SHEAR AND MOMENT BEHAVIOR OF

COMPOSITE CONCRETE BEAMS

A THESIS
SUBMITTED TO THE COLLEGE OF ENGINEERING
OF THE UNIVERSITY OF BAGHDAD IN
PARTIAL FULFILLMENT OF THE
REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF
PHILOSOPHY IN CIVIL
ENGINEERING

MOHANNAD HUSAIN MOHSIN AL-SHERRAWI

(M.Sc.)

SHAWAL 1421 DECEMBER 2000

I certify that the preparation of this thesis was made under my supervision
at the University of Baghdad in partial fulfillment of the requirements for
the degree of Doctor of Philosophy in Civil Engineering.

Signature:
Name: Prof. Dr. Khalid S. Mahmoud
Date: / 12 / 2000.
 ACKNOWLEDGEMENT

I hereby wish to express my sincere gratitude to my supervisor Professor

Dr. Khalid S. Mahmoud for his continued encouragement and help
throughout the preparation of this work.

Special thanks are due to staff members of college of engineering

library for their cooperation and help.

i
 ABSTRACT

To study the nonlinear response of composite concrete beams, a finite

element analysis is presented in this work. 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.
The validity of the proposed modeling and the capabilities of the
computer program written are examined by analyzing several published
experimental reinforced concrete specimens.
Comparison between the results obtained by the finite element
computer program and available experimental results of composite
concrete beams is made. The analytical results compare satisfactorily
with the experimental ones.
A parametric study deals with shear and bending moment capacity
of composite concrete beams is presented.

ii
 NOTATIONS

[A] : the displacement transformation matrix.

[B] : strain-displacement matrix.
[D] : stress-strain matrix.
[Ds]G : the reinforcement stiffness matrix in global coordinate.
[Ds]L : the incremental constitutive matrix for the reinforcement in the
material coordinate x’ and y’.
[I] : identity matrix of order (2×2).
[K(σ,ε)]: stress-strain dependent global stiffness matrix.
[K] : element stiffness matrix.
[Kb]G: steel element stiffness matrix in global coordinates.
[T] : rotation transformation matrix between principal strain axes and
global axes.
[Ts] : the steel transformation matrix.
{F}: external load vector.
{U}: unknown displacement vector.
Ab : the area of the steel bar element.
Ac : area of contact surface.
Av : the area of shear reinforcement within a distance s.
Avf : reinforcement across the assumed crack.
bv : the width of cross section at contact surface being investigated for
horizontal shear.
d: the distance from extreme compression fiber to centroid of tension
reinforcement for entire composite section.
db : diameter of dowel bar.
E: the modulus of elasticity.
E1 and E2: secant moduli of elasticity in directions 1 and 2, respectively.
Eb : the secant modulus for the steel reinforcement.
Ec : concrete modulus of elasticity.
Eci: initial modulus of elasticity of concrete.
Ecn: secant modulus of elasticity of concrete measured at stress of 0.4 fc’.
Eo: open crack strain.
Es’: steel hardening modulus.
Esx’ and Esy’ the secant modulus for the reinforcements in x’ and y’ directions,
: respectively.
fb : bearing on loaded area.
fc : extreme fiber stress in compression.
fc’ : the 28-day specified compressive strength of the concrete under
consideration.
fcr: concrete cracking stress.
fcu: ultimate compressive stress in concrete.
feq: cracking strength.
fr : modulus of rupture.
fs : the tensile stress in the reinforcement.
fsx’ : the average stress in the reinforcement.
ft : extreme fiber stress in tension for plain concrete.
fy : the specified yield strength of nonprestressed reinforcement.

iii
G: secant shear modulus.
Gf: fracture energy exerted.
I: moment of inertia of the member.
i and j : specify the location of the integration points.
Id: desired number of iterations.
It-1: number of iterations required at last (t-1)th increment.
J: the Jacobian matrix relating the natural coordinate derivatives to the
local coordinate derivatives.
j: iteration counter.
kd : secant stiffness of dowel action against core.
Kh : local horizontal direction stiffness.
ks and kn : the shear and the normal stiffness coefficients, respectively.
Kv : local vertical direction stiffness.
Lb : the length of the steel bar element.
Mn : ultimate flexural strength.
Mu : factored moment at section.
n = Es/Ec modular ratio.
N: total number of nodal points in the problem.
Ni : the shape functions.
Nuc : the factored tensile force.
s: the spacing of shear reinforcement.
t: subscript denotes the current load step.
u: bond stress.
v: shear carried by concrete plus shear reinforcement.
vc : shear carried by concrete.
Vc and Vs: nominal shear strength provided by concrete and shear reinforcement,
respectively.
Vd : the total vertical shear at the point considered due to the design load.
vdh : design horizontal shear stress.
Vn : horizontal shear strength.
Vn: ultimate shear strength.
Vnh : nominal horizontal shear strength.
vu : average shearing stress.
Vu : factored shear force at section considered.
wi and wj : the corresponding weight factors (wi = wj = 2).
x’ and y’: material (local) coordinates.
Αs: reinforcement for direct tension.
ΔF: load increment.
Δ: dowel displacement.
Εs: steel modulus of elasticity.
αf : angle of reinforcement inclination.
εc: average principal compressive strain in concrete.
εcr: concrete cracking strain.
εcu: ultimate compressive strain in concrete.
εh and εv : relative displacements between points (i) and (j) in the local
horizontal and vertical directions.
εm: strain corresponding to the maximum stress.
εo : uniaxial strain corresponding to the ultimate strength.
εsu: ultimate steel strain.

iv
εsx’ : the average strain in the reinforcement.
ε t: average principal tensile strain in concrete.
ε1f and ε2f: equivalent uniaxial strains.
φ: strength reduction factor (0.85 for shear).
λ: the correction factor related to unit weight of concrete.
μ: coefficient of friction.
ν: Poisson’s ratio.
θ: the angle between global and local coordinate of the element.
θc: angle between global and local coordinates of the concrete element.
θs: angle between the x’-direction of the steel reinforcement and the local
x-axis of the concrete element.
ρ: ratio of reinforcement.
ρv : the ratio of tie reinforcement area to area of contact surface.
ρx and ρy : the reinforcement ratio in x’ and y’ directions, respectively.
σ1 : major principal stress.
σ2 : miner principal stress.
σc: average principal compressive stress in concrete.
σh and σv: relative linkage forces between point (i) and (j) in the local horizontal
and vertical directions.
σ t: average principal tensile stress in concrete.
σx : normal stress component.
σy : normal stress component.
τxy : shearing stress component.
ζσ and ζε: stress and strain softening coefficients of concrete, respectively.
{δ}: displacement vector.
{ε}: strain vector.
{σ}: stress vector.

v
 CONTENTS

ACNOWLEDGEMENT..............................................................................................i
ABSTRACT................................................................................................................ii
NOTATIONS.............................................................................................................iii
CONTENTS...............................................................................................................vi

CHAPTER ONE
INTRODUCTION
1.1 General..................................................................................................................1
1.1.1 Composite concrete beam...............................................................................1
1.1.2 The shear connector........................................................................................2
1.1.3 Partial composite action..................................................................................3
1.2 Advantages and disadvantages.............................................................................5
1.2.1 Advantages.....................................................................................................5
1.2.2 Disadvantages.................................................................................................6
1.3 Objective and scope..............................................................................................6

CHAPTER TWO
REVIEW OF LITERATURE
2.1 Introduction..........................................................................................................8
2.2 Composite concrete beam tests............................................................................8
2.3 Shear transfer tests...............................................................................................9
2.3.1 Characteristics of the shear plane................................................................14
2.3.2 Characteristics of the reinforcement............................................................14
2.3.4 Concrete strength.........................................................................................17
2.3.5 Direct stress parallel to the shear plane........................................................17
2.3.6 Direct stress transverse to the shear plane....................................................18
2.3.7 Cyclic shear transfer.....................................................................................19
2.3.8 Slant shear test..............................................................................................20
2.3.9 Strut-and-tie model.......................................................................................21
2.3.10 Dynamic shear transfer...............................................................................21

CHAPTER THREE
RECOMMENDATIONS FOR CONSTRUCTION AND DESIGN
3.1 Introduction........................................................................................................23
3.2 Design of composite concrete beam...................................................................23
3.2.1 Vertical shear strength..................................................................................24
3.2.2 Horizontal shear strength..............................................................................25
3.2.3 Ties for horizontal shear...............................................................................26

vi
 3.2.4 Design of slab...............................................................................................26
3.2.5 Effective flange width..................................................................................26
3.2.6 Deflections...................................................................................................27
3.2.7 Uplifts..........................................................................................................28
3.2.8 Continuity....................................................................................................29
3.2.9 Allowable stresses, moduli of elasticity, and load factors...........................30
3.2.10 Differential shrinkage and creep................................................................30
3.3 Shear friction......................................................................................................31
3.3.1 General considerations.................................................................................31
3.3.2 Shear-friction concept .................................................................................31
3.3.3 Shear-transfer design methods.....................................................................32
3.3.4 Shear-friction design method.......................................................................32
3.3.5 Coefficient of friction...................................................................................33
3.3.6 Maximum shear-transfer strength.................................................................35
3.3.7 Normal forces...............................................................................................35
3.3.8 Additional requirements...............................................................................36
3.4 Analysis of composite section............................................................................36
3.5 Construction methods.........................................................................................37
3.5.1 Intentional roughening..................................................................................37
3.5.2 Shoring and unshoring..................................................................................38

CHAPTER FOUR
MECHANICAL PROPERTIES AND BEHAVIOR OF MATERIALS
4.1 General...............................................................................................................41
4.2 Concrete in uniaxial state...................................................................................41
4.2.1 Concrete stress-strain behavior in uniaxial compression.............................41
4.2.1.1 General characteristics...........................................................................41
4.2.1.2 Review of analytical relations................................................................43
4.2.2 Concrete stress-strain behavior in uniaxial tension......................................43
4.2.2.1 General...................................................................................................43
4.2.2.2 Review of experimental work and analytical relations..........................44
4.2.3 Present model...............................................................................................46
4.3 Concrete in biaxial state.....................................................................................48
4.3.1 Poisson’s ratio..............................................................................................49
4.4 Stress-strain behavior of reinforcing steel..........................................................50
4.4.1 Review of analytical relations......................................................................52
4.4.2 Estimation of reinforcing steel properties....................................................52
4.4.3 Present model...............................................................................................54
4.5 Tension related phenomena in reinforced concrete............................................54

vii
 4.5.1 Tension stiffening.........................................................................................54
4.5.1.1 Definition................................................................................................54
4.5.1.2 Modeling.................................................................................................56
4.5.2 Bond and bond slip........................................................................................57
4.5.2.1 Definition................................................................................................57
4.5.2.2 Bond stress-slip relation.........................................................................58
4.6 Shear transfer across the crack...........................................................................61
4.6.1 Aggregate interlock......................................................................................61
4.6.2 Dowel action.................................................................................................62
4.6.3 Concrete strength parallel to crack...............................................................65
4.7 Friction slip and separation.................................................................................65

CHAPTER FIVE
FINITE ELEMENT IDEALIZATION
5.1 Introduction........................................................................................................68
5.2 Finite element idealization of reinforced concrete members.............................69
5.3 Representation of concrete.................................................................................70
5.3.1 Four-node isoparametric quadrilateral element............................................71
5.3.2 Modeling of cracked concrete......................................................................72
5.4 Modeling of the reinforcement...........................................................................73
5.4.1 Discrete reinforcement.................................................................................73
5.4.2 Embedded reinforcement.............................................................................74
5.4.3 Distributed reinforcement............................................................................75
5.4.4 Modeling of the reinforcement in the present study....................................75
5.5 Linkage element.................................................................................................77
5.6 Interface element................................................................................................79
5.7 Nonlinear analysis..............................................................................................81
5.8 Model description...............................................................................................82
5.8.1 Biaxial strength envelope.............................................................................83
5.8.1.1 Biaxial compression region....................................................................83
5.8.1.2 Compression-tension region...................................................................84
5.8.1.3 Biaxial tension region.............................................................................85
5.8.2 Material stiffness matrix...............................................................................87
5.8.2.1 Poisson’s ratio........................................................................................89
5.8.3 State determination.......................................................................................89
5.9 Solution algorithm...............................................................................................90
5.10 Load increment.................................................................................................92
5.11 Convergence criterion......................................................................................93

viii
 5.12 Verification of the program..............................................................................94
5.12.1 Shallow beam OA1.....................................................................................94
5.12.2 Tension specimen UC-15...........................................................................97
5.12.3 Shear transfer specimen CB2M..................................................................98

CHAPTER SIX
ANALYSIS OF COMPOSITE BEAMS
6.1 Introduction.......................................................................................................102
6.2 Revesz tests.......................................................................................................102
6.3 Saemann and Washa tests.................................................................................104
6.3.1 Finite element idealization..........................................................................107
6.3.2 Discussion of results...................................................................................109
6.4 Parametric study...............................................................................................117

CHAPTER SEVEN
CONCLUSIONS AND RECOMMENDATIONS
7.1 Conclusions......................................................................................................124
7.2 Recommendations............................................................................................125

REFERENSES.........................................................................................................126

ix
 CHAPTER ONE
INTRODUCTION

1.1 General:
Composite construction consists of using two materials together in one
structural unit and using each material to its best advantage. The number
of combinations is almost endless; steel and concrete, timber and
concrete, timber and steel, precast and cast-in-place concrete, etc.
Composite construction is generally economic when a floor or
bridge deck is desired to be in situ, as opposed to precast, and its total
depth is required to be less than for in situ reinforced concrete
construction, or when durability (absence of cracks at working loads) is
required (e.g. bridge decks).

