Basappa - 2013 PDF

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 16

Journal of Structural Engineering

Vol. 40, No. 2, June - July 2013 pp. 169-184 No. 40-23

Modeling of CFRP strengthened RCC beam using the nonlinear finite element
method
Umesh Basappa* and Amirtham Rajagopal*,
 Email: rajagopal@iith.ac.in

*Department of Civil engineering, Indian institute of Technology Hyderabad, 502 205, India.

Received: 31 August 2012; Accepted: 29 October 2012

Computational modeling of fracture in reinforced cement concrete (RCC) beam considering various phenomena has
been a challenging task over the years. This paper presents a crack modeling methodology in three dimensions for
carbon fiber reinforced polymer (CFRP) strengthened RCC beam by performing a three dimensional nonlinear finite
element analysis of the beam subjected to four point loading. The concrete is modeled as inelastic material. Various
concrete failure parameters such as shear transfer coefficients, uniaxial tensile and compressive strengths, biaxial
compressive and crushing strengths and stiffness reduction in cracked concrete in tensile region are considered. In
numerical studies an unstrengthened beam with and without hanger bars are considered for the analysis and results
are compared with experimental results. In the next case, a CFRP strengthened RCC beam is considered for analysis. A
parametric study is performed considering different length of CFRP used for flexural strengthening of beams, modeling
CFRP as Isotropic and orthotropic and varying the area of steel reinforcement in tension region. The study indicates
that the proposed method is able to accurately predict the behavior, crack patterns and load carrying capacity. The
results are comparable with the experimental results available in the literature.

Keywords: Nonlinear FE analysis; crack in RC beams; flexural strengthening; CFRP.

Reinforced cement concrete (RCC) structures has loss in the human life and assets. Thus there is a need
found a wide range of applications like for instance in for accurate and reliable method to assess the safety
high rise buildings, bridges, pre-cast structures, dams, and serviceability of RCC structure.
tunnels are some amongst others. These structures are Experiments are extensively conducted to predict
designed to satisfy serviceability criteria (prediction the response of RCC structures providing a real life
of cracks and deflection under service load) and safety response. However experiments are time consuming,
criteria (prediction of load deformation behavior of costly and include improper simulation of loading and
RCC structures and estimation of ultimate load). In the support conditions of the actual structure. With the
present scenario, the construction of modern structures advent of digital computers and analysis method such
and loading histories are more complex together as finite element method (FEM), it is possible to predict
with an increase in the cost of construction seeking the nonlinear behavior of RCC structures numerically.
innovative design without negotiation for safety of
Classically two methods are available for numerical
structure because any structural failure would include
modeling of cracking in concrete namely discrete

Journal of Structural Engineering 169


Vol. 40, No. 2, June - July 2013
crack model and smeared crack model1-2. Among these concrete by adding fibers to matrix called as steel
methods smeared model has become more popular and fiber reinforced cementations composites (SFRCC’s).
researchers started to develop constitutive models for Recently, computational modeling of SFRCC’s gained
concrete under the triaxial behavior3. The outcome of interest in research. Discrete or lattice crack model is
these studies include two major parameters with respect extensively used to study the behavior of matrix-fiber
to cracking phenomenon namely shear retention factor and matrix-aggregate interface and both meso/macro
and tension stiffening4. It was observed in some of level of approach are used for the analysis11-13.
these studies that tension stiffness (brittle crack model) Many studies have been carried out in developing
must be replaced with tension softening behavior efficient renovation or retrofitting techniques for
(linear or gradual weakening behavior), to match with post strengthening of concrete structures in bridges.
the experimental data. However the tension softening Because of deterioration of strength concrete structures
behavior was found to be prone to mesh sensitivity5 due to corrosion of steel reinforcement and continuous
and does not consider some concepts from fracture increase in the traffic volume. Even, structures like
mechanics6. beam, column and slab amongst others require post
Later models proposed included a tension softening strengthening. Traditional approach of strengthening
behavior that was mesh independent5. Discrete crack concrete structures was by external post tensioning
modeling assumes a predefined crack in the structure and steel plate bonding. However bonded steel plates
and uses concepts from linear elastic fracture mechanics are also found to exhibit corrosion due to continuous
(LEFM), together with adaptive mesh refinement and exposure to weathering and it was recommended that
or interface elements to predict the crack propagation. partial substitution of steel plates with polymer matrix/
The advantage of this method is that it gives a near fiber composites. A comparative study earlier made
correct prediction of crack propagation path. However, concluded that carbon fiber reinforced polymer (CFRP)
the method is applicable only when a single dominant are best suitable for the post strengthening of concrete
crack in the structure leads to complete structural failure. structures14-15.
Smeared crack models predict the cracking behavior Effective application of CFRP to post strengthen
by making changes in the constitutive relation. In this the concrete is not possible until the different failure
work we present a smeared crack model to study the mechanisms of CFRP reinforced concrete are studied.
cracks in reinforced cement concrete. In one of the earlier works various possible modes
Classical approaches are more related to global of failure of the retrofitted concrete beam has been
stress-strain relation in defining fracture mechanisms identified as flexural compression failure, shear failure
of concrete materials. These techniques do not include in concrete beam, delamination of CFRP and debonding
material structure at different length scale (nanometer of a layer of concrete at the flexural steel level16.
to millimeter). Bazant and Schlangen studied 2D In the present work a three dimensional nonlinear
fracture mechanisms in heterogeneous materials like finite element analysis is performed to predict the
concrete, using random particle or lattice model7-8. flexural cracking behavior of CFRP strengthened RCC
The method consists discretization of material beams. A smeared crack approach was assumed and
with the use of truss or beam element as lattice and analysis is performed using commercially available
assigning the different properties to map the material software ANSYS 2010. The behavior RCC beam with
microstructure. The fracture mechanism is captured and without hangers is compared with experimental
by deleting lattice elements based on the linear elastic results. Later, the crack patterns for the various area of
approach. It is observed that, the microstructure is the steel reinforcement were simulated to judge the failure
one which extensively leads to fracture mechanisms in pattern of RCC beam with hanger bars. A parametric
heterogeneous material like concrete9-10. study is made for various lengths of carbon reinforced
Concrete structures have very low tensile strength, fiber polymer (CFRP) used for post strengthening of
fatigue resistance, fracture toughness etc. In earlier the flexural capacity of RCC beam with and without
days, researchers improved the mechanical behavior hanger bars.

