the multilinear stress-

8 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

strain relationship for the concrete was the model as perfectly straight. Ideally, the
plotted beyond the ultimate compressive bond strength between the concrete and
strength to indicate the reduction of steel reinforcement should be considered.
compressive strength of the concrete after However, in this study, perfect bond is
failure. This relation is illustrated in figure assumed between steel and concrete
9. The Poisson ratio used for all concrete is (Kachlakev 2002). A Poisson ratio of 0.3
0.2. is used for all reinforcing steel. The stress-
(6) strain relationship for the steel used in the
model is shown in figure 8.
The self-weight of all the beams were not
included in the finite element models.

Fig 7; Simplified uniaxial stress-strain

curve for concrete

Material Model number 2 was used for the

unconfined concrete. This element was Fig 8; Stress-strain curve for steel
modelled as a linear isotropic element. The reinforcement
constants also guiding the properties of the
failure surface were also defined as
suggested by (Willam and Warnke 1974) .

Material Model number 3 was used for the

longitudinal reinforcement. Elastic
modulus and yield stress for the steel
reinforcement used in this the study
follows the experimental material
properties. The steel is assumed to be a
perfectly elastio-plastic material with
identical properties in both tension and
compression (Kachlakev and Yim 2001).
The material was modelled as a bilinear
isotropic element with elastic modulus,
yield strength, Poisson ratio and tangential
modulus shown in table 2. Fig 9; Multilinear Stress-Strain curve

Material Model 4 was used for the shear Modeling

links and some longitudinal reinforcement. The model consists of confined and
The material was modelled as a linear unconfined concrete, steel reinforcements
isotropic element with elastic modulus, and supports. The concrete and steel
yield strength, Poisson‟s ratio and tangent elements were modelled by plotting nodes
modulus shown in table 2. and joining neighboring nodes to form
elements. The confined concrete properties
The steel reinforcements are simplified in were used for concretes in the region

9 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

outside the shear links and that for 12. The solution controls as set for the
unconfined concrete were used for analysis is shown in Figure 13. The default
concretes inside the region of the shear nonlinear convergence criteria were based
links. Each concrete element was modelled on the force and moments with tolerance
to have dimensions of 25mm x 25mm x values set to 0.001 each.
20mm in the x,y,z directions respectively.
These dimensions were chosen so that
nodes will coincide perfectly with
supports. Figure 10 shows a 3-D model of
a sample beam modelled as confined and
unconfined concrete.
Reinforcements were modeled as line
elements by joining nodes. Figure 11
shows a 3-D model of the reinforcement.

Fig 62 Load application and boundary

conditions at supports

Fig 10; Concrete elements (confined and

unconfined)

Loading and Boundary Conditions

The beams were simply supported with a
hinge at one end and a roller at the other.
All restrained nodes at the support plates
were assigned a displacement boundary Fig 137; Basic Solution controls
value of zero. This is done to ensure that
the finite element model is a true
representation of the experimental model
which was simply supported so as to
produce a unique solution which will be a
true representation of the experimentally
tested beams.

A total point load of 230 KN was evenly

distributed and applied as a line load
across the centre of the beam. The total
load was applied in 2 KN increments. The Fig 148; Nonlinear Solution controls
finite element model showing loads and
boundary conditions is shown in Figure
10 Page 1-17 © MAT Journals 2017. All Rights Reserved
 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

RESULTS AND DISCUSSIONS mechanisms at the cracked faces (Shah

Failure Loads and Swartz 1995), i.e. The grain bridging
Using BS8110 for the theoretical analysis, process, interlocking between the cracked
the failure loads were compared with the face, crack tips blunted by voids and the
experimental failure loads obtained from cracked branching process may also
laboratory testing and that obtained from slightly extend the failures of the
the finite element model. The theoretical experimental beams before complete
failure loads were derived for three failure collapse. The finite elements models do
modes; flexure, compression and shear. not have these mechanisms (Kachlakev
Table 3 shows the different failure loads 2002). Also, the material properties which
for the experimental, theoretical and finite are assumed in this model may be
element analysis. The Theoretical Failure imperfect.
Loads (shear failure) of the beams
happened to be lower than the However, finite element models predict
Experimental and Finite Element Failure correctly the failure mode (shear failure)
Loads. The difference between the for the deep beams as it occurred in the
experimental and the finite element failure experimental beams.
loads may be as a result of the toughening

