Number 3 Volume 8 Journal of Engineering

NONLINEAR FINITE ELEMENT ANALYSIS

OF COMPOSITE CONCRETE BEAMS
Prof. Dr. Khalid S. Mahmoud Dr. Mohannad H. Al-Sherrawi
Dept. of Civil Engineering,
University of Baghdad, Iraq

ABSTRACT
To study the nonlinear response of composite concrete beams, a finite element analysis is presented.
Material nonlinearities as a result of nonlinear response of concrete in compression, crushing and
cracking of concrete, strain softening and stiffening after cracking, yielding of reinforcement, bond-
slip, shear-slip, and dowel action between the precast concrete beams and the cast-in-situ slabs are
considered.
A biaxial concrete model is adopted. Concrete is treated as an orthotropic material with smeared
rotating crack model. The steel reinforcement is assumed to be in a uniaxial stress state and is
modeled as a bilinear material.
A two-dimensional plane stress finite element type is used to model the concrete. Reinforcement is
represented by one-dimensional bar elements. Bond-slip and dowel action is modeled by using
fictitious linkage elements with two springs at right angles. Shear-slip is modeled by using shear
transfer interface elements with appropriate stiffness values.
Comparison between the results obtained by the finite element and available experimental results of
composite concrete beams is made. The results compare satisfactorily with the experimental ones.