170 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013
Material Behavior
Modeling of Concrete

Concrete is a heterogeneous material (composed of


fine and coarse aggregates) characterized by highly
nonlinear and ductile stress-strain relationship. The
nonlinear behavior is attributed to the formation and
gradual growth of microcracks under loading. The
microcracks can be categorized as bond cracks and
mortar cracks. Bond cracks occur along the interface
between the mortar and coarse aggregate and exist due
to the improper bonding or due to the differences in
the stiffness between the coarse aggregates and mortar.
These cracks also occur due to shrinkage of concrete
and at low load levels. Mortar cracks are present in the
mortar between the pieces of aggregate and develop at
high load levels or stresses.
The uniaxial stress-strain behavior of concrete in
compression is shown in Fig. 1a. Various mathematical
models are available to approximate this nonlinear
behavior namely linearly elastic-perfectly plastic
model, inelastic-perfectly plastic, Hognestad and
piecewise linear model. In the present study a modified Fig. 1 (a) Uniaxial stress strain behavior of concrete (b) Modified
Hognestad model used in the present study
Hognestad mathematical model (see Fig. 1b) has been
used for the approximation of the stress-strain behavior
The stress-strain behavior of concrete under tension
of concrete21-22.
includes raising part and descending part. The raising
1. Initial tangent modulus of elasticity increases with part is slightly curved, approximated either as straight
an increase in compressive strength. So, the elastic lines or parabola. The descending part drops rapidly
modulus (EC) is given by with increased elongation after the maximum stress is
E C = 57000 fC (1) crossed. The tensile strength of concrete varies 8-15%
of the compressive strength.
where, fc compressive strength of concrete at 28 days
(Psi) The various tensile crack models are available in the
literature for crack modeling in concrete: Brittle, linear
The stress-strain relation initially must satisfy the and nonlinear models6. In case of brittle crack model
Hooke’s law. crack occurs when maximum principal stress exceeds
2. The strain at maximum stress increases as the the tensile strength of concrete. In linear and nonlinear
compressive strength increases. crack models, crack occurs when the principle stress
2 fC exceeds the minimum tensile strength and residual
  ε 0 = (2) tensile strength as given in Eq. (4) through Eq. (6).
EC
3. The raising portion of the stress strain curve resem-
bles a parabola with vertex at the maximum stress.
EC ε
  f = (3)
ε 2
1+
ε0
Where, f is stress for an value of strain in stress-strain
relationship of concrete.

Journal of Structural Engineering 171


Vol. 40, No. 2, June - July 2013
σ3

σ2
where, ft is tensile strength of the concrete, eult is strain
at failure of concrete, f is the stress at an value of strain
e in the stress strain relation of concrete.
f(σ) σ1
Stress, f

Stress, f
Stress, f

ft ft ft

Et Et
Et σ1 = σ2 = σ3
Strain,ε Strain,ε Strain,ε
Fig. 3 Failure surface of plain concrete under triaxial compression
ε1 ε ult ε1 εult ε1 εult (Willam and Warnke 197422)
(a) (b) (c)
Fig. 2 (a) Brittle crack model (b) Linear tensile softening crack The condition for the failure of plane concrete
model (c) Nonlinear tensile softening crack model member under the triaxial stresses is given by23-24,
F
The idealization of finite elements and size of -S≥0 (7)
mesh chosen clearly indicates the continuum level fc
of approach to modeling concrete material. Willam where, F is a function of principle stresses (sxp, syp,
and Warnke derived a mathematical model to plot szp), S is the failure surface expressed in terms of
the failure surface under triaxial behavior of concrete principal stresses and five input parameters, fC is the
material (see Fig. 3). Concrete is assumed as isotropic uniaxial crushing strength and sxp, syp, szp are the
material, as a result the failure surface are expressed in principal stresses in the principal directions.
principle space23.
Failure surface (S) is plotted using five input
The four aspects considered in deriving the parameters namely: Ultimate uniaxial tensile strength
mathematical model of failure surface: (ft), Ultimate uniaxial compressive strength (fC),
1. The close fit of experimental data in the operating Ultimate biaxial compressive strength (fcb), Ambient
range is considered. The principle stresses are hydrostatic stress state (sh), Ultimate compressive
ordered as s1  s2  s3, and then the failure surface strength for state of biaxial compression superimposed
can be expressed as a function of hydrostatic on hydrostatic stress state (f1), The ultimate compressive
and deviatoric stresses. The hydrostatic section strength of a state of uniaxial compression superimposed
contains equisectrix s1 = s2 = s3 as an axis of on hydrostatic stress state (f2). The failure surface can
revolution. The deviatoric section lies in a plane also be specified with two parameters ft and fC keeping
normal to the equisectrix. other parameters to default values.
2. Simple identification of model parameters from fcb = 1.2 fC (8)
the standard test data. The parameters are included
in the mathematical model such that they are easily f1 = 1.45 fC (9)
identified through the standard test (Uniaxial
compression, tension and biaxial compression). f2 = 1.725 fC (10 )
3. Smoothness in failure surface i.e., continuous
surface with continuously varying tangent planes These values are valid only for the condition,
4. Convexity in failure surface i.e., monotonically sh √3 fc.
curved surface without inflection points