Table 3; Theoretical, Experimental and Finite Element Analysis Failure Loads

Beam ID Experimental Failure Theoretical Failure Finite Element Analysis (ANSYS)
Loads (KN) Loads (KN) Failure Loads (KN)
She Flex Compres
ar ure sion
N3/15/T10/L 163. 312.8
H4 222 52 6 26673.15 230
N3/15/T10/L 163. 312.8
H8 228 52 6 26673.15 228
N3/15/T10/L 327. 312.8
V4H4 228 27 6 26673.15 214

Load-Deflection Behaviour therefore show that the Finite Element

The mid-span deflection of the models were generally stiffer than the
experimental setup measured with a dial Experimental Setup.
gauge was compared to the mid-span The stiffer nature of the Finite Element
deflection of the finite element model. models could be attributed to the absence
Figures 14 to 16 show the load-deflection of imperfections such as micro cracks and
plots of the beams N3/15/T10/LH4, weaker bond strength between the concrete
N3/15/T10/LH8 and N3/15/T10/LV4H4 and the steel reinforcements. The presence
for both experimental and finite element of these in the experimental beams slightly
models (ANSYS) respectively. The reduces their stiffness, hence leading to
maximum deflections for all three beams higher deflections.
are shown in Table 4. The load-deflection
curves from the Finite Element models Beam N3/15/T10/LV4H4 was also stiffer
generally correlate well with that of the in the linear range at the onset of loading.
Experimental Setup. For all three models, However, after the first crack, the Finite
the mid-span deflections recorded from the Element models demonstrated slightly
Finite Element Models were lesser than lesser stiffness which reduced below its
the mid-span deflections recorded from the experimentally tested beam and increased
laboratory tested beams. The results slightly afterwards. Generally, the Finite

11 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

Element Model N3/15/T10/LV4H4 the experimental beam very well.

mimicked the load-deflection pattern of

Table 4; Deflections at Failure Loads

Beam Maximum experimental Maximum Finite Element Analysis
Number Deflection (mm) (ANSYS) Deflection (mm)
x100%
N3/15/T10/L
H4 2.65 2.39 90%
N3/15/T10/L
H8 3.30 2.95 89%
N3/15/T10/L
V4H4 2.67 2.37 89%

BEAM N3/15/T10/LH4
250

200
LOAD / KN

150

100

50

0
0 0.5 1 1.5 2 2.5 3

MIDSPAN DEFLECTION /mm

ANSYS EXPERIMENTAL

Fig 95; Load-Deflection curve for Beam N3/15/T10/LH4

12 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

BEAM N3/15/T10/LH8
250

200
LOAD / KN

150

100

50

0
0 0.5 1 1.5 2 2.5 3 3.5
MIDSPAN DEFLECTION / mm
ANSYS EXPERIMENTAL

Fig 106; Load-Deflection curve for Beam N3/15/T10/LH8

250
BEAM N3/15/T10/LV4H4
200

150
LOAD / KN

100

50

0
0 0.5 1 1.5 2 2.5 3
-50
MIDSPAN DEFLECTION / mm
ANSYS EXPERIMENTAL

Fig 117; Load-Deflection curve for Beam N3/15/T10/LV4H4

Failure Mode observed that shear cracks crossed the

The beams in this study failed by shear compression zone of the beam and passed
either by diagonal tension or splitting of through the whole concrete section as
the bond between the concrete and flexural shown in Figure 17. The ultimate load for
steel reinforcement which is also known as diagonal tension failure is almost equal to
shear bond failure. The mode of failure of the load at which diagonal cracks occur
the beams is a factor of concrete strength, (Kong and Evans, 1998). The load
percentage of main reinforcing bars and however for shear-bond failure is slightly
the application of the load. It was beyond that for diagonal cracking.

