Qi Zhang
Faculty of Engineering
April, 2004
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Contents
Introduction__________________________________________________________ 3
Reference__________________________________________________________ 43
Page 2 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Introduction
In recent years there has been increased interest in the use of fiber-reinforced polymers (FRP)
for concrete structures. As one of the new promising technologies in construction, FRP material
solves the durability problem due to corrosion of steel reinforcement. Glass fiber reinforced
polymer (GFRP) composites have been used for reinforcing structural members of reinforced
concrete bridges. Many researchers have found that GFRP composite reinforcement instead of
steel is an efficient, reliable, and cost-effective means of reinforcement in the long run.
In the structural elements, the flat slab has the large surface exposed to the outside corrosive
environment, such as bridge decks, ocean oil flat slabs and parks. The flat slab system offers
advantages for efficient design, the overall construction process, notably in simplifying the
installation of services and the savings in construction time. However, a slab-column connection
in the flat slab system is frequently subjected to the significant transverse shear forces, which
can produce a punching shear failure. Punching in the vicinity of a column is a possible failure
mode for reinforced concrete flat slabs. Many researchers have exerted their efforts to
investigate the punching strength of slabs, due to the undesirable suddenness and catastrophic
results of punching failure. However, there is no generally accepted treatment for the punching
resistance of flat slabs because of the complicated dependence of shear strength on their
flexural behavior as well as the fact that it is difficult to observe their internal inclined cracks.
Page 3 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Furthermore, most FRP elements exhibit a relatively low modulus of elasticity and lack the
material coefficient to the models and equations based on steel reinforcement. Thus, it is
motivated to improve the understanding of the punching failure mechanisms and to establish a
reliable method for predicting the punching strength of column-slab connections reinforced with
GFRPs.
This paper, therefore, will present the application of the finite element method for the numerical
modeling of punching shear failure mode using a widely accepted code, ANSYS [ANSYS, 1998].
Based on properly modeling and simulating the experiments [Rashid, 2004] carried out in the
Page 4 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
A n extensive description of previous studies on the application of the finite element method to
the analysis of reinforced concrete structures and the underlying theory and the application of the
finite element method to the analysis of linear and nonlinear reinforced concrete structures is
presented in excellent state of-the-art reports by the American Society of Civil Engineers in 1982
[ASCE 1982]. The results from the FEA are significantly relied on the stress-strain relationship of
the materials, failure criteria chosen, simulation of the crack of concrete and the interaction of the
Because of these complexity in short- and long-term behavior of the constituent materials, the
ANSYS finite element program (ANSYS 1998), operating on a Windows 2000 system, introduces
a three-dimension element Solid65 which is capable of cracking and crushing and is then
combined along with models of the interaction between the two constituents to describe the
behavior of the composite reinforced concrete material. Although the Solid65 can describe the
reinforcing bars, this study uses an additional element, Link8, to investigate the stress along the
reinforcement because it is inconvenient to collect the smear rebar data from Solid65. Due to the
general experiment process in the structural lab, the edge of slab is free to lift, which is different
with the simple support along all four edges according to the lines of contra flexure. Then, a
spring element, Link10, along the edge, is included in this study to reflect the actual setup of
Page 5 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
slab-column connection. The constructed model using ANSYS is shown in Fig.1. The detail of
Page 6 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
The assumptions made in the description of material behavior are summarized below:
• The stiffness of concrete and reinforcement is formulated separately. The results are then
• The smeared crack model is adopted in the description of the behavior of cracked concrete.