172 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013
2.2 Modeling of Reinforcement steel
C
The mechanical behavior of reinforcing steel bar is B
obtained by testing the bar under monotonic tension

Load
loading. The steel bar initially exhibits linear elastic
portion followed by a yield plateau, strain hardening
and then stress drops till fracture occurs. The behavior
of steel bar remains same in compression and tension 0A– uncracked elastic stage
loading.
AB– Crack propagation stage
The stress-strain behavior of steel bar is independent
of environmental conditions and time and also the A BC – Plastic stage
reinforcements are used as bars in concrete. Hence, three
dimensional mechanical behavior of the reinforcement
steel bar is ignored unlike the triaxial behavior of
Deformation
concrete considered in the RCC beam25.
The model adopted to represent the reinforcement is Fig. 5 Load-Deformation behavior of RCC beam
elastic-perfectly plastic as shown in Fig. 4. The stress
for any value of strain is given as follows6 Initially, the strains are very small at the bottom
fs = ± E s ε s -ε y ≤ ε ≤ + ε y (11)
fiber resulting in uncracked elastic stage. As the stress
Liner elastic-perfectly reaches the tensile strength of concrete, primary
plastic model cracks are initiated. At discrete cracked locations the
fs = ± fy ±ε y ≤ ε ≤ ± ε ult (12)
tensile stresses are entirely carried by the steels and
also less concrete sections are effective in resisting
where, fs is the stress for any value of strain es in stress the load, resulting reduction in the stiffness of the
strain relation of steel reinforcement. Es and fy is the beam. The decrease in the stiffness of beams increases
modulus and yield value of the steel reinforcement. with increase in loading. Thus, the slope of the load-
deformation behavior also decreases as shown by the
Stress, f Compression
crack propagation stage.
-fy ,-ε y
When the stresses in reinforcement reach the yield
value deformation increases quickly with little increase
ε ult in loading. Finally, beam fails due to crushing of
ε ult Strain, ε
concrete as shown by the plastic stage21.

Material Property of CFRP


Tension
+fy ,+ε y
Carbon fiber reinforced polymer (CFRP) consists of
carbon fiber embedded in a matrix of polymer resin.
Fig. 4 Stress strain curve for the steel reinforcement Carbon fiber provides strength and stiffness to composite
and matrix protects the fiber from environmental impact
Reinforced Concrete and ensures proper sharing of the load among fiber.
Typical stress-strain behavior of carbon fiber, matrix
Reinforced concrete structures are made up of two and CFRP is shown in Fig. 6b. The strain of the carbon
materials, concrete and steel. The load deformation fiber is less than the matrix26-27.
behavior of reinforced cement concrete beam is as
The material property of CFRP is either considered
shown in Fig. 5.
as linear isotropic or linear orthotropic property. The
The non-linear behavior of RCC is divided into thickness of CFRP is taken as 1.2mm. The material
three stages: uncracked elastic stage, crack propagation properties for the CFRP are shown in Table 1.
and plastic stage.

Journal of Structural Engineering 173


Vol. 40, No. 2, June - July 2013
Table 1
Material properties of CFRP
Isotropic property Linear orthotropic property
Parameter Values Parameter (MPa) Values Parameter Values Parameter (MPa) Values
Exx (MPa) 165000 Exy 165000 Pxy 0.30 Gxy 5200
Pxx 0.3 Exz 9650 Pxz 0.45 Gxz 3400
Eyz 9650 Pyz 0.30 Gyz 5200

Solid
element
Solid element
Link/Beam
nodes
element

Matrix Fibers
(a)
Shared nodes
Fig. 7 Discrete method of FEM of RCC beam
Stress

Fiber
4
P O
5 6
Composite
M N

Y, v 3

Matrix
L K
X, u
Z, w 2 1
Strain
(b) I J

Fig. 6 (a) Schematic diagram of CFRP (b) stress strain plot for Fig. 8 3DSolid Concrete element
CFRP
Linear isotropic and nonlinear inelastic multi-linear
Finite Element modeling of RCC beam isotropic based on Von Mises failure criterion material
properties were used for the solid element. The cracking
Finite element modeling comprises of using an idealized and crushing behavior of concrete is based on William
element and meshing of elements to replicate the RCC and Warnke model. The inputs required to implement
beam. In this study discrete method is used to model the these behaviors are listed in Table 2, Table 3 and Table 4.
RCC beam (See Fig. 7). A perfect bond is considered The value of the shear transfer coefficient as suggested
between the concrete and steel reinforcement28. by Wolanski, 0.3 and 1 for open and closed crack were
A 3D finite element was used to model the concrete. considered to avoid the convergence problem. The
The element contains 8 nodes with 3 degrees of freedom element is also used for modeling of loading and support
at each node: translation in X, Y and Z directions plates. Linear isotropic material model was considered
(see Fig. 8). The element is capable of cracking in for the loading and supporting plates. The input
three orthogonal directions in tension, crushing in parameters required are modulus of elasticity equal to
compression and plastic deformation. modulus elasticity of steels and Poisson’s ratio (0.3).