1.1.1 Composite concrete beam:

The composite beam is actually a built-up member, using two different
materials and using each material to its best advantage. The usual
composite beam consists of three main components: the structural beam
(of whatever material), the plate or slab which is added, and some type of
connector to hold the beam and slab together.
The structural beam is usually a material, which carries tensile
stresses efficiently, and the slab is concrete, which has good compressive
strength.
The composite concrete beam or girder usually consists of a precast
or cast in situ reinforced or prestressed reinforced concrete beam and
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. (1.1). The principles involved
are, of course, the same for other types of sections, such as those shown
in Fig. (1.2). Primarily because of the better quality control of precast
units, it is usually economical to use higher concrete strengths for the
precast girders than for the cast-in-place decks (or toppings).

1
 cast-in-place slab be

h
precastbeam
precast beam

bw

Fig. (1.1) Typical composite reinforced concrete beam.

1.1.2 The shear connector:

In the composite beam, the purpose of the shear connector is to tie the
beam and slab together and force them to act as a unit. In order to
accomplish this, the shear connector must do two things. It must resist the
horizontal force that develops between the beam and slab, as the
composite member is loaded. Second, it should hold the slab down
against the beam and prevent uplift of the slab as the unit is loaded. If the
beam and slab are allowed to slip relative to each other as the beam is
loaded, there is little or no composite action.
In building works, where loads are primarily static, the natural bond
between the beams and slab furnished a valuable reserve of strength. In
bridge works, the moving loads and impact probably destroy the natural
bond after a very few load cycles.
Johnson (1975) discussed the behavior of flitched beams with no
shear connectors with full slip and shear connectors with no slip (i.e., full
interaction), as shown in Fig. (1.3). It may be shown that the maximum
bending stress is reduced by 50% by net, providing shear connection,
while the maximum shearing stress is unchanged. Also, the mid-span
deflection with fully effective shear connectors (no slip at the interface) is
25% of that in the case with no shear connectors (in this special case).
Thus the provision of shear connectors increases both the strength and
stiffness of a beam of given size, which leads to a reduction in the beam
size for a given loading.

2
 Cast-in-place
Cast-in-place

Precast
Precast

Cast-in-place Cast-in-place

Precast Precast

Cast-in-place Cast-in-place

Precast Precast

Cast-in-place Cast-in-place

Hollow
Precast box
Precast

Fig. (1.2) Various types of composite concrete sections.

1.1.3 Partial composite action:

In actual practice with concrete beam and slab construction, zero
composite action is impossible because there is always some degree of
bond and friction between the slab and the beam. Similary, 100%
composite action is impossible because there is always some small degree
of slip, no matter how rigidily the shear connection may be designed.
3
 b f v
w

x No interaction
L Full interaction

a. elevation b. section c. bending stress d. shearing stress

ds slip
dx

x x
δmax

L x
No interaction
Partial interaction

e. deflected shape f. slip strains g. slip

Fig. (1.3) Slip and partial interaction for two rectangular beams
(Johnson, 1975).

Fig. (1.4) shows the strain diagram in composite concrete beam with
different types of interaction. It can be seen that for the case of no
interaction, each of the beam and slab act separately (Fig. (1.4a)). While
the composite beam with full interaction behave as a single monolithic
beam (Fig. (1.4c)). The behavior of a partially composite beam is in
between (Fig. (1.4b)).
The evaluation of the strength of the joint between precast concrete
beams and cast-in-place concrete slabs has been the subject of
considerable research. However, the behavior of the joint during the
loading has not been well studied by previous researchers. When the joint
in a composite concrete girder is unable to transmit all internal forces
from one part of the section to the other part in the same manner as if the
entire section were structural concrete cast in one piece, the girder is only
partially composite with stiffness characteristics between those of a fully
composite and a two-piece girder.
Current design methods rely heavily on empirically derived
equations based on experimental results of ultimate strength tests of the
joint.

4
 (a) (b) (c)

Fig. (1.4) Strain diagrams in (a) noncomposite, (b) partially composite,

(c) composite beam.

1.2 Advantages and disadvantages:

1.2.1 Advantages:
Compared to in-situ concrete construction, the advantages of composite
construction may be expressed as follows:
1. Construction will be rapid.
2. Generally there will be a reduction in site costs as scaffolding,
shuttering, and other temporary supports will not be needed in such
quantities (if at all) as for in-situ concrete work.
3. There will be considerably less in-situ concrete work thus reducing the
amount of wet work on site, which in turn reduces the demand for
local site labor, and the import of raw materials.
4. Casting in the precast works is usually unimpeded by adverse weather
conditions.
5. Precast beams can be made by mass production methods and there
should be rapid re-use of molds. It should be said here that molds can
be made to a precession not possible on site and more intricate work
can be carried out. However, such molds are expensive to make and
there should be a considerable amount of repetition in the design to
enable economic use to be made of them.
6. Beams can be made to a good, even excellent, standard due to the use
of a trained and specialized labor force working under factory
conditions.

5
7. Precast beam can be inspected before it is erected and there is an
opportunity to reject any substandard work before incorporation in the
structure.
In addition to that, the principal advantage of composite construction
is that significantly stiffer and stronger beams can be obtained as
compared to the same members without composite action. Also when
using precast construction in seismic zones, the composite slab or topping
is normally required in order to serve as a diaphragm connecting the
varies units.

1.2.2 Disadvantages:
There are, of course, disadvantages, which may be summarized as
follows:
1. The joint between the precast beam and the cast-in-situ deck needs a
special design and a preparation before casting the deck.
2. Some additional reinforcement may be required for lifting and
transporting precast beams. It has to be appreciated that a precast
beam has to be designed not only to function as part of a total
structure but also for the load condition pertaining during lifting and
transport.
3. If the beams are large in size problems can arise concerning
transportation and lifting costs.

1.3 Objective and scope:

Different codes and specifications present rules for design of composite
concrete beam. While such design rules are adequate for many situations,
on occasions a more detailed method of analysis is required which takes
account of most the principal parameters.
The finite element method is one of the methods widely used to
explore the behavior of composite steel-concrete beam, but no attempt
has been done yet to analysis composite concrete-concrete beam by the
finite element method. This method has found increasing use in the
design office with the increasing efficiency of computers. However,
analysis of composite beam is complicated by a structural discontinuity
with slip and vertical separation between the two components (beam and
slab) on their common interface. Finite element representation of those
movements is usually achieved by using special elements.

6
 The present work is devoted to the study of the behavior of
reinforced concrete composite beams during short term loading using
finite element method to analyze them.

1.4 Thesis layout:

This work consists of seven chapters. Chapter one introduces the problem
to be solved in the present study. Chapter two reviewed the available tests
on composite concrete beams. Experimental works done by others on
shear transfer between two concretes are also reviewed. Chapter three
presents the recommendations for construction and design of composite
concrete beam. Chapter four is concerned with the review of the various
material models presented in the literature together with the models
adopted in this study to identify and represent material properties and
behavior for concrete, steel, bond-slip, dowel action, and shear friction
between two concretes, cast in different times, throughout the range of
load applications. The formulations of the element matrices are presented
in Chapter five. The nonlinear finite element formulation of the
reinforced concrete problem, and the numerical algorithm adopted in this
study for static nonlinear analysis is also presented with suitable
convergence criteria. In Chapter six, several examples of experimental
tests available in the literature are analyzed. A parametric study deals
with shear and bending moment capacity of a composite beam is
presented. Finally, a summary of the conclusions drawn from this study
and recommendations for further research are given in Chapter seven.

7
 CHAPTER TWO
REVIEW OF LITERATURE

2.1 Introduction:
In this chapter a review is made in the available previous experimental
works conducted on composite concrete beams in which a cast-in-place
concrete slab is attached to a precast concrete beam by means of shear
connectors. Experimental works done on shear transfer between two
concretes are also reviewed.

2.2 Composite concrete beam tests:

Several experimental investigations of composite concrete beams are
done in literature. Unfortunately, some of these investigations are not
available in Iraq. Tests were made as long ago as 1914 by Johonson and
Nichols Grossfield and Birnstiel, 1962).
Preliminary tests were made by Revesz (1953) at the Imperical
Collage of Science and Technology London University, London,
England, on five different composite T-beams of 4.267 m (14 ft) span to
destruction to observe the behavior of the beams under loads. No
provision had been made for stirrups, or serration at the top of the precast
beam. Age of concrete of cast-in-place flange and precast web at time of
test was 7 days and 29 days, respectively. The test load was applied at
third-points of the span. Deflections, strains and crack widths were
measured and recorded. Design and ultimate loads were examined in the
light of estimated values based on simplified assumptions. Observations
were drawn regarding warning of failure. A conclusion has been attained
that the variation in the quality of the cast-in-place concrete of T-beams
did not exert appreciable influence on the load capacity of composite
beams.
Tests carried out by Hanson (1960) on the problem of shear
connections between precast beams and cast-in-place slabs indicated that
the ultimate horizontal shear strength of a smooth bonded joint was about
2.069 MPa (300 psi) and that of a rough bonded joint was 3.449 MPa
(500 psi). In addition, it was found that the shear strength of a joint could
be increased approximately 1.207 MPa (175 psi) for each percent of
reinforcing steel crossing the joint. These values are substantially higher
than the tentative recommendations of ACI–ASCE Committee 333
(1960).

8
 A pilot test program of limited scope was undertaken by Grossfield
and Birnstiel (1962) to study the effect of three joint treatment methods
and the problems of instrumentation. Six composite beams and two
monolithic beams were tested. Two levels of horizontal shearing stress at
the joint were produced by varying the width of the contact surface
between the web and flange. The average age of the beam specimens at
testing was 5 months. Load was applied to the specimens at two points.
Vertical deflection, strains and slip measurements were made after each
increment of load until beam collapsed.
Saemann and Washa (1964) tested 42 T-beams. The tests, at the
University of Wisconsin, were designed as an attempt to provide
information on several variables:
1. Degree of roughness of contact surface.
2. Length of shear span.
3. Percentage of steel across the joint.
4. Effect of shear keys.
5. Position of the joint with respect to the neutral axis.
6. Concrete compressive strength.
All beams were tested 28 days after the slabs were cast, 35 days after
the webs were cast. Center deflections, strains, and slips along the joint
were measured. Results obtained indicated complex relations between
roughness of surface joint, percent steel across joint, and shear span.
An investigation of the strength of the joint, between a precast
concrete beam and a cast-in-place slab, when the composite beam was
subjected to repeated loading, has been done by Badoux and Hulsbos
(1967). The test program included 29 beams and the principal variables
were the amount of the ratio of the shear span to the effective depth of the
beam. Equations have been presented which yield a conservative
allowable stress for the horizontal shear in composite members under
repeated loads.