Cracking in more than one direction is represented by a system of orthogonal cracks, and the
• The reinforcing GFRP is assumed to carry stress along its axis only and the perfect bond
Solid65, an eight-node solid element, is used to model the concrete with or without reinforcing
bars. The solid element has eight nodes with three degrees of freedom at each node
translations in the nodal x, y, and z directions. The element is capable of plastic deformation,
cracking in three orthogonal directions, and crushing. The geometry and node locations for this
Page 7 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
The eight nodal points is also along with a 2X2X2 Gaussian integration scheme. This is used for
the computation of the element stiffness matrix. The element’s displacement field in terms of the
(1)
(2)
(3)
mechanical properties scatter widely and cannot be defined easily. For the convenience of
analysis and design, however, concrete is often considered a homogeneous material in the
Page 8 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
compression, the time-dependent effects of creep, shrinkage and temperature variation, the
nonlinear response of concrete members can be observed (Fig.2). This highly nonlinear
response can be roughly divided into three ranges of behavior: the uncracked elastic stage, the
Fig.2: Typical uniaxial compressive and tensile stress-strain curve for concrete (Bangash 1989)
Before concrete cracking takes place, the behavior of concrete could be regarded as a linear
(4)
The nonlinear response of concrete is mainly controlled by progressive cracking that results in
Page 9 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
localized failure. After concrete cracking takes place, the critical section of the structural member
is weakened, and then the stress of concrete and reinforcement will be redistributed. Different
with the discrete cracks model [Ngo and Scordelis, 1967], in which crack occurs during a load
cycle but the need to change the topology and the redefinition of nodal points of concrete model
greatly reduce the speed of the process, Solid65 element automatically generates cracking
without redefining the element mesh. These models depict the effect of many small cracks that
are “smeared” across the element in a direction perpendicular to the principal tensile stress
Observing the Fig.2, the stress-strain curve for concrete is linearly elastic up to about 30 percent
of the maximum compressive strength. Above this point, the crack is developed in the concrete
and the stress increases gradually up to the maximum compressive strength. The stress-strain
relations are modified in this stage to represent the presence of a crack in the concrete. A plane
of weakness in a direction normal to the crack face and a shear transfer coefficient t are
introduced in the solid65 element. The shear strength reduction for those subsequent loads
which induce shear across the crack surface is considered by defining the value of t. This is very
important to accurately predict the loading after cracking, especially when calculating the
strength of concrete member dominated by shear, such as slab-column connection and shear
wall. The detailed analysis of the shear transfer coefficient will be presented later.
The stress-strain relations for a material that has cracked in one direction only become:
Page 10 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
(5)
where the superscript ck signifies that the stress strain relations refer to a coordinate system
parallel to principal stress directions with the xck axis perpendicular to the crack face. Rt, the
slope as defined in Fig.3, works with adaptive descent and diminishes to 0.0 as the solution
converges.
If the crack closes, then all compressive stresses normal to the crack plane are transmitted
across the crack and only a shear transfer coefficient c for a closed crack is introduced. Then
Page 11 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
(6)
(7)
The matrices and equations of the stress-strain relations for concrete that has cracked in more
than one direction are detailed in the theory reference of ANSYS, 1998.
In addition to cracking, the concrete will be failed in uniaxial, biaxial, or triaxial compression. The
concrete is assumed to crush in that condition. A three-dimensional failure surface for concrete is
shown in Fig. 4. The most significant nonzero principal stresses are in the x and y directions,
represented by
xp and yp, respectively. Three failure surfaces are shown as projections on
the xp- yp plane. The mode of failure is a function of the sign of zp (principal stress in the z
Page 12 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Fig.4 Failure Surface in Principal Stress Space with Nearly Biaxial Stress
ANSYS requires input data for material properties of Solid65 element as elastic modulus (Ec),
ultimate uniaxial compressive strength ( f c' ), ultimate uniaxial tensile strength (modulus of
rupture, ( f r ), poisson’s ratio ( ), density ( γ ), shear transfer coefficient for an open crack ( t),
shear transfer coefficient for an close crack ( c), compressive uniaxial stress-strain relationship
for concrete. If the Solid65 element includes the representation of reinforcement, up to three
rebar could be defined in Solid65 element through the real constants: reinforcement material
number (Matn), volume ratio (VRn) and orientation angles (THETAn and PHIn), where n
MacGrewgor (2000) defined the elastic modulus and the ultimate uniaxial tensile strength ( f r )
as follows:
Page 13 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
γ
1.5
Ec = 3600 f c + 6700
'
(7)
2300
f r = 0.6 f c
'
(8)
The density ( γ ) and the poisson’s ratio ( ) of concrete are considered as 2300 kg/mm3 and 0.2.
The ultimate uniaxial compressive strength ( f c' ) was measured through cylinder concrete test
(Rashid, 2004).