174 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013
Table 2 In a pure compression failure of concrete secondary
Concrete Elastic Properties tensile strains are induced by Poisson’s effect
perpendicular to the load. Because of concrete weak in
Linear Isotropic
tension these cracks eventually lead to failure. Therefore,
E 30000 N/mm2
in this study, the crushing capability was turned off30.
 0.2
3D spar element was used to model the reinforcement.
The element has 3 degrees of freedom at each node:
Table 3
translation in X, Y and Z directions. The element is
Concrete Inelastic Properties capable experiencing uniaxial tension, compression
Nonlinear inelastic- Multi-linear isotropic and plastic deformation. The geometry and dof at each
Strain Stress node are shown in Fig. 9. Two real constant was used
Pt. 1 0.0003 09.000 for the longitudinal reinforcement and stirrups.
Pt. 2 0.0006 16.514 Another two real constant are required because
Pt. 3 0.0009 22.453 of the symmetry considered along longitudinal and
Pt. 4 0.0012 26.476
transverse direction, which reduces the reinforcement
area by half. The material model required to replicate
Pt. 5 0.0015 28.800
the reinforcement behavior is nonlinear inelastic
Pt. 6 0.0020 30.000
bilinear isotropic based Von Mises failure criterion.
Pt. 7 0.0022 30.000 This material model requires specification of modulus
of elasticity (2105 MPa), poisons ratio (0.3) and yield
Table 4 stress (420 MPa)
Concrete failure Properties
Constants - Value Y, v
Shear transfer coefficient for open crack bt 0.35
X, u
Shear transfer coefficient for closed crack bc 0.95 Z, w
Uniaxial tensile strength of concrete st 3 Fig. 9 3D Spar Reinforcement Element
Uniaxial compressive strength of concrete st -1
Biaxial compressive strength of concrete sbc 0 A 3D shell element having in plane membrane
stiffness but no out plane bending stiffness has been
Ambient hydrostatic stress state sh 0
used for modeling the CFRP. The element has 4 nodes
Biaxial crushing stress under the ambient s1 0
with 3 dof at each node: translation in the nodal X, Y
hydrostatic stress state
and Z directions (see Fig. 10). The element is capable
Uniaxial crushing stress under the ambient s2 0
large deflection, stress stiffening, variable thickness
hydrostatic stress state
and cloth option.
Stiffness multiplier for cracked tensile - 0.6
condition Z ij 5 K
L
Z
Whenever RCC beam cracks at discrete locations Y ij 1
2
as load exceeds the tensile strength, stiffness reduces at Y 6 4
this discrete location compared to surrounding concrete X
elements. In FEA, this stiffness is suddenly reduced to I 3 J
zero at cracking creating convergence problem. A stress X ij
Y, v
relaxation technique was used to gradually reduce the
stiffness surrounding the cracked element and solve
the convergence problem. The value of the stiffness X, u
multiplier for cracked tensile condition was taken as Z, w
0.629. Fig. 10 3D Shell CFRP Element

Journal of Structural Engineering 175


Vol. 40, No. 2, June - July 2013
Nonlinear Finite Element Stiffness matrix modified to include the material
formulation nonlinearity.

The strong form of the differential equation for continua  dN k dN i dN i


    K ij  Cijkl 1  k U k  (24)
is given by  dx  dx dx

. [ a(U) U] + f = 0 = Residual (13) Writing in matrix form,

The residuals are minimized using one of the   K (U) U = f (25)


weighted residual approaches as follows: The Newton Raphson method used to trace the
nonlinear material behavior.
 {. [ a (U ) U ]}W + 
 
fW = 0 (14)
r (U) = K (U) U – f (26)
where W is the weighting function and U is the trail r (U) = 0
function. After simplifying the integral form results in
a weak form.    Ur+1 = Ur – T-1(Ur) r (Ur) (27)
where,
 
a(U) U W =  
fW +  RW
r
(15)
dr (Ur ) d[ K (U )U  f ]
T (Ur )   (28)
Expressing in terms of Bilinear and Linear du du
functions,
d[ K (Ur )]
B (U, W) = L (W) (16) T (Ur )  K (Ur )  Ur (29)
du
The material is discretized as,
Writing in matrix form,
U ( x , y , z )  U h ( x, y, z )   j U j N j ( x, y, z ) K im
n
  Tij  K ij   m1 Um (30)
W ( x, y, z )  i Wi N i ( x, y, z ) (17) u
Here, the iteration process is terminated when
Including the discretization process in weak the residual r(U) is very small or the successive
formulation, displacement is small compared to user specified
 dN j dN i tolerance. An L2 norm of residual or successive
 a(U )  j U j    f N i r R N i (18) displacement was considered in the comparison with
 dx  dx
the tolerance values.
Writing in matrix form: The L2 norm expression for the residual or successive
  Kij Uj = fi (19) displacement are given as,
Where,   L2 norm for the residual ||r||2  Tolerance (31)
dN i dN j L2 norm for the successive displacements,
K ij   a(U ) (20)
dx dx Ur1 Ur
  2
 Tolerance (32)
fi   f N i  r R N i (21) Ur1
Material Nonlinearity: Expressing modulus as a In Newton-Raphson (N-R) method, end load
function of displacement. step are divided into a number of small increments
called substep to predict the nonlinear changes in the
 dU 
a(U )  Cijkl1  (22) structural behavior. N-R method provides convergence
 dx  requirement at the end of each load step with user
Material non-linearity including discretization. specified tolerance. To achieve the convergence of
a solution tolerance was increased by 5 times to the
 dN k 
  a(U )  Cijkl1  k U k  (23) default values due to the nonlinear behavior of RCC
 dx  beam.