‫اﻟﺨﻼﺻﺔ‬
‫ﻟﻐﺮض ﻣﻌﺮﻓﺔ ﻣﻘﺪار ﺗﺤﻤﻞ اﻷﻋﺘﺎب اﻟﺨﺮﺳﺎﻧﻴﺔ اﻟﻤﺮآﺒﺔ و اﺳﺘﺠﺎﺑﺘﻬﺎ اﻟﻼﺧﻄﻴﺔ ﻟﻸﺣﻤﺎل اﻟﺴﺘﺎﺗﻴﻜﻴﺔ اﻟﻤﺆﺛﺮة‬
‫ ﻋﺪة ﻋﻼﻗﺎت ﻻﺧﻄﻴﺔ‬.‫ ﺗﻢ ﺗﺤﻠﻴﻠﻬﺎ ﺑﻮاﺳﻄﺔ اﻟﺤﺎﺳﺒﺔ اﻹﻟﻜﺘﺮوﻧﻴﺔ و ﺑﺎﺳﺘﻌﻤﺎل ﻃﺮﻳﻘﺔ اﻟﻌﻨﺎﺻﺮ اﻟﻤﺤﺪدة‬، ‫ﻋﻠﻴﻬﺎ‬
‫ اﻟﻌﻼﻗﺔ ﺑﻴﻦ ﻗﻮة‬، ‫ آﺬﻟﻚ ﻟﻤﺎدة اﻟﺤﺪﻳﺪ‬، ‫ آﺎﻟﻌﻼﻗﺔ اﻟﻼﺧﻄﻴﺔ ﺑﻴﻦ اﻹﺟﻬﺎد و اﻻﻧﻔﻌﺎل ﻟﻤﺎدة اﻟﺨﺮﺳﺎﻧﺔ‬، ‫اﻋﺘﻤﺪت‬
‫ اﻟﻌﻼﻗﺔ ﺑﻴﻦ ﻗﻮى اﻟﻘﺺ و اﻻﻧﺰﻻق‬، ‫اﻟﺮﺑﻂ و اﻻﻧﺰﻻق ﺑﻴﻦ ﺣﺪﻳﺪ اﻟﺘﺴﻠﻴﺢ و ﻣﺎدة اﻟﺨﺮﺳﺎﻧﺔ اﻟﻤﺤﻴﻄﺔ ﺑﻪ‬
‫ و آﺬﻟﻚ اﻟﺘﺄﺛﻴﺮ‬، ‫اﻟﺤﺎﺻﻞ ﻋﻨﺪ ﺳﻄﺢ اﻟﺘﻤﺎس ﺑﻴﻦ اﻟﻌﺘﺒﺔ اﻟﻤﺴﺒﻘﺔ اﻟﺼﺐ و ﺧﺮﺳﺎﻧﺔ اﻷرﺿﻴﺔ اﻟﻤﺼﺒﻮﺑﺔ ﻣﻮﻗﻌﻴًﺎ‬
.‫اﻟﻮﺗﺪي ﻟﺮاﺑﻄﺎت اﻟﻘﺺ‬
‫ و ﺗﻢ ﺗﻤﺜﻴﻞ‬.‫ﺑﺎﻟﻨﺴﺒﺔ ﻟﻠﺨﺮﺳﺎﻧﺔ ﺗﻢ اﻋﺘﻤﺎد ﻧﻤﻮذج اﻟﻌﻼﻗﺎت اﻟﺘﻜﻮﻳﻨﻴﺔ ﻟﻠﻤﺎدة اﻟﻤﺨﺘﻠﻔﺔ ﺑﺎﻻﺗﺠﺎهﻴـﻦ اﻟﻤﺘﻌﺎﻣﺪﻳـﻦ‬
‫ ﺗﻢ اﻋﺘﺒﺎر ﺣﺪﻳﺪ اﻟﺘﺴﻠﻴﺢ ذو ﻣﻮاﺻﻔﺎت أﺣﺎدﻳﺔ اﻻﺗﺠﺎﻩ‬.‫اﻟﺘﺸﻘﻖ ﻓﻲ اﻟﺨﺮﺳﺎﻧﺔ ﺑﻄﺮﻳﻘﺔ اﻟﺸﻘﻮق اﻟﻤﻮزﻋﺔ اﻟﺪوارة‬
.‫و ﺑﺎﺗﺠﺎﻩ ﻗﻀﺒﺎن اﻟﺘﺴﻠﻴﺢ‬
.‫ﻟﺘﻤﺜﻴﻞ اﻟﺨﺮﺳﺎﻧﺔ ﺗﻢ اﺳﺘﻌﻤﺎل ﻋﻨﺼﺮ اﻹﺟﻬﺎد اﻟﻤﺴﺘﻮي ﻋﻨﺪ ﺗﺤﻠﻴﻞ اﻟﻌﺘﺒﺔ اﻟﻤﺮآﺒﺔ ﺑﻄﺮﻳﻘﺔ اﻟﻌﻨﺎﺻﺮ اﻟﻤﺤﺪدة‬
‫ و ﻟﺘﻤﺜﻴﻞ اﻟﺮﺑﻂ واﻻﻧﺰﻻق و اﻟﺘﺎﺛﻴﺮ اﻟﻮﺗﺪي ﺗﻢ اﺳﺘﻌﻤﺎل ﻋﻨﺼﺮ‬.‫ﺣﺪﻳﺪ اﻟﺘﺴﻠﻴﺢ ﻣُﺜﻞ ﺑﻌﻨﺼﺮ أﺣﺎدي اﻻﺗﺠﺎﻩ‬
.‫ ﻗﻮى اﻟﻘﺺ و اﻻﻧﺰﻻق ﻣُﺜﻠﺖ ﺑﻌﻨﺼﺮ اﻟﺴﻄﺢ اﻟﺒﻴﻨﻲ‬.‫وﺻﻠﻲ ذو ﻧﺎﺑﻀﻴﻦ ﻣﺘﻌﺎﻣﺪﻳﻦ‬
273
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

‫ﺗﻢ ﻋﻤﻞ ﻣﻘﺎرﻧﺔ ﺑﻴﻦ اﻟﻨﺘﺎﺋﺞ اﻟﻤﺴﺘﺤﺼﻠﺔ ﻓﻲ هﺬﻩ اﻟﺪراﺳﺔ ﻣﻊ ﻣﺎ هﻮ ﻣﺘﻮﻓﺮ ﻣﻦ ﻧﺘﺎﺋﺞ ﻟﻠﻔﺤﻮص اﻟﻤﺨﺘﺒﺮﻳﺔ‬
.‫ اﻟﻨﺘﺎﺋﺞ اﻟﻤﺴﺘﺤﺼﻠﺔ آﺎﻧﺖ ﻣﻘﺎرﺑﺔ ﻟﻠﻨﺘﺎﺋﺞ اﻟﻤﺨﺘﺒﺮﻳﺔ‬.‫ﻻﻋﺘﺎب ﺧﺮﺳﺎﻧﻴﺔ ﻣﺮآﺒﺔ‬

KEYWORDS
composite beams, nonlinear analysis, orthotropic material, finite element, shear-slip.