13 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

Fig 128; Beam N3/15/T10/LH4

Cracking Behaviour For a concrete structure subjected to

Cracking is one result of shrinkage of uniaxial compression, cracks generally
concrete and the action of shear forces in propagate parallel to the direction of the
concrete. The presence of cracks reduces applied compressive load since the cracks
the overall strength of the concrete. Initial result from tensile strains developed due to
cracks start to form at the bottom of the the Poisson‟s effect (Shah and Swartz
beam at the end of its linear response. 1995). Similar behaviour is seen in the
These cracks normally occur at the mid- ANSYS model. Circles indicating the
span bottom face of the beam. They are formation of cracks appear perpendicular
referred to as Flexural Cracks. More to the principal tensile strains at
cracks are formed and the widths of integration points in the concrete element
existing cracks are increased with increase close to the loading points.
in loading. Diagonal Tension Cracks are
now seen at the face of the beam close to The cracking behaviour obtained by using
the support and gradually progress towards the crack/crushing plot option in ANSYS
load on the loading face. At this stage, the indicates that the initial cracking of the
response of the beam is non-linear and beams in the finite element models.
deflections and stresses in the beam now Figures 18 and 19 show the formation of
become difficult to predict (Wolanski cracks in the ANSYS model and
2004). Table 5 shows the loads at which experimental beams respectively. Table 5
flexural and diagonal tension cracks summarizes the loads at which the first
started occurring in the beams for this flexural and diagonal tension cracks
study. develop in both the experimental and finite
element models.

Table 5; 3 Loads at formation of Flexural and Diagonal Tension Cracks

Beam Number At First Flexural Crack Load (KN) At First Diagonal Tension Crack Load (KN)
Experimental ANSYS Experimental ANSYS
N3/15/T10/LH4 100 44 148 80
N3/15/T10/LH8 162 40 166 66
N3/15/T10/LV4H4 86 40 120 60

14 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

Fig 19; Formation of Diagonal Tension Cracks in Beam N3/15/T10/LH4

Fig 20; Crack Pattern in Beam N3/15/T10/LH4 from Experiment

CONCLUSIONS AND observed in the finite element model.

RECOMMENDATIONS Generally, the Finite Element models
Conclusion demonstrate slightly higher stiffness as
The Finite Element Method was used to compared to the experimentally tested
analyse the Shear Behaviour of laboratory beams. This may be as a result of material
tested Deep Reinforced Concrete Beams imperfections which could not be modeled
and the results were compared. The and the formation of micro cracks on the
following conclusions can be drawn based experimental beams
on the experimental data and the finite
element analysis made. Recommendations
Variations in the results between the Finite
The midspan deflections of the Finite Element Models and Experimental Models
Element Models correlate very well with are as a result of the inability to perfectly
those Experimental Models, since the model conditions in the experimental
former was not less than 89% of the latter. beams. These include formation of micro
cracks due to shrinkage of the concrete,
The failure mechanism of the material imperfections and loss of
experimentally tested reinforced concrete composite action between concrete and
beams is well modelled using finite steel reinforcement due to bond slip. These
element model and the failure loads were conditions should further be investigated
well predicted. so that they can be incorporated in future
Finite Element Models.
From the experiment, beams failed by
formation of diagonal cracks that either REFERENCES
formed from around the support to the 1. ACI 318. 2008. “Building Code
loading point or propagated from already Requirement for Structural Concrete
formed vertical flexural cracks to the and Commentary.” American Concrete
loading point. Such crack patterns were Institute.