176 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013
Numerical studies analysis. In the second numerical example a reinforced
concrete beam without hanger bars and strengthened
In the present work, three dimensional nonlinear with a layer of CFRP is considered for analysis. In the
finite element analysis of RCC beam has been made third example a parametric study is performed by varying
to capture the crack formation and growth in flexure32. the amount of reinforcement steel and considering
A four point loaded beam is considered for nonlinear the beam to be strengthened with CFP. The loads are
analysis. The geometry and loading details of the beam gradually applied in small increments till the failure of
considered in the present work are taken from earlier the beam and crack pattern observed at the end of each
experimental works done by the Blockhouse and Foley load increment are recorded. The numerical examples
and are available in the literature32. are discussed in detail in the following sections.
Nonlinear analysis is performed in two cases namely
with and without hanger bars to simulate the crack Numerical Example 1
pattern. Considering symmetry in both longitudinal A flexural beam with four point load was considered for
(loading) and transverse direction (geometry), only one the FE analysis. The geometry and reinforcement details
fourth of the beams were considered in making the FE of beam are shown in Fig. 11. Here, the beam RB1
model. Both 3D solid concrete and 3D spar element and RB2 represents the beam without and with hanger
formulations are used in the concrete and reinforcement bars considered for the study. The shear reinforcement
for the discretization of the beam. Three numerical provided is sufficient enough for the shear strength of
examples have been considered. In the first example the beam. The longitudinal reinforcement is sufficient
a simple reinforced concrete beam with and without enough to show sufficient cracks or warning before the
hanger bars subjected to four point load is considered for failure of the beam.

P P varies

3 - 16 φ
10 φ 10 φ 10 φ
@ 200mm c/c spacing @ 250mm c/c spacing @ 200mm c/c spacing

4800 mm

(a)

250 mm

2 - 12 φ

500 mm

Experiment (1998) RB2


(b)
Fig. 11  Geometrical details of RCC beam (a) Longitudinal section details (b) Cross section details Foley and Buckhouse Experiment 1998)

Journal of Structural Engineering 177


Vol. 40, No. 2, June - July 2013
Cracking patterns in RB1 Flexural diagonal cracks are identified at 35kN
The entire behavior of RCC beam is divided into three load. In general, these cracks are identified between
stages: linear elastic stage, crack propagation and the loading and support plate as shown by an incline
plastic stage. The linear elastic stage depends on the red circle outlines in Fig. 15. At this stage entire load
tensile strength of concrete. Once applied load exceeds in tension region is carried by the reinforcement. The
tensile strength cracks are initiated at the bottom fiber yielding of steel is identified at 57.0 kN load.
of the RCC beam. In this study, the initial cracks are The second and third cracks, green and blue circle
identified at load 18.735 kN as shown in Fig. 12. outline in tension region indicating the yielding of steel
reinforcement as shown in Fig. 16. The deflection of
| First cracks
| Second cracks the beam increases at a faster rate after the yielding of
| Third cracks
reinforcement showing large numbers of cracks in the
constant moment region. These cracks are shown in
Detail A Fig. 17 by a large numbers of green and blue cracks.
RCC beam reaches ultimate failure at 84kN load shown
in Fig. 18.
Detail A
| First cracks
Fig. 12 first crack at load 18.735kN for RB1 | Second cracks
| Third cracks

The plot of stress distribution at first cracking is


shown in Fig. 13. Experimentally it is difficult to find Cracking at 44.64 kN (a)
the load corresponding to first crack. The first crack is Diagonal cracks
identified at a distance of 2.075m from the left support | First cracks
(Fig. 12). As the load crosses 18.735kN flexural cracks | Second cracks
| Third cracks
start propagating in the vertical direction as shown by the
red circle outline in Fig. 14. At this stage the loads are
carried by the reinforcement at the cracked element. Cracking at 54.00 kN

MN Fig. 15 Flexural diagonal cracks for RB1 at load (a) 44.64kN


(b) 54.00 kN

| First cracks
| Second cracks
| Third cracks

0.03741 0.09863 6.19352 9.28842 12.3833


1.55119 4.64608 7.74097 10.8359 13.9308
Cracking at 57.60 kN (a)
Fig. 13 Stress plot at first cracking for RB1
| First cracks
Flexural cracks parallel to applied load | Second cracks
| Third cracks
| First cracks
| Second cracks
| Third cracks

Cracking at 64.48 kN (b)


Yielding of steel reinforcement
Cracking at 18.84 kN (a) | First cracks
| Second cracks
| First cracks
| Third cracks
| Second cracks
| Third cracks 25

Cracking at 68.52 kN (C)


(b)
Fig. 16 RB1 - Cracks during yielding of steel reinforcement at load
Fig. 14 Flexural diagonal cracks for RB1 at load (a) 18.84kN (a) 57.60 kN (b) 64.48 kN (c) 68.52 kN
(b) 35.76 kN

178 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013
| First cracks
100
| Second cracks
| Third cracks

75
Cracking at 69.96 kN a)
Crack propagation in to compression

Load (kN)
| First cracks 50
| Second cracks
| Third cracks

25
Cracking at 78.24 kN RB2
(b) RB1
Foley & Buckhouse Experiment
Fig. 17 RB1 - Cracks after yielding of steel reinforcement at load 0
(a) 69.96 kN (b) 78.24 kN 0 50 100 150
Deformation (mm)
78.24 kN
| First cracks Fig. 19 Load deformation behavior of RCC beam with and withour
| Second cracks
| Third cracks hanger bars