2.3 Shear transfer tests:

In the 1960’s Birkeland and Birkeland (1966) and Mast (1968) developed
a philosophy of connection design in which cracks are assumed to have
occurred disadvantageous locations within the region of the connection.
Reinforcement is then designed to transfer shear, normal force and
moment across these cracks when the connection is loaded. No
dependence is placed on the tensile strength of the concrete. They

9
proposed that the reinforcement needed to transfer shear across the cracks
be designed using the “Shear-Friction” method of design.
Provisions for the shear transfer reinforcement design using the
shear-friction method were subsequently included in the American
Concrete Institute, ACI, Building Code 318M-95. These provisions were
based on the test data obtained in monotonic loading tests of specimens
made from normal weight natural aggregates. Subsequent tests (Mattock
et al., 1976) showed that the shear transfer strength of lightweight
concrete under monotonic load is inferior to that of normal weight
concrete of the same compressive strength. It was, therefore, proposed
(Mattock et al., 1976) that the shear transfer strength of all-lightweight
concrete and of sand-lightweight concrete be taken, respectively, as 0.75
and 0.85 times the shear transfer strength of normal weight concrete of
the same compressive strength and having the same reinforcement.
Shear transfer across a definite plane must frequently be considered
in the design of precast concrete connections (Birkeland and Birkeland
1966, Mast 1968). A continuing study of the factors affecting shear
transfer strength was in progress at the University of Washington. Factors
so far included in the study were as follows:
1. The characteristics of the shear plane.
2. The characteristics of the reinforcement.
3. The concrete strength.
4. Direct stresses acting parallel and transverse to the shear plane.
5. Cyclic shear transfer.
The influence of the first three factors has been studied in tests
(Hofbeck et al., 1969) of monolithically cast “push-off” specimens as
seen in Fig. (2.1a). Tests (Chatterjee 1971, Vangsirirungruang 1971) to
study the influence of direct stresses acting parallel and transverse to the
shear plane were made on the “pull-off” and modified puss-off specimens
shown in Figs. (2.1b) and (2.1c), respectively.
Mattock et al. (1975) studied the effect of moment and normal force
in the shear plane on single direction shear-transfer strength. Tests were
reported of corbel type push-off specimens and of push-off specimens
with tension acting across the shear plane (Fig. (2.2)). To study the
influence of cyclic shear transfer, tests (Mattock, 1981) were made on a
crack in monolithic concrete or an interface between concretes cast at
different times. A typical specimen and the arrangements for test are
shown in Figs. (2.3) and (2.4), respectively.

10
 P P Shear transfer P
reinforcement

Shear plane

P P P

(a) (b) (c)

Fig. (2.1) Shear transfer test specimens: (a) push-off; (b) pull-off;
(c) modified push-off (Mattock et al., 1975).

Shear plane
V V
Shear plane

Bolts

V
(a) (b)
Fig. (2.2) Push-off specimens with moment or tension across shear plane
(a) corbel type, (b) tension type (Mattock et al., 1975).

11
 190.5 mm 190.5 mm

Shear plane
381 mm 254x127 mm

Closed stirrup shear

transfer

Secondary reinforcement, # 6 bars

# 3 bars

Shear transfer
reinforcement
190.5 mm

Faces “A”
Section

Fig. (2.3) Shear transfer specimen CB2M (Mattock, 1981).

In all cases, the shear transfer reinforcement crosses the shear plane
at right angles and is securely anchored so that it can develop its yield
strength in tension. Additional reinforcement was provided away from the
shear plane, to prevent failures other than along the shear plane. For
convenience, the ultimate shear strengths were expressed as average
shearing stresses (vu), obtained by dividing the ultimate shear force (Vu)
by the area of the shear plane.

12
Fig. (2.4) Arrangement for test (Mattock, 1981).

13
2.3.1 Characteristics of the shear plane:
Mast (1968) pointed out the need to consider the case where a crack may
exist along the shear plane before shear is applied. Such cracks occur for
a variety of reasons unrelated to shear, such as tension forces caused by
restrained shrinkage or temperature deformations or accidental dropping
of a member.
A crack in the shear plane reduces the ultimate shear strength of
under-reinforced specimens (Fig. (2.5)). The decrease is greater in the
push-off specimens than in the pull-off specimens. The shear strength of
the initially cracked specimens is not directly proportional to the amount
of reinforcement.

2.3.2 Characteristics of the reinforcement:

The reinforcement parameter (ρfy) can be changed by varying either the
reinforcement ratio (ρ), the reinforcement yield strength (fy), or both.
Also, for a given shear planes the reinforcement ratio can be changed by
changing the bar size and/or the bar spacing. In Fig. (2.6), the results are
compared to determine whether the way in which the reinforcement ratio
is changed has any effect on the relationship between ultimate shear
strength and the reinforcement parameter (ρfy). Fig. (2.7) shows that the
way in which ρ is changed does not affect the relationship between shear
strength and the reinforcement parameter (ρfy).
It was found (Mattock and Hawkins, 1972) that for given value of fy,
the specimens with 464 MPa (66 ksi) steel had slightly higher shear
strength than the specimens reinforced with the 350 MPa (50 ksi) steel.
This appears to indicate that at ultimate strength the higher strength steel
stirrups developed a stress greater than their yield point, i.e., strain
hardening had occurred. This is quite possible, as the yield plateau of the
higher strength reinforcement is considerably shorter than that of the
intermediate grade reinforcement. It therefore appears conservative to
assume that the relationship between ρfy and vu is the same for higher
strength reinforcement as for intermediate grade reinforcement, provided
the yield strength does not exceed 464 MPa.

14
 ρfy (MPa)
0 1 2 3 4 5 6 7 8 9 100 1 2 3 4 5

10
1400 PUSH-OFF TESTS PULL-OFF TESTS
9
1200 Uncracked 8
Uncracked
1000 7

vu 6
800 vu
(psi) Initially cracked 5 (MPa)
600
in shear plane
4

3
400

Initially cracked 2
200
fc’ = 4000 psi (345 MPa)
fy = 50 ksi (345 MPa) in shear plane 1

0 0
0 200 400 600 800 1000 1200 1400 0 200 400 600 800
ρfy (psi)

Fig. (2.5) Variation of shear transfer strength reinforcement parameter

(ρfy), with and without an initial crack along the shear plane
(Mattock and Hawkins, 1972).

ρfy (MPa)
0 1 2 3 4 5 6 7 8 9 10

10
1400 PUSH-OFF TESTS
9
1200
Bar size varies, 8
spacing constant 7
1000

vu 6
vu
800
(psi) (MPa)
5

600 4
Bar size constant,
spacing varies 3
400

fc’ = 4000 psi (345 MPa) 2

200 fy = 50 ksi (345 MPa)
1
Specimens initially cracked
0 0
0 200 400 600 800 1000 1200 1400
ρfy (psi)

Fig. (2.6) Effect of stirrup bar size and spacing on the shear transfer
strength of initially cracked push-off specimens (Mattock and
Hawkins, 1972).
15
 ρfy (MPa)
0 1 2 3 4 5 6 7 8 9 10

10
1400 PUSH-OFF TESTS
9
1200
fc’ =4000 psi (27.5 MPa) 8

1000 7

vu 6
vu
(psi) 800
5 (MPa)
600
fc’ = 2500 psi (17.2 MPa) 4

3
400
2
fy = 50 ksi (345 MPa)
200
Specimens initially cracked 1

0 0
0 200 400 600 800 1000 1200 1400
ρfy (psi)

Fig. (2.7) Effect of concrete strength on the shear transfer strength of

initially cracked push-off specimens (Mattock and Hawkins,
1972).
ρfy (MPa)
0 1 2 3 4 5 6 0 1 2 3 4 5 6

10
1400 UNCRACKED INITIALLY CRACKED
9
Push-off tests
1200
8
Push-off tests
1000 7

vu 6
(psi) 800 vu
5
(MPa)
600 4

Pull-off tests 3
400
Pull-off tests
2
200
1

0 0
0 200 400 600 800 0 200 400 600 800 1000
ρfy (psi)

Fig. (2.8) Effect on shear transfer strength of direct stress acting

parallel to the shear plane (Mattock and Hawkins, 1972).

16
2.3.4 Concrete strength:
The effect of variation in concrete strength on the shear strength of
initially cracked push-off specimens is illustrated in Fig. (2.8). For values
of ρfy below about 4.14 MPa (600 psi) the concrete strength does not
appear to affect the shear transfer strength. For higher values of ρfy the
shear strength is lower for the lower strength concrete. The concrete
strength therefore appears to set an upper limit value of ρfy, below which
the relationship between vu and ρfy established for 28.1 MPa (4000 psi)
concrete would hold for any strength of concrete equal to or greater than
the strength being considered, and above which the shear strength
increases at a lesser rate for concrete strength being considered.

2.3.5 Direct stress parallel to the shear plane:

In an earlier research (Hofbeck et al., 1969) a method was proposed for
the calculation of the shear transfer strength of initially uncracked
concrete. This was based on the average shear and normal stresses acting
on a concrete element in the shear plane, and made use of the failure
envelope for concrete proposed by Zia (1961). This approach predicted
the relationship between vu and ρfy very closely for the tests of initially
uncracked push-off specimens reported by Mattock and Hawkins (1972)
and also for tests of larger initially uncracked composite push-off
specimens reported by Anderson (1960). In the push off tests, direct
compressive stresses are exist parallel to the shear plane, and these were
taken into account in the calculation.
The ultimate shear strengths of the pull-off and push-off specimens
are compared in Fig. (2.9). For initially uncracked specimens, the pull-off
tests gave lower shear strengths than the push-off tests, indicating that a
direct tension stress parallel to the shear plane is detrimental to shear
transfer strength in initially uncracked concrete. However, the reduction
in shear strength appears to be due to a reduction in the cohesion
contribution of the concrete, and the rate of increase in vu with increase in
ρfy is approximately the same in both the pull-off and push-off tests. This
indicates that the method of calculation proposed by Hofbeck et al.
(1969) is faulty and cannot be extrapolated to the case of the pull-off test.

17
 ρfy (MPa)
0 2 4 6 8 10 12 14 16 18 20 22

16
Initially cracked,
Concrete failure Modified push-off tests
2000 14
envelope
12

vu 1500
10 vu
(psi)
Uncracked, (MPa)
Modified push-off tests 8
1000
6
Uncracked,
Push-off tests
4
500
Initially cracked, fc’ =4000 psi (27.5 MPa) 2
Push-off tests fy = 50 ksi (345 MPa)
0 0
0 500 1000 1500 2000 2500 3000
ρfy (psi)
Fig. (2.9) Effect on shear transfer strength of direct stress acting transverse
to the shear plane (Mattock and Hawkins, 1972).

For specimens cracked along the shear plane before being loaded in
shear, the shear strengths of the push-off and the pull-off specimens were
essentially the same for any given value of ρfy. This is important
practically, since it indicates that direct stresses parallel to the shear plane
may be ignored in design for shear transfer, if the design is based on the
relationship between vu and ρfy obtained in tests of initially cracked
specimens.

2.3.6 Direct stress transverse to the shear plane:

The effect of compressive stresses acting transverse to the shear plane
was studied by Chatterjee (1971) and Vangsirirungruang (1971) and
reported by Mattock and Hawkins (1972). Modified push-off specimens
were used. The ultimate shear strengths of the modified push-off
specimens which had shearing type failures are compared in Fig. (2.9)
with results from the push-off tests.
The effect of moment or tension force acting on the shear plane, on
the shear which can be transferred across a shear plane by a given

18
quantity and arrangement of reinforcement has been studied by Mattock
et al. (1975). Corbel type push-off specimens and push-off specimens
with tension acting across the shear plane have been used. It was found
that:
1. Moments in the shear plane less than or equal to the flexural ultimate
moment of the shear plane do not reduce the shear transfer strength.
2. Tension across the shear plane results in a reduction in shear transfer
strength equal to that, which would result from a reduction in the
reinforcement parameter (ρfy) by an amount equal to the tension
stress.

2.3.7 Cyclic shear transfer:

Mattock (1981) reviewed cyclically reversing and monotonic shear
transfer tests for companion initially cracked specimens of normal weight
and lightweight reinforced concrete. Both composite and monolithic
specimens were tested. The interfaces of all composite specimens were
roughened as required by Section 11.7.9 of the ACI Building Code 318-
77. In some cases the bond at the interface was deliberately broken. The
following conclusions has been obtained from the test results:
1. In the case of a shear plane in normal weight or lightweight
monolithic concrete and a rough interface between concretes cast at
different times, when good bond has been obtained at the interface,
the strength under cyclically reversing shear is about 80 percent of
the shear transfer strength under monotonic loading.
2. If the interface is roughened as specified in Section 11.7.9 of ACI
Code and if good bond is obtained at the interface, then after
cracking shear transfer behavior under both monotonic and cyclic
loading will be essentially the same as in the case of shear transfer
across a crack in monolithic concrete.
3. If bond at the interface between concretes cast at different times is
destroyed, shear transfer behavior under cyclic loading deteriorates
rapidly and the shear transfer strength is only about 0.6 of the shear
transfer strength under monotonic loading.