15 Page 1-17 © MAT Journals 2017. All Rights Reserved

 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

2. Adhikary, Satadru Das, Bing Li, and 10. EN 1992-1-1. 2004. “Eurocode 2:
Kazunori Fujikake. 2013. “Strength Design of Concrete Structures - Part 1-
and Behavior in Shear of Reinforced 1 : General Rules and Rules for
Concrete Deep Beams under Dynamic Buildings.”
Loading Conditions.” Nuclear 11. Gandomi, A. H., A. H. Alavi, D.
Engineering and Design 259 (June): Mohammadzadeh Shadmehri, and M.
14–28. G. Sahab. 2013. “An Empirical Model
doi:10.1016/j.nucengdes.2013.02.016. for Shear Capacity of RC Deep Beams
3. Adom-Asamoah, Mark, and Russell Using Genetic-Simulated Annealing.”
Owusu Afrifa. 2012. “Artificial Archives of Civil and Mechanical
Neutral Network Model for Low Engineering 13 (3): 354–69.
Strength RC Beam Shear Capacity.” doi:10.1016/j.acme.2013.02.007.
Journal of Science and Technology 32 12. IS 456. 2000. “Indian Standard Code
(2): 119–32. of Practice Plain and Reinforced
4. ANSYS, Inc. 2015. “ANSYS (2015) Concrete IS 456:2000-. Bureau of
ANSYS User‟s Manual Revision Indian Standards, Manak Bhavan, New
15.0.” ANSYS, Inc. Delhi, India.”
5. Ashour, Ashraf, and Keun-Hyeok 13. Kachlakev, Damian. 2002. “Finite
Yang. 2007. “Application of Plasticity Element Ananlysis and Model
Theory to Reinforced Concrete Deep Validation of Shear Deficient
Beams.” Morley Symposium on Reinforced Concrete Beams
Concrete Plasticity and Its Strengthened with GFRP Laminates.”
Application, July, 1–16. Carlifonia Polytechnic State
6. Birrcher, David, Robin G. University.
Tuchscherer, Matt Huizinga, Oguzhan 14. Kachlakev, Damian, and Solomon
Bayrak, Sharon Wood, and James Yim. 2001. “Finite Element Modeling
Jirsa. 2009. “Strength and of Reinforced Concrete Structures
Serviceability Design of Reinforced Strengthened with FRP Laminates.”
Concrete Deep Beams.” 0-5252–1. FHWA-OR-RD-01-XX. Oregon
Texas Department of Transportation Department of Transportation and
and the Federal Highway Federal Highway Administration.
Administration. 15. Kim, H. S., M. S. Lee, and Y. S. Shin.
7. BS 8110. 1997. “British codeS of 2011. “Structural Behaviors of Deep
Practice B.S.8110 (BRITISH RC Beams under Combined Axial and
STANDARDS, Structural Use of Bending Force.” Procedia
Concrete Code of Practice for Design Engineering, The Proceedings of the
and Construction).” Twelfth East Asia-Pacific Conference
8. CIRIA Guide 2. 1977. “CIRIA Guide 2 on Structural Engineering and
(1977), „The Design of Deep Beams in ConstructionEASEC12, 14: 2212–18.
Reinforced Concrete‟, Over Arup and doi:10.1016/j.proeng.2011.07.278.
Partners and Construction Industry 16. Metwally, Ibrahim M. 2015. “Three-
Research and Information Association, Dimensional Nonlinear Finite Element
London, (Reprinted 1984), 131p.” Analysis of Concrete Deep Beam
9. CSA-A23.3. 2004. “CANADIAN Reinforced with GFRP Bars.” HBRC
CODE (CSA-A23.3- Journal, February.
2004)(CANADIAN STANDARD doi:10.1016/j.hbrcj.2015.02.006.
ASSOCIATION Design of Concrete 17. Mohammad, Mohammadhassani,
Structures).” Mohd Zamin Bin Jumaat, Mohamed
Chemrouk, Ali Ghasemi, S. J. S.
16 Page 1-17 © MAT Journals 2017. All Rights Reserved
 Journal of Ceramics and Concrete Sciences
Volume 2 Issue 2

Hakim, and Rafieipour Najmeh. 2011. and the Effect of Shear

“An Experimental Investigation of the Reinforcement.” In . Singapore.
Stress-Strain Distribution in High 21. Shah, Surendra P., and Stuart E.
Strength Concrete Deep Beams.” Swartz. 1995. Fracture Mechanics of
Procedia Engineering, The Concrete; Applications of Fracture
Proceedings of the Twelfth East Asia- Mechanics to Concrete, Rock and
Pacific Conference on Structural Other Quasi-Brittle Materials. United
Engineering and States of America: John Wiley and
ConstructionEASEC12, 14: 2141–50. Sons, Inc., New York.
doi:10.1016/j.proeng.2011.07.269. 22. Wight, James K., and James G.
18. Niranjan, B. R., and S. S. Patil. 2012. MacGregor. 2012. REINFORCED
“Analysis of R.C Deep Beam by Finite CONCRETE Mechanics and Design.
Element Method.” International 6th ed.
Journal of Modern Engineering 23. Willam, K. J., and E. P. Warnke. 1974.
Research (IJMER) 2 (6): pp-4664- “Constitutive Model for the Triaxial
4667. Behaviour of Concrete.”
19. NZS-3101. 2006. “NEWZEALAND 24. Wolanski, A. J. 2004. “Flexural
CODE (NZS-3101-2006)New Zealand Behaviour of Reinforced and
Standard CONCRETE STRUCTURES Prestressed Concrete Beams Using
STANDARD Part-1 The Design Of Finite Element Analysis” (2004).”
Concrete Structures.” 25. Zhang, Ning, and Kang-Hai Tan. 2007.
20. Sazzad, Md M, M M Younus, and S M “Direct Strut-and-Tie Model for Single
Nizamud-Doulah. 2002. “Study of Span and Continuous Deep Beams.”
Behaviour of Reinforced Concrete Engineering Structures 29 (11): 2987–
Deep Beam under Two Points Loading 3001.
doi:10.1016/j.engstruct.2007.02.004.

17 Page 1-17 © MAT Journals 2017. All Rights Reserved

View publication stats