Cracking at 84.00 kN (a) Table 5


Comparisons of FEA and experimental results
| First cracks
| Second cracks Beams Avg. load Deflection Avg. Avg.
| Third cracks
at first at first failure Deflection
cracking cracking load at failure
Cracking at77.00 kN (b) Foley & 20.430 kN 1.34 mm 74 kN 093 mm
Buckhouse
Fig. 18 Cracks at ultimate failure (a) RB1 at load 84.00 kN (b)
RB2 at load 77.00 kN RB1 18.735 kN 0.95 mm 84 kN 128 mm
RB2 19.300 kN 1.00 mm 77 kN 026 mm
The load-deformation behavior of RCC beam is as The study of crack patterns reveals the load at
shown in Fig. 19. All the three stages: linear elastic, which initial crack, flexural crack, flexural diagonal
crack propagation and plastic stage are simulated cracks is started. Finite element software is evaluated
exactly. The point at which the RCC beam collapse by comparing the mechanical behavior of RCC
is identified by insoluble convergence failure of the beam with and without hanger with the Buckhouse
solution. experiment. The crack pattern depends on the area of
The outcome of the analysis is compared with the longitudinal reinforcement provided. The larger the
experimental results as shown in Fig. 19. It is observed area of reinforcement beam fails to show larger cracks
that 3D nonlinear FE analysis of RCC beam almost (more warning) and smaller the area of reinforcement
matches with the experimental results, showing stiff beam fails showing a few cracks (less warning). Hence,
and ductile response. beam with hanger bar was considered to study the
Table 5 gives a numerical comparison between RCC crack pattern for various area of reinforcement and at
and Experimental results. The result also indicates that the same time under, balanced and over sections are
RCC beam losses it’s ductility provision of hanger bars. identified by comparing the mechanical behavior of the
In the next numerical example, RCC beam RB1 is post beam. Later, the flexural capacity of RCC beam is post
strengthened with CFRP attached in tension region. strengthened using CFRP. A parametric study was also
Also, a parametric study is done for various lengths of made considering various lengths CFRP.
CFRP. The behavior of RCC beam RB2 is studied by A 3D shell element was used for the modeling of
varying the area of steel reinforcement in the tension CFRP. Beam with and without hanger bar considered
region. Also, beam RB2 are strengthened with CFRP for the post strengthening. The interface between beam
for different cases of area steel reinforcement. and CFRP is considered as perfect bond. The crack

Journal of Structural Engineering 179


Vol. 40, No. 2, June - July 2013
pattern simulated and mechanical behavior of beam property (Unidirectional property), ultimate strength
post strengthened with CFRP are compared. is higher for the RBC2 compared to RBC1 indicating
CFRP losses stiffness as the length exceeds beyond the
Numerical Example 2 critical length. Also, the beam becomes less ductile by
In the second study, RCC beam RB1 are strengthened provision of CFRP with isotropic property compared
using CFRP attached in tension region. Using same with orthotropic property.
geometry and reinforcement shown in RB1, details of CFRP with Orthotropic property (Fig. 23a, c and e):
the CFRP are shown in Fig. 20. Two different material Cracks in RCC beam is directly depends on the length
properties were considered for the CFRP: isotropic and of the CFRP, larger the length of CFRP more the cracks
orthotropic. It is assumed that a perfect bonding exists in the beam showing larger load carrying capacity and
between the RCC beam and CFRP. ductility. But, beam losses its ductility with decrease in
Three different lengths of CFRP are considered: length of CFRP.
Effective length, half of the clear span and three fourth CFRP with Isotropic property (Fig. 23b, d and f):
of the clear span. CFRP are attached at the bottom Cracks in RCC beam is not directly depends on the length
corner of the RCC beam, as it is observed that initiation of the CFRP. More cracks are identified for CFRP length
of cracks starts in this region. Only one CFRP was used equal to half the clear span of the beam. Indicating beam
FE modeling taking advantage of symmetry. loses its stiffness beyond the critical length.
The behavior of RCC beam strengthened with CFRP Comparing both isotropic and orthotropic property
is shown in Fig. 21 and Fig. 22. It is obvious that, as of CFRP it is observed that, for smaller length of CFRP
the length of CFRP increases ultimate strength of the with isotropic property shows higher load carrying
RCC beam increases. In case of CFRP with isotropic capacity and ductility than the orthotropic property.

P P

CFRP: 50mmx1.2 mmx4650 mm

4800 mm
(a)

CFRP : CFRP: CFRP:


50mmx 50mmx 50mmx
1.2mmx 1.2mmx 1.2mmx
4650 mm 3600 mm 2400 mm

RBC1 RBC2 RBC3

(b)

Fig. 20 RCC beam strengthened with CFRP at tension region (a) Longitudinal details of RBC1 (b) cross sectional details for various lengths
of CFRP

180 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013
200 ultimate and ductile with the increase in the area steel
reinforcement up to a critical value. Beyond this value
the load deformation behavior becomes brittle with an
150
increase in the yield and ultimate strength value. Hence,
area of reinforcement equal to the critical value is termed
Load (kN)

100 as balanced reinforced section, below termed as under


reinforced section and above termed as over reinforced
50 RB1 section. Area steel reinforcement considered in the case
RBC1 + Orthotropic CFRP + Perfect Bond RB1 and RB2 indicates an under reinforced section.
RBC2 + Orthotropic CFRP + Perfect Bond
RBC3 + Orthotropic CFRP + Perfect Bond
0
0 50 100 150
Deformation (mm)

Fig. 21 Load deformation behavior of RCC beam strengthened | First cracks | Second cracks | Third cracks
with CFRP of varying length (a)

200

150
| First cracks | Second cracks | Third cracks
Load (kN)

(b)
100

50 RB1
RBC1 + Isotropic CFRP + Perfect Bond
RBC2 + Isotropic CFRP + Perfect Bond
RBC3 + Isotropic CFRP + Perfect Bond | First cracks | Second cracks | Third cracks
0 (c)
0 50 100 150
Deformation (mm)
Fig. 22 Load deformation behavior of RCC beam strengthened
with CFRP of varying length