19
2.3.8 Slant shear test:
The most important aspect of the joining of two concretes is the strength
of the bond that can be achieved. This bond is crucial, as it determines
what forces can be transferred across the junction between the two
concretes. These forces arise mainly from strains in the additional
longitudinal reinforcement due to external loading; however, they may
also be caused by shrinkage and temperature differentials. The strength of
a junction between concretes cast at different times can be investigated by
a slant shear test. Impact tools are frequently used to roughen concrete in
practice (Cheong and MacAlerey, 2000).
Climaco (1990) has shown that the degree of roughening is relatively
unimportant as long as a reasonably rough surface is obtained, and
excessive damage to the concrete (i.e., cracking of the matrix or
dislodging of aggregate particles) is avoided.
Cheong and MacAlerey (2000) presented a description on an
experimental investigation into the behavior of reinforced concrete beams
strengthened by jacketing. Static and dynamic loads tests to failure were
carried out on 61 slant shear prisms and 13 jacketed reinforced concrete
T-beams. The concrete used in jacket was preplaced aggregate concrete.
The strength of the bond between preplaced aggregate concrete and plain
concrete was assessed by slant shear tests and a Mohr-Coulomb-type
failure envelope was derived. An example of the test results plotted for
fcu = 45 N/mm2 concrete is shown in Fig. (2.10).

Shear stress
τ 14
2
(N/mm )
12 φ angle of
friction
10

Cohesion
2
c
0
0 1 2 3 4 5 6 7 8 9
Normal stress (N/mm2)

Fig. (2.10) Slant shear test parameters – grade 45 concrete

(Cheong and MacAlerey, 2000)
20
2.3.9 Strut-and-tie model:
A softened strut-and-tie model for determining interface shear capacity
was proposed by Hwang et al. (2000). Contrary to the shear-friction
concept, the proposed theory predicts that ultimate failure is caused by
the crushing of concrete in the compression struts formed after cracking
of the concrete. The shear strength predictions of the proposed model and
the empirical formulas of the ACI 318-95 building code were compared
with collected experimental data from 147 specimens. Examination of
existing experimental data indicated that the softened strut-and-tie model
was capable of predicting the interface shear strengths of both the push-
off and the pull-off specimens with or without the precracked shear
planes. The comparison showed that the performance of the softened
strut-and-tie model was better than the ACI code approach for the
parameters under comparison. The parameters reviewed included
concrete strength and amount of shear transfer reinforcement.

2.3.10 Dynamic shear transfer:

Displacement controlled shear tests on concrete lift joint specimens with
different surface preparations are conducted by Fronteddu et al. (1998) to
compare dynamic sliding joint behavior with the static behavior. Fig.
(2.11) shows the specimen geometry used in the shear test. Experimental
results indicate that the coefficient of friction decreases with the increase
in normal load, under both static and dynamic shear. The shear strength is
dependent on surface preparation.
Monolithic specimens and water-blast joints show higher shear
strengths than untreated joints and plane independent joint surfaces. An
empirical lift joint constitutive model was developed as a function of a
basic friction coefficient and a roughness friction coefficient that was
dependent upon the type of surface preparation.

21
Fig. (2.11) Specimen geometry (Fronteddu et al., 1998).

22
 CHAPTER THREE
RECOMMENDATIONS FOR CONSTRUCTION
AND DESIGN

3.1 Introduction:
It is always a big boost for any new construction method when a major
specification writing body incorporates the new construction method into
its latest code revision. For composite construction, this breakthrough
came in 1944. The American Association of State Highway Officials
included composite construction into its specifications.
Today several specifications and codes are dealing with and have
recommendations for the design and construction of composite concrete
beams. These recommendations apply to composite concrete beams
subjected primarily to static loads.
This chapter presents methods of construction and rules for design of
composite concrete beams. These rules are based generally on the ACI
Code “Building Code Requirements for Reinforced Concrete (ACI
318M-95)”.

3.2 Design of composite concrete beam:

Design of the composite concrete-concrete member should be in
accordance with Chapter 17 of “Building Code Requirements for
Reinforced Concrete (ACI 318M-95)”. Other sections of the code which
affect composite members are noted in Chapter 9 which includes control
of deflections and the sections on shear-friction (Section 11.7 of the
Code), T-beam construction (Section 8.10), and deep beams (Sections
10.7 and 11.9).
The ACI Code states that if the following four criteria are met,
horizontal shear need not be calculated.
1. Contact surfaces between the beam and slab is clean and intentionally
roughened.
2. Minimum ties are provided.
3. Web members are designed to resist the entire vertical shear.
4. All stirrups are fully anchored into all intersection components.
Full transfer of the shear forces should be provided at the interface,
or there is no provision for partial composite action. If the four
requirements above are met, then full transfer of horizontal shear can be
23
assumed. If the horizontal shear stress is to be calculated, the computation
is quite simple. The horizontal shear stress is found from:
V
v dh = u .......(3.1)
φb v d
where vdh is the design horizontal shear stress; Vu is the factored shear
force at section considered; φ is a strength reduction factor (0.85 for
shear); bv is the width of cross section at contact surface being
investigated for horizontal shear; and d is the distance from extreme
compression fiber to centroid of tension reinforcement for entire
composite section.
Horizontal shear is transferred by direct bond, shear friction, and
reinforcement across the interface. At working loads, the direct bond is
usually intact. Shear friction becomes active when the bond is broken, as
by cyclic loading, or at high load levels.
Provisions of Chapter 17 in ACI Code may be applied for design of
composite concrete flexural members defined as precast and/or cast-in-
place concrete elements constructed in separate placements but so
interconnected that all elements respond to loads as a unit.
Individual elements should be investigated for all critical stages of
loading.
In strength computations of composite members, no distinction may
be made between shored and unshored members.
Tests have indicated that the strength of a composite member is the
same whether or not the first element cast is shored during casting and
curing of the second element.
When used, shoring should not be removed until supported elements
have developed design properties required to support all loads and limit
deflections and cracking at time of shoring removal.

3.2.1 Vertical shear strength:

When an entire composite member is assumed to resist vertical shear,
design may be in accordance with requirements of Chapter 11 as for a
monolithically cast member of the same cross-sectional shape.
Shear reinforcement should be fully anchored into interconnected
elements in accordance with 12.13.
Extended and anchored shear reinforcement may be permitted to be
included as ties for horizontal shear.

24
3.2.2 Horizontal shear strength:
In a composite member, full transfer of horizontal shear force should be
assured at contact surfaces of interconnected elements.
Unless calculated in accordance with 17.5.3, design of cross sections
subjected to horizontal shear shall be based on:

Vu ≤ φVnh .......(3.2)

where Vnh is nominal horizontal shear strength in accordance with the

following:
a. When contact surfaces are clean, free of laitance, and intentionally
roughened, shear strength (Vnh) shall not be taken greater than
0.6bvd in newtons.
b. When minimum ties are provided and contact surfaces are clean
and free of laitance, but not intentionally roughened, shear strength
shall not be taken greater than 0.6bvd.
c. When ties are provided and contact surfaces are clean, free of
laitance, and intentionally roughened to a full amplitude of
approximately 5 mm, shear strength shall be taken equal to
(1.8+0.6ρvfy)λbvd but not greater than 3.5 bvd.
Where ρv is the ratio of tie reinforcement area to area of contact
surface; fy is the specified yield strength of nonprestressed
reinforcement; and λ is the correction factor related to unit weight of
concrete.
When factored shear force (Vu) at section considered exceeds φ(3.5
bvd), design for horizontal shear shall be in accordance with 11.7.4 of the
ACI Code, which covers shear friction design.
As an alternative to the above, horizontal shear shall be determined
by computing the actual change in compressive or tensile force in any
segment, and provisions shall be made to transfer that force as horizontal
shear to the supporting element. The factored horizontal shear force shall
not exceed horizontal shear strength (φVnh) where area of contact surface
(Ac) shall be substituted for bvd.
When ties provided to resist horizontal shear, the tie area to tie
spacing ratio along the member shall approximately reflect the
distribution of shear forces in the member.
When tension exists across any contact surface between
interconnected elements, shear transfer by contact shall be permitted only
when minimum ties are provided.
25
3.2.3 Ties for horizontal shear:
When ties are provided to transfer horizontal shear, tie area shall not be
less than
1 b vs
Av = .......(3.3)
3 fy
where Av is the area of shear reinforcement within a distance s; and s is
the spacing of shear reinforcement; and tie spacing shall not exceed four
times the least dimension of supported element, nor 600 mm.
Ties for horizontal shear shall consist of single bars or wire, multiple
leg stirrups, or vertical legs of welded wire fabric (plain or deformed). All
ties shall be fully anchored into interconnected elements.

3.2.4 Design of slab:

The slab may be designed as either a one-way or two-way slab in
accordance with ACI Code. In the design of the slab, the stresses caused
by composite action may be neglected.
The ACI Code specifies temperature and shrinkage reinforcement in
slabs where the principal reinforcement extends in one direction only. In
a composite beam system, the slab is designed as a one-way slab
spanning transversely between the beams. The temperature and shrinkage
steel required by the code, in addition to serving its primary function, is
quite effective in maintaining the structural integrity of the slab overhang
portion of the T-beam. The code limits the spacing of shrinkage and
temperature steel to 5 times the slab thickness, or 500 mm (7.12.2.2).

3.2.5 Effective flange width:

The object in designing the composite member is to have a member at
least as strong as a monolithic T-beam. Consequently, the restrictions on
the effective flange width of T-beams also apply to composite beams
(Cook, 1977).
It should be obvious that if the beams are spaced quite a apart. As shown
in Fig. (3.1), the entire slab will not be effective as a compression flange
for the beam. This same Figure should also make it apparent that the slab
thickness and web (stem) width also affect the effective flange width.
The effective flange width, according to the ACI Code (Section
8.10), shall be the least of the following three values:
1- Width of stem plus 16 times the slab thickness.
2- Distance center to center of beams.

26
3- One-quarter of the span length.
4- Isolated beam shall have a flange thickness not less than ½ width
of web and an effective flange width not more than four times the
width of web.

Fig. (3.1) Poor proportions.

3.2.6 Deflections:
The composite beam is stiffer, and consequently it deflects less than the
noncomposite beam of the same size. Deflections are computed by elastic
analysis using formulas found in specifications.
Live load deflection is always carried by the composite section.
Dead loads are not so simple. Dead load deflections largely depend on the
construction method. If the beam is shored, the dead load is carried by the
false work. If the beam is not shored, the dead load is carried by the bare
beam and dead load deflections should be checked.
In concrete-concrete composite member, long-term loading must be
considered. If shoring is used, the composite member may be considered
as equivalent to a monolithic cast-in-place member for deflection
purposes (Cook, 1977). The effects of shrinkage and creep can be
accounted for by using an amplification factor to compute the additional
deflection due to long-term effects. If shoring is not used ACI Code states
that if the member meets minimum thickness requirements, deflections
occurring after the beam become composite need not be calculated.
However, the precast beam must be investigated for long-term effects
prior to the beginning of effective composite action.
Prior to the adoption of the strength method of design by ACI Code,
concrete sections were larger and stiffer and consequently had smaller
deflections. With the use of the strength method, plus higher strength
steels and concrete, members have become more slender, so that
deflection and crack control have become serous considerations.
The familiar formulas for deflection are used with concrete.
However, those formulas include E, the modulus of elasticity, and I, the

27
 moment of inertia of the member. When designing and building with
most other materials, the values of E and I are constant. However, in
concrete structures, both of these factors become variables.

Live load deflections- Live load deflections should be computed based

on the moment of inertia (I) of the transformed composite section using
the full value of the concrete modulus of elasticity (Ec).

Dead load deflection

1. For beams shored during construction, the dead load deflections should
be computed based on the moment of inertia of the transformed
composite section using one-half the value of the modulus of elasticity
(Ec).
2. For beams not shored during construction, the dead load deflections
should be computed on the basis of the moment of inertia of the
prefabricated beam alone except that deflections due to dead loads
applied after the concrete slab has attained 75 percent of the specified
28-day strength should be computed according to Point 1.

3.2.7 Uplifts:
Virtually every specification writing body recognizes the tendency of the
slab virtually to separate from the beam under some types of loading.
During investigations of bonded-aggregate composite beams, Cook
(1977) found those concentrated loads at mid-span often caused a visible
loss of bond at the beam-ends at relatively low load levels.
Although most codes recommend the use of a connector designed to
prevent uplift, none specifies any computation or recommended limits for
the uplift connectors.
The ACI Code permits a closed-loop stirrup (Fig. (3.2a)), which
furnishes good uplift resistance. Other configurations of the stirrups can
also be used such as extended stirrups (Fig. (3.2b,c&d)) to provide
adequate embedded length, truss type (Fig. (3.2g)), welded wire fabric
(Fig. (3.2e)), inverted stirrup (Fig. (3.2f)), and single bar (Fig. (3.2h)).
The shear connection in the concrete-concrete composite beam
comes from two sources:
1- Intentional roughening of the top surface of the precast unit.
2- Ties or extended stirrups across the beam-slab interface.