The shear transfer coefficient for a closed crack c is widely accepted within 0.9 to 1.0. The shear
transfer coefficient, t represents conditions of the crack face. The value of t ranges from 0.0 to
1.0, with 0.0 representing a smooth crack (complete loss of shear transfer) and 1.0 representing
a rough crack (no loss of shear transfer) [ANSYS 1998]. The value of t used in many studies of
reinforced concrete structures, however, varied between 0.05 and 0.25 [Bangash 1989; Huyse,
et al. 1994; Hemmaty 1998]. A number of preliminary analyses were attempted in this study with
various values for the shear transfer coefficient within 0.125-1.0, the results are shown in the
Fig.5. Although there is no significant difference of the ultimate loading with the different t,
accompanying with the increase of t, the ultimate loading is slightly increased after the
Page 14 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
250000
200000
Bt0.8_Bc1.0
ultimate loading (N)
150000 Bt0.5_Bc1.0
Bt1.0_Bc1.0
100000
Bt0.3_Bc0.9
Bt0.125_Bc0.9
50000
0
0.0 2.0 4.0 6.0 8.0 10.0 12.0
displacem ent (m m )
It is possible that the ultimate loading will have larger divergence if the larger displacement takes
place. GFRPs with the lower modulus of elasticity will lead to the larger deflection than the
normal structure member reinforced with steel. The experiment conducted by Rashid [Rashid,
2004] proves this assumption. Thus, the ultimate loading should be sensitive to t. Fig. 6
simulated the load-deflection curve according to the specimen GSHS2, which proves the larger
difference of ultimate loading and deflection due to different t in the condition of displacement
more than 15mm. Since convergence problems were encountered at low loads with t less than
0.1. Therefore, the shear transfer coefficient t used in this study was equal to 0.2, which gives a
good agreement with the change of stiffness of the specimen as well as the ultimate loading.
Page 15 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
formulas have been proposed. These are summarized in ASCE (1982). Fig.7 shows the simplest
of the nonlinear models, the linearly elastic-perfectly plastic model, which was used by Lin and
Scordelis (1975) in a study of reinforced concrete slabs and walls. Fig.7 also shows a piecewise
straight-line segments. Although this is the most versatile model capable of representing a wide
range of stress-strain curves, the softening branch of the concrete stress-strain relation is found
Page 16 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
40
30
stress (MPa)
20
piecew ise linear
model
10 elastic-perfectly
plastic model,
0
0 0.001 0.002 0.003 0.004 0.005
strain
Due to this unique problem in the finite element analysis of slab-column connections, two
solution strategies have been applied in this study. The one is applying the piecewise linear
model without defining the crush of concrete, the ultimate uniaxial compressive strength ( f c' ); the
another is involved with the crush of concrete but using the elastic-perfectly plastic model. After
comparing this two approach with the simple model (Fig. 11), the results in the Fig. 8 shows the
advantage of the latter. The elastic-perfectly plastic model gives a good prediction of ultimate
Page 17 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
400
GSHD2
Displacement (mm)
300
piecew ise
linear model
200
elastic-
perfectly
plastic model,
100
0
0 5 10 15 20 25 30 35 40
with three degrees of freedom at each node: translations in the nodal x, y, and z directions. As in
a pin-jointed structure, no bending of the element is considered. The element is also capable of
plastic deformation, stress stiffening, and large deflection. The geometry, node locations, and the
Page 18 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
GFRP reinforcement in the experimental slab-column connections was made of typical GFRP
material. Through the tension test of GFRP reinforcement, the properties including elastic
modulus and ultimate tensile stress in this FEM study is as Fig. 10.
300
200
100
0
0 0.004 0.008 0.012 0.016
Strain
Page 19 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Beside the discrete reinforcement model adopted in this study, the reinforcement could be also
defined using the smeared reinforcement option of the Solid65 element. The amount of
reinforcement is defined by specifying a volume ratio (VRn) and the orientation angles (THETAn
and PHIn) of the rears. This approach is easy to construct the reinforced concrete model, which
has the uniform or simple reinforcement arrangement. In practice, the reinforcement in flat slabs
is normally arranged in a uniform form in different slab strips. Therefore, it is suggested to use
this smear reinforcement option when simulating the large slab model. However, the stress and
location of reinforcement is difficult to obtain in the smear model. In this study, the behavior of
slab-column connections was investigated by examining the stress distribution in the whole
specimen. The stress of the reinforcement is critical to this investigation. The discrete
reinforcement model is used in the study. On the other hand, it is motivated to construct smear
reinforcement model for comparing the discrete one, then strengthen the understanding the
difference between these two models and prepare for the future application.