Numerical Example 3 | First cracks | Second cracks | Third cracks


(d)
The objective in this numerical example is identifying
the under, balanced and over reinforced section by
increasing the area of steel reinforcement for the beam
RB2. Later the flexural capacity of typical under,
balanced and over a section of RCC beam RB2 is | First cracks | Second cracks | Third cracks
post strengthened using CFRP in the tension region to (e)
compare the material behavior.
Hence, geometry and reinforcement details of RB2
are used in this study as shown in Fig. 26. In the earlier
study it is observed that CFRP with orthotropic property
is more effective than isotropic property. The length of | First cracks | Second cracks | Third cracks
(f)
CFRP is taken as the effective length of the RCC beam
comparing the earlier studies. Fig. 23 Cracks at ultimate failure for beam (a) RBC1 + Orthotropic
property (b) RBC1+ Isotropic property (c) RBC2 +
The load deformation behavior of RCC beam with Orthotropic property (d) RBC2+ Isotropic property (e)
increase area of steel reinforcement is shown in Fig.24. RBC3 + Orthotropic property (f) RBC3 + Isotropic
Load deformation behavior shows increased yield, property

Journal of Structural Engineering 181


Vol. 40, No. 2, June - July 2013
200 a large numbers of green and red circle outlines in the
constant moment region. In balanced reinforced section
the failure of the beam is due to compression failure
150 of the concrete. It is shown by a red circle outline in
the compression region. In over reinforced section
Load (kN)

the failure of the beam is due to the shear failure of


100 concrete. It is shown by a large numbers of inclined red
circle outline between the support and loading points.

Over Section - 1972 sq. mm 2 - 12 ø


50 Over section - 1294 sq. mm
Balanced Section - 1080 sq. mm
Under Section - 824 sq. mm
RB2 - Under Section - 603 sq. mm CFRP:
0
0 12.5 25 37.5 50 50mmx
Deformation (mm) 1.2mmx

Fig. 24 Load Deformation plot for RCC beam with varying area of 4650 mm
varies
steel reinforcement

300 Fig. 26 RCC beam with CFRP attached

250

200
Load (kN)

| First cracks | Second cracks | Third cracks Tensile failure of concrete


150 (a) beam

100

50
| First cracks | Second cracks | Third cracks Compression failure of concrete
(b) beam
0
0 25 50 75 100
Deformation (mm)
Over Section - 1972 sq. mm + Orthotropic Over Section - 1972 sq. mm
CFRP
Balanced Section - 1080 sq. mm + Balanced Section - 1080 sq. mm | First cracks | Second cracks | Third cracks Compression failure of concrete
Orthotropic CFRP
(c) beam
Under Section - 603 sq. mm + Orthotropic RB2 - Under Section - 603 sq. mm
CFRP Shear failure of concrete beam

Fig. 25 Load deformation behavior of RCC beam strengthened


with CFRP

The crack pattern for the under, balanced and over | First cracks | Second cracks | Third cracks

reinforcement are shown in Fig. 27a through Fig. 27d (d)

and Fig. 28e through Fig. 28h. It is observed that, Fig. 27 crack at ultimate failure (a) Under reinforced section – 603
failure of the beam in under reinforced section is due to mm2 (b) under reinforced section – 824 mm2 (c) balanced
section – 1080 mm2 (d) over reinforced section – 1294
flexure dominated cracks, tensile failure. It is shown by mm2

182 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013
Shear failure of concrete 3. Suidan M, Schnobrich W C. “Finite Element
Analysis of Concrete”. ASCE Jl. of the Structs.
Divi. 1973, Vol. 99, pp 3257–3276.
4. Cope. R.J, Rao P.V., Clark L.A., Norris P.,
| First cracks | Second cracks | Third cracks “Modeling of Reinforced Concrete Behaviour
(a) for Finite Element Analysis of Bridge Slabs”.
Numerical Methods for Non-Linear Problems
1980, Vol. 1, pp 457–470.
5. De Borst R., “Some recent development in
| First cracks | Second cracks | Third cracks
computational modeling of concrete fracture”.
(b) Intl. Jl. of Fracture, 1997, Vol. 86, pp 5–36.
6. Menini R.C.G, Trautwein L.M, Bittencourt T.N.
“Smeared crack models for reinforced concrete
beams by finite element method”. IBRACON
Structs. and Mat. Jl. 2009, Vol. 2, (2), pp 166-200.
| First cracks | Second cracks | Third cracks
(c) 7. Bazant P.Z, Tabbara M.R, Kazemi M.T., Pijaudier-
Cabot G. Random Particle Model for Fracture of
Aggregate or Fiber Composites. Jl. of Engg. Mech.
1990, Vol. 119, (8), pp 1686–1705.
8. Schlangen. E, Garboczi. E.J., “Fracture Simulation
| First cracks | Second cracks | Third cracks
(d)
of Concrete using Lattice Models: Computational
Aspects”. Engg. Fracture Mech., 1997, Vol.57
Fig. 28 crack at ultimate failure (e) over reinforced section – 1972
(2/3), pp 319–332.
mm2 (f) under reinforced section 603 mm2 with CFRP
(g) balanced section – 1080 mm2 with CFRP (h) over 9. Van Mier, J.G.M, Van Vliet M.R.A, Wang, T.K.,
reinforced section 1972 mm2 with CFRP “Fracture Mechanisms in Particle Composites:
Statistical Aspects in Lattice type Analysis”.
Summary Mech. of Mat., 2002, Vol. 34, pp 705–724.
10. Van Mier, J.G.M, and Van Vliet M.R. A.
In this present work a crack modeling methodology in “Influence of Microstructure of Concrete on Size/
three dimensions for carbon fiber reinforced polymer Scale Effects in Tensile Fracture”. Engg. Fracture
(CFRP) strengthened RCC beam is made by performing Mech. 2003, Vol. 70, pp 2281–2306.
a three dimensional nonlinear finite element analysis of 11. Leite J.P.B, Slowik, V., Apel J., “Computational
the beam subjected to four point loading. The results Model of Mesoscopic Structure of Fracture
are comparable with the experimental results available Processes”. Comp. and Structs., 2007, Vol. 85,
in the literature. FE analysis of RCC beams without pp 1293–1303.
(RB1) and with (RB2) hanger bars are carried out and
12. Brighenti, R., Carpinteri, A, Spagnoli, A., Scorza
the results are compared with the experiments.
D., “Cracking Behavior of Fibre-Reinforced
Cementations Composites: A Comparison between
References
a Continuous and a Discrete Computational
1. Ngo. D, and Scordelis A.C. Finite Element approach”. Engg. Fracture Mech., 2012,
Analysis of Reinforced Concrete Beams. “Jl. of doi:10.1016/j.engfracmech.2012.01.014.
the Ame. Inst”. 1967; Vol. 67, pp 152–163. 13. Meier U. Strengthening of structures using
2. Rashid Y R. “Analysis of Prestressed Concrete carbon fiber/epoxy composites. Construction and
Pressure Vessels”. Nuclear Engineering and Building Materials 1995; 9(6): 341-351.
Design, 1968, Vol. 7, pp 334–344. 14. Etse G, Caggiano A, Vrech S., “Multiscale Failure