INTRODUCTION
The composite concrete beam or girder usually consists of a precast or cast-in-situ reinforced or
prestressed concrete beam and a cast-in-place reinforced concrete slab tied together to act as a unit.
The shear connectors between the beam and the slab should tie the components together well
enough so that they act as a monolithic T-beam. Typical section of building and bridge composite
reinforced concrete flexural beam is shown in Fig. (1).

cast-in-place slab be

t
h
precastbeam
precast beam

bw

Fig. (1) Typical composite reinforced concrete beam.

MECHANICAL PROPERTIES
To model a reinforced concrete specimen, it is not only necessary to model the concrete and the
steel correctly but their interaction as well. This involves the tension stiffening effect in concrete
due to bond slip between the concrete and steel and the effect of dowel action and shear due to the
reinforcement.

Concrete
To model nonlinear concrete response, the constitutive relation contained in the modified
compressive field theory (Vecchio and Collins, 1986), Fig. (2) has been adopted. Thus, for concrete
in compression, the relation is:
⎡ ⎛ ε ⎞ ⎛ ε ⎞2 ⎤
σ c = f c ⎢2⎜⎜ c ⎟⎟ − ⎜⎜ c ⎟⎟ ⎥
'
(1)
ε ε
⎢⎣ ⎝ o ⎠ ⎝ o ⎠ ⎥⎦
where σc and ε c are the average principal compressive stress and strain in concrete, respectively;

ε o is the strain in concrete cylinder at peak stress f c' .

For concrete in tension, prior to cracking, a linear relation is used:
274
 Number 3 Volume 8 Journal of Engineering

σt = Ec ⋅ εt 0 ≤ ε t ≤ ε cr (2)
where
2f c' f cr
Ec = ε cr = f cr = 0.33 f c' (in MPa)
εo Ec

and σt and ε t are the average principal tensile stress and strain in concrete, respectively; Ec is the
modulus of elasticity of concrete ( initial tangent stiffness); fcr and ε cr are concrete cracking stress
and strain, respectively.
After cracking, concrete in tension is made to reflect tension softening and tension stiffening effects
together through the following relation:
f cr
σt = (3)
1 + 200ε t

ζ σ : stress softening coefficient of concrete. εi

ζ ε : strain softening coefficient of concrete. fcr= feq
εo ζ εε o
σi
ε cr
(compression) (tension)

cracked ζ σ fc

uncracked
fc’

Fig. (2) Stress-strain relation adopted.

Reinforcing steel
The bilinear representation of the stress-strain relation of the steel reinforcement is found adequate
for the proper simulation of the actual behavior since the elastic-plastic behavior with or without the
introduction of the strain-hardening region is easily simulated by controlling the slope of the second
line.

Friction slip and separation

In current design practice, the surface between two concrete parts cast at different ages is a plane of
weakness in an otherwise monolithic beam. In addition to tensile strength, this surface is usually
characterized by the Coulomb friction parameters, namely cohesion and a coefficient of friction. In
the ACI Code “Building Code Requirements for Structural Concrete (ACI 318M-95)”, for a crack
being assumed along an interface, tensile strength and cohesion are considered to be zero while the
coefficient of friction ( μ ) is assumed dependent on the concrete placement. For example, μ =1.4
for concrete placed monolithically, μ =1.0 for concrete placed against hardened concrete with
275
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

surface intentionally roughened, μ =0.6 for concrete placed against hardened concrete not
intentionally roughened.
For normal stresses ( σn ) ranging up to 2 MPa, Fronteddu et al. (1998) utilized their experimental
results from displacement controlled shear tests on concrete lift joint specimens with different
surface preparations, to propose an empirical interface constitutive model based on the concept of
basic friction coefficient ( μ b ) and roughness friction coefficient ( μ i ):