For the comparison purpose, four simple models are constructed for predicting the
time. The simple model doesn’t consider the layer mesh, spring support and the column stub. In
Fig. 11 there are four models. The first three are the smear reinforcement models with different
thickness of the reinforcement layer, including the thickness of 20mm, 40mm and 100mm. The
Page 20 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
last one is the discrete reinforcement model with the reinforcement spacing of 240mm. The test
It is obvious that the four models obtain the similar displacement and ultimate loading, but the
smear models got a smoother curve. In the general building codes in the world, the major
concerns of punching strength of slabs are the reinforcement ratio, the load area, the effective
depth and compressive strength of concrete. The effect of the arrangement of reinforcement on
the punching strength is ignored, which is verified in this test. Therefore, the reinforcement
spacing of the simulated models is taken as 120mm and 240mm for constructing the model
consistent with the mesh size and column size. This is different with the reinforcement spacing of
the actual tested specimens, but the reinforcement ratio is kept same.
Page 21 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
a).Smear Model with 20 mm thickness Rebar Layer b).Smear Model with 100 mm thickness Rebar Layer
Page 22 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
140000
120000
Ultimate loading (N)
100000
80000
60000
40000
20000
0
0 5 10 15 20
displacement (mm)
LINK10 element is a three-dimensional spar element having the unique feature of a bilinear
stiffness matrix resulting in a uniaxial tension-only (or compression-only) element. With the
tension-only option, the stiffness is removed if the element goes into compression (simulating a
slack cable or slack chain condition) [ANSYS, 1998]. The geometry, node locations, and the
Page 23 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
In this study, since the effect of gravity of slab-column connection on the punching strength of
specimens is little, the simulated model is constructed in the form of one quarter of the slabs due
to the two axes of symmetry. Thus, the boundary condition of these two edges is defined as the
Because the specimens are simply supported on four edges that are free to lift during the tests,
the proper simulation of the boundary condition is taken into account. Link10 elements, the
special nonlinear spring elements in the transverse direction are employed along four edges of
labs in the numerical model, which is illustrated in Fig.1. The stiffness of spring elements is
Furthermore, the simple support in the test is a roller support made of the steel tube and covered
with 3mm-thickness rubber. Both the steel tube and rubber will be deformed during the test, the
stiffness of spring would have been chosen carefully to represent the test simple support in the
Page 24 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
condition of relatively large deflection, so that the similar load-deflection curve to the tested
specimens could be obtained. However, the test setup tries to simulate the slab-column
connection around the conflexure line. In spite of the careful setup, the specimen could not
reflect the real deformation perfectly, while the model using ANSYS could exactly simulate this
theoretical specimen. Therefore, simply simulating the test specimen spring support could not
represent the real situation. In this study, the comparison between the specimens and models,
including the change and development of stiffness of slabs, the crack loading and the ultimate
loading are the major concern. This model can be regarded as a good reference to the real
After constructed a model with volumes, areas, lines and key points, a finite element analysis
requires meshing of the model. The model is then divided into a number of small elements, and
after loading, stress and strain are calculated at integration points of these small elements (Bathe
1996). Since the FEM could approximate the real situation as close as possible depending on the
selection of the mesh density, it is a significant step in finite element modeling to choose an
appropriate mesh size to meet the requirement of accuracy and speed. A convergence of results
is obtained when an adequate number of elements are used in a model. This is practically
achieved when an increase in the mesh density has a negligible effect on the results (Adams and
Askenazi 1998). Therefore, in this finite element modeling study the appropriate mesh size is
Page 25 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Volume elements can often be either hexahedral (brick) or tetrahedral shaped, but the solid65
hexahedral shape.
The ACI code (ACI, 1995) assume that the control perimeters of punching failure is located at a
distance of 0.5 times the effective depth from the edge of load (column), while the British code
(BS8110, 1985) considers a larger control perimeter, 1.5d. Furthermore, because the spring
support is used in this study, the support failure will not occur. Therefore, the mesh size in this FE
study must be small in the area around the column and the larger size can be acceptable near
the edge.