28
 (a) (b) (c) (d)

(e) (f) (g) (h)

Fig. (3.2) Stirrup configurations (a) closed-loop stirrup, (b) out extended
stirrup, (c) in extended stirrup, (d) right extended stirrup,
(e) welded wire fabric, (f) inverted stirrup, (g) truss type, and
(h) single bar.

3.2.8 Continuity:
Determination of moments, shears, and thrusts- Moments, shears, and
thrusts produced by external loads should be determined by elastic
analysis. For the purposes of such analysis, the moment of inertia of the
gross composite section may be used throughout the length of the beam.

Sections resisting negative moments

1. In negative moment regions of continuous or cantilever beams, the
bending moment may be assigned either to the prefabricated section
alone or the composite section composed of the prefabricated section
and the slab reinforcement.

29
2. When the slab is continuous at the supports of beams, reinforcement
should be provided sufficient to prevent excessive cracking of the
slab.

Shear connection in regions of negative moments

When the negative moments are assigned to the prefabricated section
alone, shear connection between the prefabricated beam and the slab need
not be provided throughout the full length of the beam.

3.2.9 Allowable stresses, moduli of elasticity, and load factors:

Allowable stresses- Allowable stresses for reinforcing steel and concrete
specified in ACI Code are recommended for the design of composite
beams. In structures composed of elements with different concrete
strength, the allowable stresses in each portion should be governed by the
concrete strength of the portion under consideration.

Tension in concrete- The tensile resistance of concrete should be

neglected.

Moduli of elasticity- The following values for the moduli of elasticity

should be used:
1. Steel: Es = 200,000 MPa
2. Concrete: E c = 4700 f c' ; where fc’ is the 28-day specified
compressive strength of the concrete under consideration.

Load factors- For designs based on ultimate flexural strength, load

factors given in the ACI Code 318M-95 are recommended.

3.2.10 Differential shrinkage and creep:

Differential shrinkage and creep between the precast beam concrete and
the slab concrete cast at different ages also produce stresses and
deflections that are additive to those due to gravity loading.
The effects of differential shrinkage and creep are usually not of
great importance, but may be of some consequence where tensile stresses
and/or deflections are critical. This is particularly true when the age of the
precast beam concrete at the time of slab casting is relatively large (i.e.,
more than a few weeks) and/or when the difference in the quality of the

30
two concretes with regard to shrinkage and creep is significant.
Differential shrinkage and creep cause a reduction in the cracking
moment under load but have a negligible effect on the ultimate strength
of a composite member (Sabnis, 1979).

3.3 Shear friction:

3.3.1 General considerations:
With the publication of ACI 318-83, Section 11.7 was completely
rewritten to expand the shear-friction concept to include applications:
(1) where the shear-friction reinforcement is placed at an angle other than
90 degrees to the shear plane, (2) where concrete is cast against concrete
not intentionally roughened, and (3) with lightweight concrete. In
addition, a performance statement was added to allow “any other shear-
transfer design methods” substantiated by tests.

3.3.2 Shear-friction concept:

The shear-friction concept that provides a convenient tool for the design
of members for direct shear where it is inappropriate to design for
diagonal tension, as in precast connections, brackets, and corbels. The
concept is simple to apply and allows the designer to visualize the
structural action within the member or joint. The approach is to assume
that a crack has formed at an expected location, as illustrated in Fig. (3.3).
As slip begins to occur along the crack, the roughness of the crack surface
forces the opposite faces of the crack to separate. This separation is
resisted by reinforcement (Avf) across the assumed crack. The tensile
force (Avffy) developed in the reinforcement by this strain induces an
equal and opposite normal clamping force, which in turn generates a
frictional force (Avffyμ) parallel to the crack to resist further slip, where μ
is the coefficient of friction.
Shear-friction design is to be used where direct shear is being
transferred across a given plane. Situations where shear-friction design is
appropriate include the interface between concretes cast at different
times, an interface between concrete and steel, and connections of precast
constructions, etc. Successful application of the concept depends on
proper selection of location of the assumed slip or crack.

31
3.3.3 Shear-transfer design methods:
The shear-friction design method presented in Section 11.7.4 is based on
the simplest model of shear-transfer behavior, resulting in a conservative
prediction of shear-transfer strength. Other more comprehensive shear-
transfer relationships provide closer predictions of shear-transfer strength.
The performance statement of 11.7.3 “…any other shear-transfer design
method…” includes the other methods within the scope and intent of
11.7.

3.3.4 Shear-friction design method:

As with the other design applications, the code provisions for shear-
friction are presented in terms of shear-transfer strength (Vn) for direct
application in the basic shear strength relation:

Required shear - transfer strength ≤ Design shear - transfer strengt

Vu ≤ φVn .......(3.4)
Note that φ is 0.85 for shear and torsion (9.3.2.3). Furthermore, it is
recommended that φ = 0.85 be used for all design calculations involving
shear-friction, where shear effects predominate. This is based on the
requirement of 11.9.3.1 which specifies the use of φ = 0.85 for all design
calculation in accordance with 11.9.
The required shear-transfer strength for shear-friction reinforcement
perpendicular to the shear plane is:

Vu ≤ φA vf f y μ .......(3.5)

The required area of shear-friction reinforcement (Avf) can be

computed directly from:
V
A vf = u .......(3.6)
φf y μ
The condition where shear-friction reinforcement crosses the shear-
plane at an angle αf other than 90 degrees is illustrated in Fig. (3.4). The
tensile force (Avffy) is inclined to the crack and must be resolved into two
components: (1) a clamping component (Avffy sinαf) with an associated
frictional force (Avffy sinαf μ), and (2) a component parallel to the crack
that directly resists slip equal to Avffy cosαf. Adding the two components
resisting slip, the shear-transfer strength requirement becomes:

32
 [
Vu ≤ φ A vf f y sin α f μ + A vf f y cos α f ] .......(3.7)

For shear reinforcement inclined to the crack, the required area of

shear-friction reinforcement (Avf) can be computed directly from:
Vu
A vf = .......(3.8)
φf y (μ sin α f + cos α f )
Note that Eq. (3.7) applies only when the shear force (Vu) produces
tension in the shear-friction reinforcement.
The shear-friction method assumes that all shear resistance is
provided by friction between crack faces. The actual mechanics of
resistance to direct shear are more complex, since dowel action and the
apparent cohesive strength of the concrete both contribute to direct shear
strength. It is, therefore, necessary to use artificially high values of the
coefficient of friction (μ) in the shear friction equations so that the
calculated shear strength will be in reasonable agreement with test results.
Use of these high coefficients gives predicted strengths that are a
conservative lower bound to test data, as shown in Fig. (3.5). The
modified shear-friction design method given in R11.7.3 is one of several
more comprehensive methods, which provide closer estimates of the
shear-transfer strength (Ghosh et al., 1996).

3.3.5 Coefficient of friction:

The “effective” coefficient of friction (μ) for the various interface
conditions include a parameter λ which accounts for the somewhat lower
shear strength of all-lightweight and sand-lightweight concretes. For
example, the μ value for all lightweight concrete (λ = 0.75) placed
against hardened concrete not intentionally roughened is 0.6(0.75) = 0.45.

33
 Shear plane slip
(assumed crack) Vu

Shear friction Vu
reinforcement
Vu

Avf fy μ
Avf fy

Avf fy
Vn = Avf fy μ

Fig. (3.3) Idealization of the shear-friction concept.

Vu

αf
Αvf fy sinαf μ
αf Avf fy
Avf fy sinαf
Avf fy
Avf fy cosαf

Vn = Avf fy sinαf μ + Αvf fy cosαf

Fig. (3.4) Idealization of inclined shear-friction reinforcement.

34
 ρfy (psi)
0 200 400 600 800 1000 1200
1600
Test Results
10
1400

Modified 1200
8
Shear-Friction
Method 1000
Vn/Ac (MPa)

Vn/Ac (psi)
Vn/Ac Limit 800

4 600
1.4 1.0 0.8 μ=0.6
400
2
Shear-Friction 200

Design Method
0 0
0 2 4 6 8

ρfy (MPa)

Fig. (3.5) Effect of shear-friction reinforcement on shear

transfer strength (Ghosh et al., 1996).

3.3.6 Maximum shear-transfer strength:

The shear-transfer strength (Vn) cannot be taken greater than 0.2fc’, nor
5.52 MPa times the area of concrete section resisting shear transfer. This
upper limit on Vn effectively limits the maximum reinforcement, as
shown by Fig. (3.5). Also, for lightweight concrete, 11.9.3.2.2 limits the
shear-transfer strength (Vn) along the shear plane for design applications
with low shear span-to-depth ratios, such as brackets and corbels.

3.3.7 Normal forces:

Equation (3.5) and (3.7) assume that there are no forces other than shear
acting on the shear plane. A certain amount of moment is always present
in brackets, corbels, and other connections due to eccentricity of loads or
applied moments at connections. Tensile reinforcement required for
bending due to eccentricity of the loads, or other causes, is determined in
the normal manner.
Joints may carry a significant amount of tension due to restrained
shrinkage or thermal shortening of the connected members. Therefore, it

35
is recommended, although not generally required, that the member be
designed for a minimum direct tensile force at least 0.2Vu in addition to
the shear. This minimum force is required for design of connections such
as brackets or corbels (see 11.9.3.4), unless the actual force is accurately
known. Reinforcement must be provided for direct tension according to
11.7.7, using As = Nuc/φfy where Nuc is the factored tensile force.
Since direct tension perpendicular to the assumed crack (shear
plane) detracts from the shear-transfer strength, it follows that
compression will add to the strength. Section 11.7.7 acknowledges this
condition by allowing a “permanent net compression” to be added to the
shear-friction clamping force (Avffy). It is recommended, although not
required, to use a reduction factor of 0.9 for strength contribution from
such compressive loads.

3.3.8 Additional requirements:

Section (11.7.8) requires that the shear-friction reinforcement be
“appropriately placed” along the shear plane. Where no moment acts on
the shear plane, uniform distribution of the bars is proper. Where a
moment exists, the reinforcement should be distributed in the flexural
tension zone.
Reinforcement should be adequately embedded on both sides of the
shear plane to develop the full yield strength of the bars. Since space is
limited in thin walls, corbels, and brackets, it is often necessary to use
special anchorage details such as welded plates, angles, or cross bars.
Reinforcement should be anchored in confined concrete. Confinement
may be provided by beam or column ties, “external” concrete, or special
added reinforcement.

3.4 Analysis of composite section:

The analysis of composite sections and the proportioning of members
varies slightly with the materials used. In concrete-concrete construction,
either load factor (ultimate strength) or service load (working stress)
design method can be used. However, the shear connection is usually
designed by the ultimate strength method, even if the service load method
is used to proportion the beam.
In the elastic analysis of the composite beam the following
assumptions are made:
1. The beam and slab materials are both elastic.
36
2. These two materials are related by the modular ratio, n.
3. The shear connection provides full interaction between the beam and
slab.
4. Any concrete below the neutral axis is considered ineffective. (This is
reversed in negative bending regions.)

3.5 Construction methods:

Camber- Necessary provisions should be taken in the design and
construction to prevent excessive dishing of the slab in beams built with
shores and excessive thickening of the slab in beams built without shores.

Treatment of beam surfaces- Surfaces of prefabricated beams in contact

with the slab should be cleaned of any foreign or loose material before
casting the slab.