λ dμ b + χiμ i
μ= (4)
1 − λ d χiμ bμ i

where μ b = 0.950-0.220 σn for σ n ≤ 0.5MPa

μ b = 0.865-0.050 σn for 0.5 ≤ σ n ≤ 2.0MPa
λ d is the dynamic reduction factor equal to 1.00 for static loading and 0.85 for dynamic loading,
and χ i is the interface roughness factor equal to 1.00 for cracked homogeneous concrete, 0.80 for
water-blasted joints, 0.15 for untreated joints, and 0.00 for flat independent concrete surfaces.
Based on the experimental results presented by Fronteddu et al. (1998), a bilinear relationship
between shearing stress and slip is adopted, Fig. (3).
Shearing stress (MPa)

0.02 Slip (mm)

Fig. (3) Adopted shearing stress-slip relationship.

Dowel action
Shear force can be transmitted across a crack in reinforced concrete by the reinforcement crossing
the crack. If the reinforcement is normal to the plane of cracking, dowel action (shearing and
flexure of the bars) will contribute to the over-all shear stiffness. With oblique reinforcement there
is also a contribution from the tangential component of the axial force in the reinforcement.
The efficiency of dowel action in shear transfer in cracked reinforced concrete structures depends
strongly on the confinement exerted by the surrounding concrete, as well as on the hoop action of
the stirrups.
It has been suggested (Paulay et al., 1974) that there are three mechanisms of shear transfer through
dowel action in cracked reinforced concrete, i.e. direct shear, kinking and flexure of the bars. If the
concrete supporting each bar were considered rigid, the first two mechanisms would predominate.
However, it has been recognized (Mills, 1975) that significant deformation of the concrete does
occur, so that flexure of the dowel bar within the concrete is a principal action.
The experimental data presented by Poli et al. (1993) will be utilized to propose the following
simplified mathematical model for the secant stiffness of dowel action against core (kd):
276
 Number 3 Volume 8 Journal of Engineering

20 + 5(d b −14)
kd = Δ ≤ 1.5 mm
1.5
20 + 5(d b − 14)
kd = Δ ≥ 1.5 mm .(5)
Δ
where db is the diameter of the bar; and Δ is the dowel displacement. These equations are shown in
Fig. (4).
100
Experimental
Proposed
80
Shear force ( kN )

db=24
60

40 db=18

20 db=14

0
0 1 2 3 4 5 6
Dowel displacement ( mm )
Fig. (4) Proposed dowel stiffness.

Bond stress-slip relation

Based on the experimental results presented by Darwin and Graham (1993), a bilinear relationship
between bond stress and bond slip is adopted, Fig. (5). The mathematical model for the secant
stiffness of bond-slip is:

k s = 25 Δ s ≤ 0.2mm
5
ks = Δ s ≥ 0.2mm (6)
Δs

where ks is the secant stiffness for bond-slip; and Δ s is the slip.

6

5
Bond stress (MPa)

4
ks (constant)
3

2 ks (varying)
1

0
0.0 0.2 0.4 0.6 0.8 1.0
Bond slip (mm)

Fig. (5) Adopted bond stress-bond slip relationship.

277
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

FINITE ELEMENT REPRESENTATION

For simplicity it is preferable to approximate the original three-dimensional problem to an
equivalent two-dimensional one when possible.
A finite element with two-dimensional plane stress element type is used to model the concrete
Fig. (6). The rotating smeared crack model is used for modeling cracking. Reinforcement is
represented by one-dimensional bar elements. Bond-slip and dowel action are modeled by using
fictitious linkage elements, Fig. (7). Shear-slip is modeled by using shear transfer interface elements
with appropriate stiffness values, Fig. (8).

Y y

y
2
1

x θ

θ
X x
(a) (b)
Fig. (6) Four-noded quadrilateral element (a) element in global
coordinates, (b) cracked element.

Y
y’ x’
y
l x
j θ

j
k
i
θ i
X

Fig. (7) Interface element. Fig. (8) Linkage element.