On the other hand, the column stub, slab and reinforcement are required to work together in the
discrete model used in this study. Then, the size of column and the spacing of reinforcement are
taken into consideration of determining the mesh size. The small change of the load area
(column size) gives the similar results of punching strength of slab-column connections using
either of ACI code or British code. Consistent with the reinforcement spacing 120mm and 240mm
determined in the above chapter, column stub size in the model is reduced from 250mm to
Page 26 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
100000
80000
60000
40000
20000
0
0 5 10 15
Displacement (mm)
Therefore, the mesh size is considered as 60mm or 120mm. Using the simple model described
in the Fig.11 to check the difference of models meshed with the mesh global size of 60mm and
120mm and the mesh thickness size of 50mm, the results are illustrated in the Fig. 14. According
to this figure, it is fair to draw a conclusion that there is no significant effect on the load-deflection
relationship of the model using controlled mesh size of either the 60mm or 120mm. However, the
120mm is relatively bigger to investigate the element stress around the column because the
column size is 120mm; while the model with the mesh size of 60mm is quite time consuming: 4
hours have been spent to run this simple model. In the end, the layer mesh is selected in this
5. Numerical Implementation
The numerical implementation of the finite element model is based on the virtual work principle or
the theorem of minimum potential energy to the assemblage of discrete elements. The following
Page 27 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
(9)
(10)
(11)
(12)
(13)
(14)
In Eqs. 9-14, the components of [N ] are the shape functions, {d } is the vector of node
displacements, {R} is the vector of applied nodal forces, {p} is the vector of surface forces
This is a system of simultaneous nonlinear equations, since the stiffness matrix [K ] , in general,
Page 28 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
depends on the displacement vector {d } . The solution of this system of nonlinear equations is
typically accomplished with an iterative method. The load vector {R} is subdivided into a
number of sufficiently small load increments. At the finishing point of each incremental solution,
the stiffness matrix of the model is adjusted to reflect nonlinear changes in structural stiffness
before proceeding to the next load increment. The ANSYS program (ANSYS, 1998) uses
equilibrium iterations provide convergence at the end of each load increment within tolerance
limits. Fig. 15 shows the use of the Newton-Raphson approach in a single degree of freedom
nonlinear analysis.
Prior to each solution, the Newton-Raphson approach assesses the out-of-balance load vector,
which is the difference between the restoring forces (the loads corresponding to the element
stresses) and the applied loads. Subsequently, the program carries out a linear solution, using
the out-of-balance loads, and checks for convergence. If convergence criteria are not satisfied,
the out-of-balance load vector is re-evaluated, the stiffness matrix is updated, and a new solution
is attained. This iterative procedure continues until the problem converges (ANSYS, 1998).
Page 29 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
In this study, for the reinforced concrete solid elements, convergence criteria were based on
force. In order to obtain fast and accurate convergence of this nonlinear analysis, “Line Search
Option” and “Predictor-Corrector Option “ are set on, and the convergence tolerance limits
increased to a maximum of 10 times the default tolerance limits (0.5% for force checking).
According to the test procedure in the structural lab, the displacement is applied to the specimen
by a hydraulic actuator. In the other word, the actuator imposes the constant displacement to the
slab, while the applied loading is adjusted according to the change of the stiffness of the
specimen. Therefore, the displacement is applied to the column stub in this study for simulating
Page 30 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
For the nonlinear analysis, automatic time stepping in the ANSYS program predicts and controls
load step sizes. Based on the previous solution history and the physics of the models, if the
convergence behavior is smooth, automatic time stepping will increase the load increment up to
a selected maximum load step size. If the convergence behavior is abrupt, automatic time
stepping will split the load increment until it is equal to a selected minimum load step size.
Although the automatic time stepping could ensures that the time step variation is neither too
aggressive nor too conservative, the amount of time step still determines the initial displacement.
In this study, the reinforced concrete is cracked, which leads to dramatically reduce the specimen
stiffness, subsequent to the small deflection of the slab. It is possible to fail to detect the cracking
load if the initial amount of time step is relatively big. Therefore, the amount of time step is set
1000 to ensure to replicate the complete loading process during the test.
Page 31 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Since the reinforced concrete slab-column connection can be regarded as the elastic plate
before the crack forms, the finite element model, simply supported, and subjected to a centrally
concentrated load, was initially investigated. These models simulated the specimen (Table 1)
had the same main parameters. Based on verification of the slabs before cracking, the slabs
were analyzed using plain concrete theory and finite element results were compared with hand
calculations.
Fig. 16a-d show relationship of the deflection and loading of the center on the top of the slabs, as
calculated by the model and by traditional hand calculations (Timoshenko, 1987). As expected,
the deflection is linear with loading, and the finite element model shows excellent correlation with
the hand calculation before the cracks form in the model. However, the models shown in the Fig.
16e and f are a little divergent from the hand calculations. This is because the applied uniform
load in the column area is regarded as central load on the slab. When the slab is relative thin, the
uniform load is well consistent with the simulated central load. However, the consistency is
weakened as the depth of slab increases, which shows in Fig 16e and f.