3.5.1 Intentional roughening:

Intentional roughening means roughening the top surface of the precast
member to full amplitude of approximately 5 mm (¼ in.). This intentional
roughening is best done in the casting yard before the beams are shipped
to the site.
Roughening can be done in several ways. Sand blasting and needle
gun both does a good job. If sand blasting is used, it makes much more
sense to do the job in the casting yard under controlled conditions.
Another possible method of accomplishing the roughening is to cast the
beam units upside down, and coat the bottom of the form with a retarding
agent. After the beams are stripped, the top surface of the beam can be hit
with a high-pressure water jet. However, this method is tricky and
requires good quality control, and it should only be tackled by an
experienced casting yard crew.
Power-driven wire brushes on cured concrete are generally
unsatisfactory for roughening. They do the job, but often leave behind a
black greasy stain, which inhibits bond.
The bush-hammer does an effective job of roughening. A simple
hand rake, which will make 5 mm grooves in the top of the concrete
before it takes its initial set, will also suffice.
An AASHTO Class 5, wire brushed or scrubbed finish will also
produce satisfactory roughening. This type of finish is produced by
scrubbing the surface of green concrete with stiff wire or fiber brushes

37
using a solution of one part muriatic acid to four parts water. When the
scrubbing has progressed enough to provide the required roughening, the
surface should be flushed with water, to which a small amount of
ammonia has been added, to remove all traces of the acid.
Raking the plastic concrete or chiseling hardened concrete surfaces
also can be used.
Intentional roughness may be assumed only when the interface is
roughened with full amplitude of approximately 5 mm. Methods of
obtaining this intentional roughness are spelled out.
Practically, a complete design for the beam stem will include
vertical ties or stirrups for web reinforcement. All that remains is to detail
the ties so that they extend for enough up into the slab to be effective in
transmitting the horizontal shear. The shear connector check then reduces
to a simple check of whether the intentional roughening is required to
supplement the shear reinforcement.
The vertical ties can have any of several configurations, but the
closed loop is usually considered the most effective because it provides
very good uplift resistance. This assumes that there is enough room to
provide the full-embedded length of the bar.

3.5.2 Shoring and unshoring:

Throughout the various codes and specifications reference is made to
shored and unshored construction. Shoring means support under the
beams so that the dead load of the beam, formwork, and wet concrete of
the slab is carried by the temporary supports. The supports should be
wedged under the beam so that the beam member does not deflect under
the dead load.
In shored construction, the temporary supports must remain in place
until the concrete has gained 75% of its design strength, which usually
means about 14 days. After the slab is a week or 10 days old, forming and
shoring can be done for the next level, but the slab for the next level
should not be cast until the composite beam and slab at the first level is
capable of supporting the dead load of the second floor.
Unshored construction goes up faster. After the slab for a given floor
level has been cast, only few days are necessary before the slab is hard
enough to be used by other trades, and forming can begin for the next
floor level (Cook, 1977).
38
 To notify the recommendations for design in other codes and
specifications, Table (3.1) contents a comparison between some available
specifications and codes with the ACI Code. These codes and
specifications are:
1. Iraqi Building Code Requirements for Reinforced Concrete (Code
1/1987).
2. American Association of State Highway and Transportation Officials
(AASHTO), 1989, specifications.
3. British Standards Institution (BS 5400: Part 4:1978).
4. Standards Association of Australia (SAA), Prestressed Concrete Code,
AS 1481, 1974.

39
 Table (3.1) Comparison between the available specifications and codes.

Shear Strength Vh (N/mm2) Reinforcement Max.

Specifications Roughening
Surface preparation Effective flange width cross-sectional reinforcement
and codes (mm)
area spacing (mm)
bw+1/7 effective span for
continuous beam,
spray of water, bw+1/5 effective span for
1.105 brush, simply supported or 0.0015 of the 4 tw
Iraqi 0.425 0.340 6
1.870 sand blasting, cantilever, contact area 600
needle gun bw+16 tf,
center to center,
4bw for isolated T-beam
Sand blasting,
Needle gun,
¼ span,
Cast upside down,
ACI 0.550 0.550 2.410 5 bw+16 tf ,
Power-driven wire brush,
center to center
Push-hammer,
Hand rake
¼ span,
bw+12 tf , 4 tw
AASHTO 0.248 0.248 1.104 6
center to center, 600
4bw for isolated T-girder
Spray of water,
bw+1/5 distance between
Stiff brush, 0.15% of the 4 tw
BS 0.380-0.640 0.360-0.500 1.220-1.450 points of zero moments,
Sand blasting, contact area 600
center to center
Needle gun
0.450 2
AS 1481 0.050fc’ 0.150fc’ 157 mm 300
 CHAPTER FOUR
MECHANICAL PROPERTIES AND BEHAVIOR
OF MATERIALS

4.1 General:
Reinforced concrete is a composite material made up of concrete and
steel. The stress-strain relationship for steel is very well defined.
Concrete, however, is heterogeneous and has completely different
properties in tension and in compression.
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.

4.2 Concrete in uniaxial state:

4.2.1 Concrete stress-strain behavior in uniaxial
compression:
4.2.1.1 General characteristics:
The experimental investigation of the uniaxial stress-strain behavior of
concrete in compression together with the analytical representation of this
behavior by proper mathematical forms were the matter of interest and
the subject of extensive study during the past century.
Several types of tests have been used to trace the stress-strain
relation in compression. Concentric tests for various specimen types
(cylinders, cubes, and prisms) were adopted as a conventional tool for
obtaining this relation even for flexural considerations. On the other hand,
eccentric tests and beam tests were utilized to produce more
representative relation for flexural compression, to derive correlation for
flexural compression, and to derive correlation with conventional tests.
Generally, the stress-strain relation is composed of a strain-
hardening portion extending to the maximum stress (fc’), which is of the
same shape for various testing conditions, followed by a strain-softening
portion extending to the ultimate compressive strain (εcu), which is much
affected by the testing machine and the aggregate characteristics, Fig.
(4.1). The essential features defining the stress-strain relation are:

41
 σ

ft
εcu εm
εcr ε

fc’

Fig. (4.1) Typical stress-strain curve for concrete.

1- Compressive strength:
It is the average maximum stress obtained from conventional testing
of concrete specimens, usually denoted by (fc’) for standard (150x300
mm) cylinder strength.
2- Initial elastic modulus (Eci):
It is the initial slope of the stress-strain curve or may be considered,
due to experimental difficulties, as the secant modulus (Ecn) measured
at a stress of 0.4fc’. The ACI Code suggested an empirical formula for
normal weight concrete to estimate this parameter:
E cn = 4700 f c′ .......(4.1)
3- The strain corresponding to the maximum stress (εm):
This parameter may not vary much with compressive strength but it is
much affected by testing procedures and strain rates. The results of
most conventional tests have provided its values in the range of
(0.0015-0.0022).
4- Ultimate unconfined strain (εcu):
It is very much affected by the type of testing machine. Investigators
have shown it to be in the range of (0.0022-0.0042) for different
strength. However, it is considered in practice to be 0.003 (ACI Code).

42
4.2.1.2 Review of analytical relations:
Since 1899, many mathematical functions have been proposed to
represent the ascending part of the stress-strain relation in compression
(e.g. parabola, hyperbola, ellipse, cubic parabola, sine wave, etc.)
(Hognestad, 1957). The importance of introducing the descending part
was recognized in early fifties.
Since then, many equations were proposed to represent the complete
stress-strain relation. These are found in several forms (parabolic-linear
relation (Hognestad, 1951), exponential relations (Smith and Young
1956, Sturman et al. 1965), fractional relation (Desayi and Krishnan,
1964), and polynomial function (Kabaila, 1964)). Other investigators
proposed similar equations with certain changes, modifications, and
refinements (Saenz 1964, Tulin and Gerstle 1964, Basu 1967, Kayal
1984, Medland and Taylor 1971, Al-Sabah and Al-Ne’aimi 1988, Smith
and Orangun 1969, Gangadharam Reddy 1980, Wang et al. 1978,
Carreira and Chu 1985, Tsai 1988). In spite of the large amount of
equations provided, several investigators used some idealized relations
due to their simplicity in nonlinear analysis programs (elastic-perfectly
plastic relations (Lin and Scordelis 1975, Seniwongse 1979), and
piecewise linear relations (Aldstedt and Bergan 1978, Zeris and Mahin
1988, Holzer et al. 1979)). A complete review of these equations is
presented by Rasheed (1990).

4.2.2 Concrete stress-strain behavior in uniaxial tension:

4.2.2.1 General:
Cracks may form soon after casting, due to the settlement of plastic
concrete. During hardening, hydration heat produces temperature
differences between internal and external portions and is likely to cause
cracking in thick elements. Rapid evaporation can also lead to plastic
shrinkage cracking.
After hardening, besides the dead load and the live load, restraint
forces in statically indeterminate structures produced by differential
settlement of foundations or by different temperatures of the top and
bottom faces of the element may cause cracking. Shrinkage often leads to
cracks between connected members of significantly different sizes or
ages (Balázs, 1993).
Cracks occur whenever the principal tensile stress from loads or
restraint forces exceeds the tensile strength of concrete. Due to the

43
formation of cracks, the compatibility of deformations between the steel
and concrete is not maintained. The accumulation of the strain differences
produces relative displacements (slip) between the steel and concrete. The
width of a crack at the steel is provided by the sum of the two slips
reaching the crack from either side. This emphasizes that the cracking
phenomenon is governed by the slip.
Cracking is the most distinctive feature of the nonlinear behavior of
reinforced concrete structures. It affects both local behavior such as bond
slip and tension stiffening and global behavior such as stiffness
distribution along the member. This process is governed by concrete
tensile properties and it needs to be clearly understood and properly
formulated together with its relevant phenomena.

4.2.2.2 Review of experimental work and analytical

relations:
The shape of the uniaxial stress-strain relation in tension is highly
dependent on the testing technique adopted. In the past, three well-known
types of tests have been commonly used for tensile testing of plain
concrete namely, the direct tensile test, the beam test, and the split
cylinder test. While the first test was almost difficult to be controlled and
hence provided the stress-strain diagram up to tensile strength followed
by sudden failure, the other two tests were conceptually based on elastic
theory. Thus none of the above tests could provide the actual stress-strain
diagram which was generally assumed to be linear with a slope of Eci up
to tensile strength followed by an abrupt drop to zero stress.
For that reason, early models used in most analytical studies were
based on such tension-brittle model which depends on tensile strength
only (Lazaro and Richards 1973, Nam and Salmon 1974, Seniwongse
1979, Al-Metwally and Chen 1989, Zeris and Mahin 1988). Other studies
even neglected the tensile capacity of concrete (Chang and Ferguson
1963, Kroenke et al. 1973, Kayal 1984). It has been reported (Lazaro and
Richards 1973, Bazant and Cedolin 1979) that the commonly (that time)
used tension-brittle model was providing inconsistent results especially
under service loads.
It is remarkable to note that two studies (Hughes and Chapman
1966, Evans and Marathe 1968) succeeded during late sixties to trace
experimentally some descending branches of the tensile stress-strain
diagram using modified testing machines.

44
 Tension softening may be defined here as the slop of the descending
branch of the stress-strain relation, which is continuously degrading as
strain increased. Analytically, attention has been given to propose
idealized tension-softening model in which concrete may exhibit gradual
decrease of stress with progressive cracking and crack widening. Such
models were rather simplified due to the lack of sufficient experimental
data. Hillerborg et al. (1976) were the first to introduce analytically the
tension-softening behavior through their fictitious crack model on the
basis of fracture mechanics and finite elements. Several parameters were
used to represent tension softening behavior namely, the tensile strength
(ft’), the initial modulus (Eci), the open crack strain (eo), the fracture
energy exerted (Gf), and other parameters defining the shape of the
softening branch.
All of the proposed σ-w (or displacement δ) relations, except for the
linear one, show qualitatively the same behavior. The first part of these
relations is very steep, whereas the second part is more flat. In addition,
most practical applications are not very sensitive to the exact shape of the
σ-w relation (Karayannis, 2000).
The simplest forms of such idealized models were of linear
descending branch (Bazant and Oh, 1983), Fig. (4.2a), of bilinear and
trilinear descending branches, Fig. (4.2b) and (4.2c). More complex
models having nonlinear descending branches are also available as
exponential functions (Gopalaratnam and Shah 1985, Reinhardt et al.
1986) and a power function (Reinhardt et al., 1986), Fig. (4.2d).
The cracking strain (εcr) cannot be easily measured during
conventional tensile tests. It is highly sensitive to many conditions (rate
of straining, effect of age, testing method, aggregate characteristics,
shrinkage, etc.). Medland and Taylor (1971) adopted a value of
(εcr= 0.0001). A good survey of the available literature measuring (εcr) by
different methods was done by Al-Rawi (1986).

45
f C1 fC1
f Cr f Cr

EC EC

ε Cr ε m ε1 ε Cr ε m ε1
(a) (b)
f C1 fC1
f Cr f Cr

EC EC

ε Cr ε m ε1 ε Cr εm ε 1
(c) (d)

Fig. (4.2) Tension softening idealization (a) linear, (b) bilinear,

(c) trilinear, (d) nonlinear.