To achieve the requirements of the present work, a computer program (MHND) has been written
based on the orthotropic concrete constitutive models (Al-Sherrawi, 2000). It is coded in Microsoft
Visual Basic 6.0. The equation solver is based upon symmetric banded Gaussian Elimination.
The nonlinear material properties considered in the present analysis are:
1- Nonlinear stress-strain relationship of concrete.
2- Cracking of concrete.
3- Yielding of reinforcement.
4- Bond slip.
5- Post-cracking shear transfer by aggregate interlock and dowel action.
6- Friction slip.
7- Separation.

278
 Number 3 Volume 8 Journal of Engineering

An incremental-iterative technique and the secant stiffness approach were utilized as a solution
algorithm. The local stiffness matrices are updated on the second iteration of every increment,
rather than the first. This is very helpful since the local stiffness matrices are calculated only after
load increment has been applied, i.e. the global stiffness matrix of the structure is evaluated on the
basis of the actual stress present at that increment, rather than on the basis of the previously reached
stress state.

ANALYSIS OF COMPOSITE CONCRETE BEAMS

Several experimental investigations of composite concrete beams are available in literature.
Although some of the available tests have concentrated only on specific aspects of the structural
problem and their experimental data are usually incomplete. In this paper, the computer program
(MHND) is used to analyze several composite concrete beams, which were tested by others, and
comparisons between the experimental and the finite element results are given.

Saemann and Washa tests

The evaluation of the strength of the joint between precast concrete beams and cast-in-place
concrete slabs has been the subject of Saemann and Washa (1964) tests.
Fig. (9) shows the beam cross section used with nominal dimensions indicated. The reduced breadth
of web produced high horizontal shearing stress in the bonded joint between the web and slab at
loads well below flexural failure. The beams were designed to have the joint 50.8 mm (2 in.) above
the neutral axis.
Fig. (10) shows the arrangement of stirrups ( φ 12 bars) in the beam having the maximum
percentage of stirrup steel across the joint. As the nominal steel percentage decreased from 1.02 to
0.51 and from 0.51 to 0.20, the reduction in the percentage of stirrup steel across the joint in each
case was accomplished by cutting off half of the stirrups crossing the joint at a level 25.4 mm (1 in.)
below the joint. The number of stirrups crossing the joint for a nominal steel percentage of 0.11 was
the same as for 0.20 but φ 10 (#3) bars were used instead of φ 12 (#4) bars. As the nominal steel
percentage decreased from 0.11 to 0.06 half of the φ 10 bars were cut off 25.4 mm (1 in.) below the
joint. In summary, the information on the steel bars crossing the joint is given in Table (1).
Percentage of steel was calculated by dividing the area of all stirrups crossing the joint by the total
joint area.
381
101.6

Neutral axis Ø 12 bar

88.9
Cage stirrup
431.8
152.4
50.
50.

152.4 Ø 25 bars

Fig. (9) Beam cross section (Saemann and Washa, 1964).

279
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

P 304.8 P
2 2

10 stirrups at 127 mm 355.6

plus 9 cage stirrups
2438.4 mm 203.2

Fig. (10) Arrangement of stirrups ( φ 12 bars) in beams having the maximum

percentage of stirrup steel across the joint (Saemann and Washa, 1964)

Table (1) Stirrup data.

Beam Nominal
percentage of Bar size Bar spacing
No. Series stirrup steel mm mm (in.)
across the joint

10 A 1.02 φ 12 127 (5)

5 C 0.51 φ 12 254 (10)
5 D 0.20 φ 12 508 (20)
12 C 0.11 φ 10 508 (20)
14 C 0.06 φ 10 1016 (40)
16 C 0.00 - -

Properties of the reinforcing steel are given in Table (2). Cage stirrups were added to reduce
shearing stresses below the joint. All steel members of the reinforcement were welded together.

Table (2) Properties of reinforcing steel.

Ultimate tensile
Size Yield point
strength
mm MPa
MPa
φ 10 370 560
φ 12 294 404
φ 25 252 401

Seven days after the webs were cast the slabs were cast. All beams were tested 28 days after the
slabs were cast (or 35 days after the webs were cast). Beams were supported and loaded as shown in
Fig. (10).