Page 32 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
120 90
80
100
70
Loading (KN)
80
Loading (KN)
60
50
60
40
40 GS1_Tested 30 GS3_Tested
ANSYS_GS1 20 ANSYS_GS3
20
10
0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 1
Deflection (mm) Deflection (mm)
Fig. 16a: Finite Elem ent Model Verification Fig. 16b: Finite Elem ent Model Verification
in the Elastic Stage in the Elastic Stage
160 140
140 120
120
100
Loading (KN)
Loading (KN)
100
80
80
60
60
GSHS1_Tested GSHS2_Tested
40
40
ANSYS_GSHS1 A_GSHS2
20 20
0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8
Deflection (mm) Deflection (mm)
Fig. 16e: Finite Elem ent Model Verification Fig. 16f: Finite Elem ent Model Verification
in the Elastic Stage in the Elastic Stage
Page 33 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
250
200
Loading (KN)
150
100
50 GSHD1_Tested
ANSYS_GSHD1
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Deflection (mm)
250
200
Loading (KN)
150
100
50 GSHD2_Tested
ANSYS_GSHD2
0
0 0.1 0.2 0.3 0.4 0.5 0.6
Deflection (mm)
Further verification of the validity of finite element models using non-linear analysis may be
demonstrated by comparing the predicted response of the model with experimental results
obtained from laboratory tests. Load deflection behavior obtained from the model, was compared
Page 34 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Based on the preparation and analysis for the slab-column connections constructed by ANSYS,
six finite element models were conducted for the comparison of the measured ultimate loads and
the associated deflection at the center of the tested slabs. All slabs were reinforced with GFRP
bars. The length and width of all the slabs were 1920 mm with the thickness of either 200 mm or
150 mm. Since the two-axis symmetry, the finite element models were only a quarter of the slabs.
A concrete cover of 50 mm was used for all the slabs. Due to the brittle flexural failure of the
concrete members reinforced with GFRP bars, all the slabs were initially designed to be
over-reinforced, using the reinforcement ratio more than the balanced reinforcement ratio b.
This also achieved the punching failure because of high reinforcement ratio (Menetrey, 1998).
The slabs of GS1 and GS3 were designed for investigating the effect of reinforcement ratio on
the punching loading on the normal slab-column connection. The size effect was examined using
the slabs of GSHD1 and GSHD2 with the high depth. And the application of the high strength
concrete to slabs reinforced with GFRP was analyzed based on the slabs of GSHS1 and GSHS2.
With the application of the three-dimensional solid element, there was reasonable agreement
between the predictions by the finite element model and the measurement by the tests in
Page 35 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
slab-column connections. The ratios of predicted-to-measured ultimate loads were in the range
form 0.82 to 1.13; the average value was 0.98 with standard deviation of 10.8. The ratios of the
associated deflections at the center of the slabs ranged from 0.68 to 0.98; and the average value
was 0.85 with standard deviation of 12.4. Table 2 presents the comparison results and Fig. 17a-d
show the comparison of the three-dimensional finite element model prediction to the test results.
It is noted from the finite element analysis that the 40% increase of reinforcement ratio only
increases around the 12% punching capacity of slab-column connections. This proves the cubic
root relationship of the reinforcement ratio and punching capacity. This relationship is defined in
the BS8110 code for the steel as reinforcement. According to the results, even although the
GFRP rears reinforcing concrete structure have the linear elastic constitutive relationship and
relatively low modulus of elasticity, different from the traditional steel reinforcement, the strength
of slab-column connections reinforced with GFRP rears could be predicted by the BS8110 code if
Page 36 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Page 37 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Page 38 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Since the finite element model provides the considerable accurate prediction of the strength of
slab-column connections, some traditional confusion for investigating two-way slabs using
experimental approach could be overcome based on the finite element approach, for example,
parameter study, the crack phenomenon and the form of punching cone, the failure mechanism
of two-way slabs, the conflict of the punching shear prediction using major codes such as ACI
Page 39 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Since the two-way slab reinforced with GFRPs are a new reinforced concrete member, there
have not been an accepted approach to predict the ultimate load in the building codes.
EI-Ghandour (EI-Ghandour, 2003) has proposed an approach to modify the equation in the ACI
code using a term referring to the reinforcement stiffness to the power of 0.33; on the other hand,
he proposed to modify the equation multiplying a correction factor . In this study, the finite
element analysis results were compared with EI-Ghandour’s modified code predictions. This is
given in the Table 3. It is obvious that the finite element results are consistent to the equation of
the modified BS8110 code for the normal concrete strength slabs, but the equation of modified
Page 40 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
A quarter of the full-size slab-column connections, with proper boundary conditions, were used in
ANSYS for modeling to reduce computational time and computer disk space requirements.