4.2.3 Present model:

To model nonlinear concrete response, the constitutive relation contained
in the modified compressive field theory (Vecchio and Collins, 1986)
(Fig. (4.3)) has been adopted. Thus, for concrete in compression, the
relation is:
⎡ ⎛ ε ⎞ ⎛ ε ⎞2 ⎤
σ c = f c ⎢2⎜⎜ c ⎟⎟ − ⎜⎜ c ⎟⎟ ⎥
'
.......(4.2)
⎢⎣ ⎝ ε o ⎠ ⎝ ε o ⎠ ⎥⎦
where
σc and εc are average principal compressive stress and strain in concrete,
respectively; εo is the strain in concrete cylinder at peak stress fc’.
For concrete in tension, prior to cracking, a linear relation is used:
σt = Ec ⋅ εt 0 ≤ ε t ≤ ε cr .......(4.3)

46
where
2f c' f cr
Ec = ε cr = f cr = 0.33 f c'
εo Ec
and
σt and εt are average principal tension stress and strain in concrete,
respectively; Ec is the modulus of elasticity of concrete ( initial tangent
stiffness); fcr and εcr 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 = .......(4.4)
1 + 200ε t

ε
Cupfer 1969
Vecchio and Collins 1986

fcr

Ec

εcr ε

Fig. (4.3) Adopted stress-strain curve for concrete.

47
4.3 Concrete in biaxial state:
Concrete is an anisotropic material, and its properties are significantly
influenced by the presence of confinement. Fig. (4.4) shows the results of
a typical concrete specimen subjected to biaxial compression. It is
observed that both an increase in strength and ductility results due to the
confinement of the orthogonal stress. For the case of tension-compression
(Fig. (4.5)), a substantial reduction in tensile strength results in the
presence of the compressive stresses. Under biaxial tension, the tensile
strength is independent of the stress ratio (Fig. (4.6)).

σ1
1.4
f c'
1.2

1.0

0.8

+σ1 0.6
σ1/σ2
-1/0
+σ2 0.4
-1/-1

0.2 -1/-0.52

-4 -3 -2 -1 0 1 2 3 4
Strain mm/m
Fig. (4.4) Stress-strain relationships of concrete under biaxial
compression (Kupfer et al., 1969).
σ1
1.2 f c'
1.0

0.8

0.6 +σ1
σ1/σ2
-1/0 0.4
+σ2
-1/0.204
-1/0.052 0.2

-3 -2 -1 0 1 2
Strain mm/m
Fig. (4.5) Stress-strain relationships of concrete under combined
tension and compression (Kupfer et al., 1969).
48
 0.10 σ1
f c'
0.08

0.06
+σ1

σ1/σ2 0.04
+σ2
1/0
1/1 0.02
1/0.55

-0.04 0.00 0.04 0.08 0.12

Strain mm/m

Fig. (4.6) Stress-strain relationships of concrete under biaxial

tension (Kupfer et al., 1969).

Experimental investigations of the failure envelope of concrete

under biaxial stress states have been reported by Kupfer and Gerstel
(1973). Their results cover the full range of biaxial stress states and are
considered to be the most reliable data available.
Fig. (4.7) shows the biaxial strength envelope obtained by Kupfer et
al. (1969). A maximum increase in strength of 27 percent over the
uniaxial compressive strength is observed at a stress ratio of σ2/σ1 = 0.5
and, for equal biaxial compression (σ1 = σ2), a 16 percent strength
increase is found, where σ1 and σ2 are the major and minor principal
stresses, respectively.
Failure of concrete occurs by tensile splitting with the fractured
surface orthogonal to the direction of the maximum tensile stress or
strain. Tensile strains are of crucial importance in the failure criterion and
failure mechanism of concrete. Failure modes of biaxially loaded
concrete are shown in Fig. (4.8).

4.3.1 Poisson’s ratio:

Poisson’s ratio for concrete shows some dependency on the stress and the
stress ratios. It increases towards the ultimate load. Kupfer et al. (1969)
obtained a value of 0.2 for biaxial compression, 0.18 for biaxial tension,
and a range from 0.18 to 0.2 for tension-compression. In general, a

49
constant value of Poisson’s ratio is usually assigned, ranging from 0.15
to 0.2.
0.2 σ2/fc’

σ1/fc’
0.0
-1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2
-0.2

-0.4

-0.6

-0.8

-1.0

-1.2

-1.4

Fig. (4.7) Biaxial strength envelope (Kupfer et al., 1969).

4.4 Stress-strain behavior of reinforcing steel:

Stress-strain behavior of reinforcing steel bars used in reinforced concrete
structures is well known. It has been fully investigated since along time
ago. Tensile loading of steel bars is the conventional test for obtaining
typical stress-strain curves, which are assumed to be identical for both
tension and compression. Generally, typical curves are composed of three
distinctive parts; a linear elastic part followed by a yield plateau after
which a strain-hardening part is observed terminating by fracture, Fig.
(4.9).
The essential properties defining the stress-strain curve are:
1- Elastic modulus (Es).
2- Yield strength (fy).
3- Yield plateau.
4- Strain-hardening region.

50
 σ2

σ1

Fig. (4.8) Failure modes of biaxially loaded concrete.

Stress

Strain

Fig. (4.9) Typical stress-strain curves for steel.

51
4.4.1 Review of analytical relations:
The simplest and the most commonly used idealization of the stress-strain
curve is the elastic-perfectly plastic relation which ignores the strain-
hardening region. This model agrees well with low strength or mild steels
used in usual structures and is therefore adopted by the ACI Code 318M-
95, Fig. (4.10a).
Another idealization adopted by several investigators (Hu and
Schnobrich 1990, Soroushian 1991, Al-Imam 1992, Ng et al. 1993,
Elmorsi 1998) is the bilinear model with the second line sloped by a
small percentage of the elastic modulus (Es) either to include the strain-
hardening effect (efficient for high strength steels in which the strain-
hardening starts soon after yielding) or to overcome some numerical
difficulties encountered in some algorithms (the use of tangent stiffness
approach), Fig. (4.10b).
More realistic representation is gained by the use of the trilinear
model (Polak and Vecchio, 1993) combining elastic, yield and strain-
hardening regions, Fig. (4.10c).
More sophisticated models (Santhanam 1979, Krauthammer and
Hall 1982, Marzouk and Chen 1993, Rex and Easterling 2000) have been
proposed but these models were of limited use mainly because they were
not necessary for analyzing usual structures and partly because the
parameters defining the strain-hardening region are not always available
from tests, besides the complexity of the models due to the several stress
states involved, Fig. (4.10d&e).
Belarbi and Hsu (1994), based on their experimental results,
formulated constitutive laws for mild steel reinforcing bars embedded in
concrete. The average stress-strain curve of mild steel bars embedded in
concrete has not show a yield plateau. The apparent yield stress is lower
than the yield stress of a bare bar, Fig. (4.10f).

4.4.2 Estimation of reinforcing steel properties:

Yield strength (fy):
This parameter could be considered as the limit of elastic range. It is quite
distinguished in low strength steels having a definite yield plateau while
it is not so in high strength steels lacking such plateau. It is usually
assumed for the latter case that yield strength is the stress corresponding
to a strain of 0.01 or the stress causing a residual strain of 0.002 after
unloading. However this parameter is usually given by conventional tests.

52
σs σs

Es’
fy fy

Es Es

εy εsu εs εy εsu εs
(a) (b)

σs σs

fy Es’ fy Es’

Es Es

εs εsu εs εy εsu εs

(c) (d)

σs σs

fy fy Es’

Es Es

εy εsu εs εsu εs
(e) (f)

Fig. (4.10) Idealization of stress-strain curve for steel (a) elastic-

perfectly plastic, (b) bilinear, (c) trilinear, (d) & (e) more
sophisticated models, and (f) steel reinforcing bars
embedded in concrete.

53
Elastic modulus (Es):
This parameter is the most important property of reinforcing steel since
steel is elastic in the working load range of reinforced concrete structures.
It has a value ranging between 186,000 – 207,000 MPa but it is usually
considered as 200,000 MPa for practical purposes.

Hardening modulus (Es’):

This parameter is used to introduce the effect of strain-hardening region.
It was usually estimated on an arbitrary way in most studies adopting the
bilinear steel model. It has a value as a percentage of the elastic modulus
(Es) ranging between 0.3 – 2.5%.

Ultimate steel strain (εsu):

It is a function of the steel grade. Actually, it has higher values for low
strength steels. It has a value ranging between 0.12 – 0.2.

4.4.3 Present model:

The bilinear representation of the stress-strain relation of the steel
reinforcement is found to be 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.

4.5 Tension related phenomena in reinforced concrete:

In the presence of reinforcing steel, concrete exhibits more complicated
tensile behavior than that of plain concrete because of concrete-steel
interaction. Two basic phenomena are involved:
1- Tension stiffening; which needs to be clearly distinguished from
tension softening.
2- Bond slip; which needs to be thoroughly studied with all the
capabilities of developing analytical consideration to incorporate its
behavior in the analysis.

4.5.1 Tension stiffening:

4.5.1.1 Definition:
As early as 1899, Considére, in testing small mortar prisms reinforced
with steel wires, observed that their tensile load-deformation response
was almost parallel to the bare bar steel response but remained well above

54
it. In 1908, Mörsch explained that cracked concrete has the ability to
decrease strain in reinforcement due to tensile stresses in the concrete
between the cracks. This phenomenon was later called “tension
stiffening“ (Abrishami and Mitchell, 1996).
It was recognized in cracked reinforced concrete members that the
tensile force carried across cracks by reinforcement is gradually
transferred to concrete by bond on each side of these cracks. The
contribution of concrete between the cracks to the overall member
stiffness is called tension stiffening which is quite significant in
reinforced concrete beams under service loads due to the significant size
of tension zone.
Bond behavior is a key aspect of tension stiffening to transfer tensile
stresses to the concrete. The distributions of steel stress (fs), concrete
stress (fc), and bond stress (u) in a tension specimen are shown in Fig.
(4.11).

Crack

N N

(a) Cross-section
fs
f

(b)
fc

(c)
u′

(d)

Fig. (4.11) Tension specimen, (a) specimen, (b) distribution of steel

stress, (c) of concrete stress, and (d) of bond stress.

55
4.5.1.2 Modeling:
The first studies done on numerical analysis of reinforced concrete
structures assumed concrete to be an elastic-brittle material in tension,
Fig. (4.12a). When cracking occurred, the stress normal to the crack
direction was immediately released and dropped to zero. It was soon
discovered that this procedure leads to great convergence difficulties and
more importantly, to results that strongly depend on the size of the finite
elements used in the analysis.
Tension stiffening depends on many factors such as the bond
characteristics and the tensile strength of concrete, the crack spacing, the
reinforcement bar sizes and arrangements, and the angle between bars
and cracks (Gilbert and Warner, 1978). It can be incorporated into the
computational model in two indirect ways:
a) Assuming that the loss of tensile strength in concrete occurs gradually
after cracking.
b) Modifying the steel-strain curve.
To account for the tension carried by the concrete, Scanlon (1971)
used a stepped response for concrete in tension (Fig. (4.12b)). Lin (1973)
modified Scanlon’s approach and used a gradual unloading model for
concrete in tension (Fig. (4.12c)).
Scanlon’s and Lin’s approaches are generally termed the concrete
referred method because the tension is taken care by lumping the tension
stiffening into the concrete. This approach has the disadvantage that
neither the concrete stress nor the steel stress is exact. Furthermore, it
does not take into account the direction and the location of the
reinforcement in the member relative to the cracks (Chan, 1983).
Another approach used by Gilbert and Warner (1978) is generally
termed the steel referred method (Fig. (4.12d)). The additional tension
carried by the concrete is lumped at the steel level resulting in an added
stiffness of the steel. The stress in the concrete is exact at the crack.
However, the steel stress is magnified to include the tension in the
concrete. Unlike the concrete referred method, this model takes into
account the direction of the steel reinforcement and the location of the
steel reinforcement in the member (Chan, 1983).
Bortolotti (1991) stated that the first cracking load is increased
considerably when reinforcement is embedded on concrete specimens
subjected to direct tension. Based on analytical approach, he suggested a
critical reinforcement ratio where beyond it the reinforced concrete

56
specimens show ductile behavior. However, this statement needs
experimental validation.
Herein, the tension stiffening effect is taken into account as
indicated in Fig. (4.3).

fC1 fC1
f Cr f Cr

EC EC

ε Cr ε1 ε Cr ε m ε1
(a) (b)
f C1
σs
f Cr

EC Es

ε Cr εm ε 1 εs
(c) (d)

Fig. (4.12) Tension stiffening modeling (a) elastic-brittle material in

tension, (b) stepped response for concrete in tension,
(c) gradual unloading model for concrete in tension,
(d) added stiffness of the steel.