280
 Number 3 Volume 8 Journal of Engineering

The measured central deflections with respect to the applied load are shown in Fig. (11) for beams
of 2.438 m (8 ft) clear span, and with different percentage of steel across the joint. The curves show,
in general, that for an intermediate surface roughness the maximum load increased from about 232
kN to 356 kN as the percent steel across the joint increased from zero to 1.02 percent. However, the
beams behavior (central deflections) were almost identical when the load is below 200 kN. This
means that the reinforcement crossing the joint (stirrups) has no observable effect on the composite
beam behavior as the bond between the two concretes is not broken, and its effectiveness begins
after a crack forms in the joint (dowel action).
400

350
1.02%
0.20% 0.51%
300 0.11%
0.06%
250
Load (kN)

0%
200

150

100

50

0
0 2 4 6 8 10 12 14 16 18
Deflection (mm)

Fig. (11) Measured central deflection with respect to the applied load
for beams with different percentage of steel across the
joint.
Finite element idealization:
Due to symmetry, both in geometry and loading in the beams tested by Saemann and Washa (1964),
only halves of the beams are idealized by restraining the horizontal movement of the nodes at the
beam centerline.
Fig. (12) illustrates the finite element mesh for 2.438 m (8 ft) beam having the maximum
percentage of stirrup steel across the joint. Concrete is idealized by using 198 four-noded plane
stress elements. Reinforcement is idealized by 178 truss bar elements. 143 bond-slip linkage
elements of zero length are used to represent the bond-slip between concrete and steel. The joint
between the web and slab is idealized by 23 shear-friction interface elements and 10 dowel linkage
elements. The total number of nodes resulting from the above idealization is 396 nodes.
An incremental load is applied at node 310. The number of increments depends on the ultimate
strength of the beam, and ranges between 40 and 50 increments.

Discussion of results:
Fig. (13) shows the experimental and the finite element results of six beams having different
percentages of steel across the joint. The same trend of behavior is seen for the numerical and the

281
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

test results, but the finite element results show less value. This may be attributed to the selfweight
effects on the stresses and strains, which were neglected in the analysis.

P
2
A
L
C

5
4
4
3
3
2
2
1
1
A
i : Node no. Interface element at joint
Bar element
L
C
i : Element no.

Fig. (12) Finite element discretization of a beam tested by Saemann

and Washa.

400
Experimental
350 Present study

300

250
Load (kN)

200

150

100

50

1.02% 0.51% 0.20% 0.11% 0.06% 0.00%

0

1 2 3 4 Deflection (mm)
0 5 mm

Fig. (13) Comparison between experimental and present study results.

282
 Number 3 Volume 8 Journal of Engineering

As the shear caused slip to develop between the flange and the web, the beam began to act as a
partially composite member. This action is shown in Figs. (14) and (15) where the computed
normal strain distribution at Section A-A refer to Fig. (12) in beams 16C and 10 A, respectively, is
plotted for a series of increasing loads. It may be noted that from the beginning a discontinuity in
the strain distribution at the joint is apparent. This discontinuity is pronounced as the load increases
and it is apparent that two-beam action exists. The neutral axis of the composite beam is remains in
its position although the load increases. This behavior results from the tendency of the beam and the
slab to act separately.
Figs. (16) and (17) show the horizontal normal stresses (in X-direction) at section A-A, they show
gradual deviation from the linear behavior, especially where the stresses are high, indicating the
nonlinear nature of concrete behavior. The crack propagation is clear with the increase of the
applied load. The stress distribution shows that the composite concrete beam acts as a partially
composite beam from the first load increment. This is due to the relative movement between the
two concretes at the surface between them.
Prediction of cracking by the finite element method at the ultimate load for beam 16C (zero
percentage of steel) is shown in Fig. (18). The overall depth and pattern of the cracks indicate that
the failure is due to shear failure between the two concrete parts (at the joint). This result agrees
with the type of failure observed experimentally by Saemann and Washa (1964).
The cracking and failure mechanism of the joint were similar for all beams analyzed in this section.
The usual crack due to a combined state of bending and shear developed in the shear span of the
precast beam. These cracks started as vertical cracks and turned more and more towards the load
point as the loading progressed. In general, the cracks stopped when they reached the horizontal
joint. However, after additional loading some of the cracks continued horizontally below the contact
surface towards the load point. Examination of failure done by Saemann and Washa indicated that
in most cases some concrete from the web adhered to the flange.