Concrete constitutive relationship included the elastic-perfectly plastic model, crack condition
and crush limit. GFRP reinforcement was defined linear elastic. Spring supports were the
Based on the comparison of smear reinforcement model and discrete reinforcement model, the
later was used in this study because it is convenient to obtain the reinforcement information.
The layer mesh was used to obtain the enough information from critical section as well as to
For nonlinear analysis in this study, the total displacement applied to a model was divided into a
number of load steps. Sufficiently small load step sizes are required, particularly at changes in
behavior of the reinforced concrete connections, i.e., major cracking of concrete, and
Page 41 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
For closed cracks, the shear transfer coefficient is assumed to be 0.9, while for open cracks it
should be in the suggested range of 0.05 to 0.5 to prevent numerical difficulties. In this study, a
The general behavior of the finite element models represented by the load-deflection plots at
center show good agreement with the test data. However, the finite element models show slightly
more stiffness than the test data in both the linear and nonlinear ranges.
The modified BS8110 code represents the cubic root relationship between GFRP rears and the
punching strength of slab-column connections. Thus, it shows good agreement with the
Page 42 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Reference
Abdel Wahab EI-Ghandour1; Kypros Pilakoutas2; and Peter Waldron (2003), “Punching Shear
Behavior of Fiber Reinforced Polymers Reinforced Concrete Flat Slabs: Experimental Study”,
Journal of Composites for Construction, ASCE, Vol. 7, No 3, August 1, 2003, pp. 258-264
American Concrete Institute (ACI), (1995) “Building code requirements for reinforced concrete
and reinforced concrete and commentary.” ACI 318-95/ACI 318R-95, Detroit.
Adams, V. and Askenazi, A., (1998) “Building Better Products with Finite Element Analysis,”
Santa Fe, New Mexico
ASCE Task Committee on Finite Element Analysis of Reinforced Concrete Structures. (1982).
State-of-the-Art Report on Finite Element Analysis of Reinforced Concrete, ASCE Special
Publications.
British Standards Institution (BS8110), (1985) “Code of practice for design and construction.”
British Standard Institution, Part 1, London.
ANSYS (1998), ANSYS User’s Manual Revision 5.5, ANSYS, Inc., Canonsburg, Pennsylvania.
Bathe, K. J., (1996) Finite Element Procedures, Prentice-Hall, Inc., Upper Saddle River, New
Jersey.
Ngo, D. and Scordelis, A.C. (1967). "Finite Element Analysis of Reinforced Concrete Beams,"
Journal of ACI, Vol. 64, No. 3, pp. 152-163.
Hemmaty, Y. (1998), Modelling of the Shear Force Transferred Between Cracks in Reinforced
Page 43 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Lin, C.S. and Scordelis, A.C. (1975). "Nonlinear Analysis of RC Shells of General Form". Journal
of Structural Division, ASCE, Vol. 101, No. ST3, pp. 523-538.
Menetrey P. (1998), “Relationships between flexural and punching failure.” ACI Structure Journal,
1998; 95(4), pp. 412-419.
Rashid, M (2004), “ The Behavior of Slabs Reinforced with GFRPs”, master thesis, preparing,
Faculty of Engineering, Memorial University of Newfoundland, St. John’s, Canada.