Journal of Structural Engineering 183


Vol. 40, No. 2, June - July 2013
Analysis of Fiber Reinforced Concrete based on a 1974, pp 174–204.
Discrete Crack Model”. Int. J. Fract. 2012; DOI: 24. ANSYS 13.0. Finite Element Analysis System.
10.1007/s10704-012-9733-z. SAS IP. 2010.
15. Teng, J.G., Chen, J.F., Smith, S.T., Lam, L., 25. Kwak, H.G, Filippou F.C., “Nonlinear FE Analysis
“Behaviour and strength of FRP-strengthened RC of R/C Structures under Monotonic Loads”. Comp.
structures: a state-of-the-art review”. Structs. and & Structs. 1997, Vol. 65, (1), pp 1–16.
Build. 2003, Vol. 156, (1), pp 51–62.
26. Yasmen T.O., Susanne, H., Ola, D., “The effect
16. Buyukozturk, O, Hearing, B., “Failure behavior of of CFRP and CFRP/concrete interface models
precracked concrete beams retrofitted with FRP”. when modeling retrofitted RC beams with FEM”.
Jl. of Compo. for Construct., 1998, Vol. 2(3), Compo. Structs. 2010, Vol. 92, pp 1391–1398.
pp 138–144.
27. Yasmen, T.O., Susanne, H., Ola, D., “Nonlinear
17. Mohamed Ali, M.S., Oeslers, D.J., Griffith, M.C., FE modeling of shear behaviour in RC beams
Seracino, R., Interfacial Stress Transfer of Near retrofitted with CFRP”. Proc. of Comput. Model.
Surface-Mounted FRP-to-Concrete Joints. Engg. of Conc. Structs., 2010, pp 671–677.
Structs., 2008, Vol. 30, pp 1861–1868.
28. Tavarez, F.A., “Simulation of Behavior of
18. Lorenzis, D.L., Teng, J.G., “Near Surface Mounted Composite Grid Reinforced concrete Beams
FRP Reinforcement: An Emerging Technique for Using Explicit Finite Element Methods”. Master’s
Strengthening Structures”. Composites Part B Thesis. University of Wisconsin-Madison: 2001.
2007, Vol. 38, pp 119–143.
29. Jacob Logan Julander. “Finite Element Modeling
19. Dhamani, L., Khennane, A., “Crack Identification of Full Depth Precast Concrete Transverse bridge
in Reinforced Concrete Beams using Ansys deck connections”. Master’s Thesis. UTAH State
Software”. Strength of Mat., 2010, Vol. 42(2), University: 2009.
pp 232–240.
30. Kachlakev, D.I., Miller, T., Yim, S., Chansawat
20. Saifullah, I., Nasir-uz-zaman, M., Uddin, S.M.K, K., Potisuk, T., “Finite Element Modeling of
Hossain, M.A., Rashid, M.A., “Experimental and Reinforced Concrete Structures Strengthened
Analytical Investigation of Flexural Behavior of With FRP Laminates”. Report SPR 316. California
Reinforced Concrete beam”. Intl. Jl. of Engg. & Polytechnic State University: 2001.
Tech., 2011, Vol. 11, (01), pp 188–196.
31. Reddy, J.N., “An Introduction to Nonlinear Finite
21. MacGregor, J.G., “Reinforced Concrete: Mechanics Element analysis”. Oxford University Press Inc.
and Design, 3rd Ed”. Prentice-Hall: 1997. 2005.
22. Anthony J. Wolanski, B.S., “Flexural Behavior 32. Foley, C.M, Buckhouse, E.R. “Strengthening
of Reinforced and Prestressed Concrete Beams Existing reinforced concrete beams for flexure
Using Finite Element Analysis”. Master’s Thesis. using bolted external structural steel channels”.
Marquette University: 2004. Report MUST-98-1. Marquette university Struct.
23. Willam, K.J., Warnke, E.P., “Constitutive Engg., 1998.
Model for Triaxial Behaviour of Concrete”. Intl.
Association of Bridge and Struct. Engg. Conf., (Discussion on this article must reach the editor before
September 30, 2013)

184 Journal of Structural Engineering


Vol. 40, No. 2, June - July 2013

You might also like