450

400
Distance above bottom (mm)

350

300 Applied load

250 30 kN

200 50 kN

150 100 kN
150 kN
100
200 kN
50
234 kN
0
-0.0008 -0.0004 0.0000 0.0004 0.0008
Normal strain (mm/mm)
Fig. (14) Normal strain distribution at section A-A in beam with 0%
steel across the joint, showing two-beam action.

283
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

450

400
Distance above bottom (mm)
350

300 Applied load

250 30 kN
50 kN
200
100 kN
150
150 kN
100 200 kN

50 250 kN
305 kN
0
-0.0012 -0.0008 -0.0004 0.0000 0.0004 0.0008 0.0012
Normal strain (mm/mm)

Fig. (15) Normal strains distribution at section A-A in beam with

1.02% steel across the joint, showing two-beam action.

450

400
Distance above bottom (mm)

350

300 Applied load

250 30 kN
50 kN
200
100 kN
150
150 kN
100
200 kN
50 234 kN

0
-16 -14 -12 -10 -8 -6 -4 -2 0 2 4
Normal stress (MPa)

Fig. (16) Normal stresses distribution at section A-A in beam with 0%

steel across the joint, showing two-beam action.

284
 Number 3 Volume 8 Journal of Engineering

450
Distance above bottom (mm) 400

350
Applied load
300
30 kN
250 50 kN

200 100 kN
150 kN
150
200 kN
100
250 kN
50
305 kN
0
-20 -16 -12 -8 -4 0 4
Normal stress (MPa)
Fig. (17) Normal stresses distribution at section A-A in beam with
1.02% steel across the joint, showing two-beam action.

P
2

5
4
4
3
3
2
2
1
1
i : Node no. Interface element
i : Element no. Bar element

Fig. (18) Prediction of cracking by the finite element method at the ultimate load for
beam 16C (zero percentage of steel).

285
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

Shearing stresses, slips, normal stresses, and separations at the joint for beam 16C (zero steel
percent across the joint) and beam 10A (1.02% steel percentage across the joint) are shown in
Figs. (19) and (20), respectively, for a series of increasing loads. The curves show that the values
increased as the load increased, that maximum shearing stresses and slips were usually located
about 900 mm from the left end of the beam, and that separations (loss of bond) started at relatively
low load levels at the beam end. This agrees well with the results of Cook’s investigation (1977) of
bonded-aggregate composite beams.
Also, it can be noticed that the shearing stresses and the normal stresses between the two concretes
decreased with the increase in the steel percentage across the joint.
Load increments
0 0.00
Shearing stress (MPa)

-0.02
-1
-0.04

Slip (mm)
-2 -0.06

-3 -0.08
-0.10
-4
-0.12
-5 -0.14
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Distance from left end of the beam (mm) Distance from left end of the beam (mm)
1 (a) 0.01 (b)
Normal separation (mm)
Normal stress (MPa)

0 0.00

-1 -0.01

-2 -0.02

-3 -0.03

-4 -0.04
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Distance from left end of the beam (mm) Distance from left of the beam (mm)
(c) (d)
Fig. (19) Distribution of (a) shearing stresses, (b) slips, (c) normal stresses, and
(d) normal separations across the joint for beam 16C (zero steel percentage) under
concentrated load.
0 0.00
S h e a rin g s tre s s (M P a )

-0.02
-1
S lip (m m )

-0.04

-2 -0.06

-0.08
-3
-0.10

-4 -0.12
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Distance from left end of the beam (mm) Distance from left end of beam (mm)
1 (a) 0.01 (b)
N o rm a l s e pa ra tio n (m m )
N o rm a l stre s s (M P a )