Page 44 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
/PREP7
!----------------------(1)MODEL GENERATION-----------------------------------
ANTYPE, STATIC
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,EX,3,,10000000
MPDATA,PRXY,3,,
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDE,EX,3
MPDE,EY,3
MPDE,EZ,3
MPDE,NUXY,3
MPDE,NUYZ,3
MPDE,NUXZ,3
MPDE,PRXY,3
MPDE,PRYZ,3
Page 45 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
MPDE,PRXZ,3
MPDE,GXY,3
MPDE,GYZ,3
MPDE,GXZ,3
MPDATA,EX,3,,1E+007
MPDATA,PRXY,3,,0
TBDATA,1,0.2,0.9,3.5,33
!the shear transfer coefficient of the open crack is 0.2
!the shear transfer coefficient of the close crack is 0.9
!the tensile strength of concrete is 3.5 MPa
!the compressive strength of concrete is 33 MPa
MPTEMP,,,,,,,,
MPTEMP,1,0
MPDATA,DENS,1,,2.3e-6 !Concrete density is 2.3e-16 kg/mm^3
Page 46 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
N,1005,960,300,-50,,,,
N,1006,960,360,-50,,,,
N,1007,960,480,-50,,,,
N,1008,960,600,-50,,,,
N,1009,960,720,-50,,,,
N,1010,960,840,-50,,,,
N,1011,960,960,-50,,,,
N,1012,0,960,-50,,,,
N,1013,60,960,-50,,,,
N,1014,120,960,-50,,,,
N,1015,180,960,-50,,,,
N,1016,240,960,-50,,,,
N,1017,300,960,-50,,,,
N,1018,360,960,-50,,,,
N,1019,480,960,-50,,,,
N,1020,600,960,-50,,,,
N,1021,720,960,-50,,,,
N,1022,840,960,-50,,,,
N,2000,960,0,-60,,,,
N,2001,960,60,-60,,,,
N,2002,960,120,-60,,,,
N,2003,960,180,-60,,,,
N,2004,960,240,-60,,,,
N,2005,960,300,-60,,,,
N,2006,960,360,-60,,,,
N,2007,960,480,-60,,,,
N,2008,960,600,-60,,,,
N,2009,960,720,-60,,,,
N,2010,960,840,-60,,,,
N,2011,960,960,-60,,,,
N,2012,0,960,-60,,,,
N,2013,60,960,-60,,,,
N,2014,120,960,-60,,,,
N,2015,180,960,-60,,,,
N,2016,240,960,-60,,,,
N,2017,300,960,-60,,,,
N,2018,360,960,-60,,,,
N,2019,480,960,-60,,,,
N,2020,600,960,-60,,,,
N,2021,720,960,-60,,,,
N,2022,840,960,-60,,,,
Page 47 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
Page 48 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
FITEM,5,-27
FITEM,5,29
FITEM,5,-31
FITEM,5,34
FITEM,5,38
FITEM,5,42
FITEM,5,46
FITEM,5,49
FITEM,5,53
FITEM,5,-55
FITEM,5,57
FITEM,5,-59
FITEM,5,62
FITEM,5,66
FITEM,5,70
FITEM,5,74
FITEM,5,77
FITEM,5,80
FITEM,5,84
FITEM,5,-86
FITEM,5,91
FITEM,5,93
FITEM,5,97
FITEM,5,-98
FITEM,5,103
FITEM,5,-104
FITEM,5,109
FITEM,5,-110
FITEM,5,115
FITEM,5,-116
FITEM,5,121
FITEM,5,-122
FITEM,5,127
FITEM,5,-128
FITEM,5,133
FITEM,5,-134
FITEM,5,139
FITEM,5,-140
FITEM,5,145
FITEM,5,-146
FITEM,5,151
Page 49 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
FITEM,5,-152
FITEM,5,157
FITEM,5,-158
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1,60, , , , , , ,1
!*
FLST,2,144,4,ORDE,2
FITEM,2,19
FITEM,2,-162
LMESH,P51X
wpoff,0,0,-50 !Define the volumes for concrete slab and column stub
BLC4,0,0,360,360,200
BLC4,360,0,600,360,200
BLC4,0,360,360,600,200
BLC4,360,360,600,600,200
wpoff,0,0,200
BLC4,0,0,120,120,600
FLST,2,4,6,ORDE,2
FITEM,2,1
FITEM,2,-4
VGLUE,P51X
WPSTYLE,,,,,,,,0
nummrg,all
numcmp,all
FITEM,5,5
FITEM,5,8
Page 50 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
FITEM,5,170
FITEM,5,173
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1,60, , , , , , ,1
!*
FLST,5,1,4,ORDE,1
FITEM,5,159
CM,_Y,LINE
LSEL, , , ,P51X
CM,_Y1,LINE
CMSEL,,_Y
!*
LESIZE,_Y1, , ,4, , , , ,1
!*
MSHAPE,0,3D
MSHKEY,1
!*
FLST,5,5,6,ORDE,2
FITEM,5,1
FITEM,5,-5
CM,_Y,VOLU
VSEL, , , ,P51X
CM,_Y1,VOLU
CHKMSH,'VOLU'
CMSEL,S,_Y
!*
VMESH,_Y1
!*
CMDELE,_Y
CMDELE,_Y1
CMDELE,_Y2
!*
nummrg,all
numcmp,all
Page 51 of 52
FE application to slab-column connections reinforced with GFRPs By Qi Zhang
!------------------------------(2)SOLUTION--------------------------------
/solu
solve
finish
Page 52 of 52