0 0.00

-1 -0.01

-2 -0.02

-3 -0.03
0 200 400 600 800 1000 1200 1400 0 200 400 600 800 1000 1200 1400
Distance from left end of beam (mm) Distance from left of the beam (mm)
(c) (d)
Fig. (20) Distribution of (a) shearing stresses, (b) slips, (c) normal stresses, and (d)
normal separations across the joint for beam 10A (1.02% steel percentage) under
concentrated load.
286
 Number 3 Volume 8 Journal of Engineering

CONCLUSIONS
The following conclusions can be drawn from the present work:
1- A good estimate can be obtained from the analysis and ultimate load calculation of
composite concrete beams that might be achieved by the Program MHND which may
contribute in more practical designs.
2- The present algorithm for the analysis of composite concrete beams proves to predict beams
behavior satisfactorily and indicates good estimates of failure loads compared with
experimental values.
3- The performance of the linkage element and interface element, used in this study to model
dowel action and shear transfer between two concretes cast in different times, is quit good.
4- Shear connectors have no observable effect on the composite concrete beam behavior at the
early load stages, and their effectiveness begins after a considerable slip occurs between the
precast beam and the cast-in-place slab.
5- From the early load stages, as the shear causes slip to develop between the web and the
flange, the composite beam behaves as a partially composite member.

REFERENCES
ACI Committee 318, (1995). “Building Code Requirements for Structural Concrete (ACI 318-95).”
American Concrete Institute, Detroit,.

Al-Sherrawi, Mohannad H. (2000). “Shear and Moment Behavior of Composite Concrete Beams.”
Ph.D. Thesis, Dept. of Civil Eng., Univ. of Baghdad, 135.

Darwin, D., and Graham, E. K. (1993). “Effect of Deformation Height and Spacing on Bond
Strength of Reinforcing Bars.” ACI Struct. J., 90(6), 646-657.

Dei Poli, S., Di Prisco, M., and Gambarova, P. G. (1993). “Cover and Stirrup Effects on the Shear
Response of Dowel Bar Embedded in Concrete.” ACI Struct. J., 90(4), 441-450.

Fronteddu, L., Léger, P., and Tinawi, R. (1998). “Static and Dynamic Behavior of Concrete Lift
Joint Interfaces.” J. Struct. Engrg., ASCE, 124(12), 1418-1430.

Mills, G. M. (1975). “A Partial Kinking Criterion for Reinforced Concrete Slabs.” Mag. Concrete
Res., 27(90), 13-22.

Pauley, T., Park, R., and Phillips, M. H. (1974). “Horizontal Construction Joints in Cast-in-Place
Reinforced Concrete.” Shear in Reinforced Concrete, Detroit, American Concrete Institute, ACI
Special Publication SP-42. 2, 599-616.

Saemann, J. C., and Washa, G. W. (1964). “Horizontal Shear Connections between Precast Beams
and Cast-in-Place Slabs.” ACI J., 61(11), 1383-1409.

Vecchio, F. J., and Collins, M. P. (1986). “The Modified Compression Field Theory for Reinforced
Concrete Elements Subjected to Shear.” ACI J., 83(2), 219-231.

NOTATIONS
db the diameter of the bar.
Ec the modulus of elasticity of concrete.
fcr the concrete cracking stress.
kd the secant stiffness of dowel action against core.
287
 NONLINEAR FINITE ELEMENT ANALYSIS
K. S. Mahmoud and M. H. Al-Sherrawi
OF COMPOSITE CONCRETE BEAMS

ks the secant stiffness of bond-slip.

Δ the dowel displacement.
Δs the slip between steel and concrete.
χi the interface roughness factor.
εc the average principal compressive strain in concrete.
ε cr the concrete cracking strain.
εo the strain in concrete cylinder at peak stress of concrete.
εt the average principal tensile strain in concrete.
λd the dynamic reduction factor.
μ the coefficient of friction.
μb the basic friction coefficient.
μi the roughness friction coefficient.
σc the average principal compressive stress in concrete.
σn the normal stresses.
σt the average principal tensile stress in concrete.
ζε the strain softening coefficient of concrete.
ζσ the stress softening coefficient of concrete.

288