4.5.2 Bond and bond slip:

4.5.2.1 Definition:
The interaction between concrete and steel is one of the most important
factors behind the wide spread use of reinforced concrete as a structural
material. Such interaction, or stress transfer, between concrete and steel is
called bond. It is made of the following components (Lutz and Gergely,
1967):

57
1- Chemical adhesion.
2- Friction.
3- Mechanical interaction between concrete and steel.
Bond of plain bars depends mainly on the first two factors.
Deformed bars, however, depend on the first two factors for bond only at
early loading stages beyond which the mechanical interaction becomes
predominant.
As a result of the transfer of stress by the mechanical interaction
relative movement or bond slip takes place. Such slip in deformed bars
could be a result of the following factors (Lutz and Gergely, 1967):
1- The ribs can push the concrete away from the bar.
2- The ribs can crush the concrete.
However, Mirza and Houde (1979) in their study of bond slip
relationship point out that no concrete crushing has taken place after
closely examining bond slip specimens used in their study. They
conclude that internal cracking of the first layer of the concrete
surrounding the bars and bending and / or cracking of the small concrete
teeth near the bar lugs are the real cause of bond slip.
What is important here, however, is the fact that the bond slip
mechanism has a great influence on the finite element solution due to its
major influence on the tensile cracks width and spacing, and on the
distribution of concrete stresses in partially cracked members.
Bond stresses are generally caused to develop in reinforced concrete
members due to either general behavior represented by flexural variation
(the presence of shear forces), or local behavior between adjacent cracks.

4.5.2.2 Bond stress-slip relation:

It has been verified experimentally that the bond stress is a basic function
of the relative displacement (slip) of steel to concrete. This is called the
bond stress-slip relationship. Although there were attempts (Russo et al.,
1990) to develop it theoretically, the available relationships are obtained
from pullout tests or by beam tests showing a nonlinear bond behavior of
deformed steel bars. Constants, linear, bilinear, or various nonlinear
approaches have been developed (Balázs, 1993).
The pullout test was a good tool utilized by several investigators to
account for bond slip relation. Lutz (1970) provided a liner relation from
such tests. Edwards and Yannopoulos (1978) presented many results of
such tests also. Bond slip relations derived from these studies are found to

58
vary significantly from these derived by direct tension studies, due to the
difference in the testing procedure resulting in highly localized slips for
the pullout tests and low averaged slips along the bars for the direct tests,
which seems more realistic to be applied to reinforced concrete members
having well developed bars.
The bond stress-slip curves may be divided into two parts, a linear
part in which the bond is developed through chemical adhesion and
friction between concrete and steel, the maximum bond stress developed
by this mechanism being in the range (1 – 2 MPa), and a nonlinear part
which the bond is developed mainly through mechanical interlock
between steel bar ribs and surrounding concrete (Abdel-Halim and
Alostaz, 1998).
Two types of bond failure was seen (Hertz, 1982):
(i) For large concrete section or confined concrete, crushing of concrete
in front of steel ribs will yield pullout failure. Thus, concrete
compressive strength is the major factor governing such type of
failure.
(ii) For thin section (i.e., with small concrete cover), concrete between
transverse cracks will be exerted to compressive stresses causing
tensile strains in the radial direction, subsequently, splitting crack will
develop in the bar direction depending on tensile capacity of concrete.
Hence, confinement of concrete may inhibit this type of failure and
bar pullout will occur (Soroushian et al., 1991).
Essa (1990) conducted three types of tests (concentric pullout test,
tension specimen test, and tension (splice) pullout test) to evaluate the
bond strength in certain types of specimens and to compare their bond
performance between those cast in hot weather conditions and those cast
in temperate weather condition. He concluded that hot ambient decrease
the bond strength of medium water-cement ratio mixes and increases
moderately the bond strength of high water-cement ratio mixes.
Studying the effect of concrete compressive strength on bond
strength, Azizinamini et al. (1993) test four pairs of beam specimens. For
each pair of test specimens, all variables were kept constant except
concrete compressive strength. The results indicate that normalized bond
strength (u / f c' ) decreases as concrete compressive strength increases.
In general, the bond behavior, of the main reinforcements in
reinforced concrete beams, improves with the increase in the embedded
length and the bottom cover. However, the test results (Abdel-Halim and

59
Alostaz, 1998) indicate that there are certain embedded length and bottom
concrete cover beyond which there is no significant improvement of bond
behavior.
For a deformed reinforced bar, deformation pattern has virtually no
effect on bond strength when a splitting failure of the concrete governs
(Darwin et al. 1992, Darwin and Graham 1993). It behaves as a wedge
due to the nature of the loading it imposes on the concrete, causing the
concrete to split. The wedging action is not sensitive to the details of the
deformation pattern. Under condition of increased confinement (as in a
standard pullout test or with the addition of transverse reinforcement or
higher concrete cover and bar spacing), the greater the rib bearing area,
the higher the bond strength. Thus, with additional confinement provided
by either transverse reinforcement or additional concrete, bond strength
increases significantly with increases in the relative rib area (Darwin and
Graham, 1993).
If a bar embedded in a large concrete piece, due to concrete
shrinkage, restrained by the steel, some self-stresses appear around the
bar. In the plane perpendicular to the bar axis, the radial stress is a
compressive stress, while the orthoradial one is a tensile stress. As the
steel bar diameter increase, the test results show that the compressive
confining stresses decrease while the tensile stresses increase. Thus, bond
between steel and concrete decreases with the increase in the bar diameter
(De Larrard et al., 1993).
Based on the experimental results presented by Darwin and Graham
(1993), a bilinear relationship between bond stress and bond slip is
adopted, Fig. (4.13). The mathematical model for the secant stiffness of
bond-slip is:

k s = 25 Δ s Δ s ≤ 0.2mm
5
ks = Δ s ≥ 0.2mm .......(4.5)
Δs

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

60
 6

5
Bond stress (MPa)
4

3 ks

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

Fig. (4.13) Adopted bond stress-bond slip relationship.

4.6 Shear transfer across the crack:

The assumption that cracked concrete can not transfer shear forces across
the crack interface is not realistic. Experiments show that a considerable
amount of shear stress can be transferred across the rough surfaces of
cracked concrete. In plain concrete, the main shear transfer mechanism is
aggregate interlock and the main variables involved are the aggregate size
and grading. In reinforced concrete, dowel action will play a significant
role, the main variables are the reinforcement ratio, the size of the bars
and the angle between crack and bars (Cervera et al., 1987).

4.6.1 Aggregate interlock:

Several models have been proposed to explain or predict the aggregate
interlock behavior. One model (Laible et al., 1977) distinguishes between
interlock due to “local roughness“ and to “global roughness“ of the crack
face. It is postulated that local roughness causes interlocking of the fine
aggregate particles, principally a bearing or crushing action, and that
global roughness causes interlocking of the coarse aggregate particles,
principally a sliding and overriding action. The predominant effect is
thought to be local roughness in specimens with initial cracks less than
0.25 mm and global roughness in specimens with wider initial cracks
(Millard and Johnson, 1984).
An alternative model proposes that aggregate interlock be entirely
due to the frictional sliding of two rigid surfaces. These surfaces have

61
been represented by a saw-tooth shape (Jimenez et al., 1978) and by a
series of parabolic segments (Fardis, 1979).
A more recent model (Walraven, 1981) suggests that concrete is a
two-phase material of aggregate and cement matrix, which can be
modeled as a distribution of rigid spheres of a range of sizes embedded to
various depths within a deformable rigid-plastic matrix. Shear forces and
direct compression forces are obtained from equilibrium when a given
shear displacement and crack widening occurs. Hence, for a known direct
stiffness restraining crack widening, the crack displacement path and the
shear stiffness can be obtained. The complex probabilistic expression
derived to predict the chances of finding a particular sized aggregate
particle at a particular embedded depth is replaced by a simpler bilinear
expression.
In this model, shear forces are resisted by a combination of crushing
and sliding of the rigid spheres into and over the softer cement matrix;
contact and interaction between spheres projection from opposite crack
faces is not considered.
Millard and Johnson (1984) devised new type tests to study
interlocking of the aggregate particles protruding from each face of a
crack. They concluded that the aggregate mechanism results from a
combination of crushing and overriding of the crack faces and can be
predicted if the normal stiffness that restrains crack widening is known.

4.6.2 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 to be rigid, the first two mechanisms

62
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 the principal action. This has been
modeled (Millard, 1983) by considering the dowel bar as a beam on
elastic foundation.
Test of a new type has been devised (Millard and Johnson, 1984) to
study the dowel action mechanism and the results were compared with
several theoretical models. The initial shear stiffness and ultimate
strength due to dowel action were successfully predicted from theoretical
models and an exponential curve was used to describe the intermediate
behavior.
In reviewing studies on dowel behavior, distinction should be made
between the behavior of dowel bars in action against concrete core and
against concrete cover. In the beam shown in Fig. (4.14), the left dowel
bars are pushing against core and the right ones against cover. In action
against core (Soroushian et al., 1986), the dowel bar behaves like a beam
on elastic foundation. The ultimate dowel load is reached when concrete
underneath the bar split under bearing stresses. In the case of dowel bars
acting against concrete cover (Soroushian et al., 1987), the bar initially
behaves like a beam on elastic foundation, but the behavior changes when
cover splits from the core. After cover spalling, the bar acts like a beam
supported on the closest stirrup to the crack. Depending on the size and
yield strength of the dowel bar, tensile strength of concrete, and location
and yield strength of the closest stirrup, the ultimate load may be reached
by one or a combination of the following action: split cracking of cover,
flexural yielding of dowel bar, or tensile yielding of stirrup.

(a) (b)

Fig. (4.14) Dowel action (a) against concrete core, (b) against
concrete cover.

63
 Cyclic test results (Soroushian et al., 1988) indicate that the stiffness
and energy dissipation capacity of dowel bars deteriorate severely with
repetition of inelastic load cycles. Degradation of strength was, however,
observed to be negligible especially in larger dowel bars. It was also
observed that the interface crack width grows progressively under
inelastic load cycles.
Tests (Poli et al., 1993) on block-type specimens, reinforced with a
single, long dowel close to concrete surface and acting against either
concrete core or cover resulted in force-displacement curves,
displacement and curvature diagrams along the axis of the bar and initial
stiffness verses bar diameter.
The experimental data presented by Poli et al. (1993) will be utilized
to proposed the following simplified mathematicl model for the secant
stiffness of dowel action against core (kd):

20 + 5(d b −14)
kd = Δ Δ ≤ 1.5 mm
1.5
20 + 5(d b − 14)
kd = Δ ≥ 1.5 mm .......(4.6)
Δ

where db is diameter of the bar; and Δ is dowel displacement. These

equations are shown in Fig. (4.15).

100
Experimental
Proposed
80
Shear force ( kN )

db=24 mm
60

40 db=18 mm

20 db=14 mm

0
0 1 2 3 4 5 6
Dowel displacement ( mm )

Fig. (4.15) Proposed dowel stiffness.

64
4.6.3 Concrete strength parallel to crack:
After cracking has taken place, the concrete parallel to crack direction is
still capable of resisting either tensile or compressive stresses. When it is
subject to tension, a pure linear elastic behavior is assumed (Hu and
Schnobrich, 1990) and a second crack normal to the first one may be
found. In other hand, when it is subjected to compression, experimental
results show that the tensile cracks have caused a degrading effect on the
compressive strength. Consequently, the concrete will be weaker and a
modification is needed on the compressive strength parallel to the crack.
Several formulas have been proposed to determine the degraded
compressive strength (Cervenka 1985, Vecchio and Collins 1986,
Vecchio 1989, Kollegger and Mehlhorn 1990). Generally, tests at the
university of Toronto (Vecchio and Collins, 1986) lead to relate the
reduction in (fc’) of cracked concrete to the transverse tensile strain. Other
tests at the University of Kassel (Kollegger and Mehlhorn, 1990) yield a
conclusion that this reduction is function of the transverse tensile stress
and do not exceed 20%.

4.7 Friction slip and separation:

In current design practice, the surface between two concretes 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 a cohesion and a coefficient of
friction. In the ACI Code, 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 surface intentionally
roughened, μ=0.6 for concrete placed against hardened concrete not
intentionally roughened.
Few interface crack models have been proposed in the literature.
There are the physical models, based on physical idealization of the crack
surfaces, and the empirical models based on fitting processes of
experimental results. The two-phase model by Walraven (1981) is an
example of a physical model. It has been extended to reversed cyclic
loading by Walraven (1994). That type of model gives a better
understanding of the mechanisms involved at the crack interface.
However, the model is quite complex and is applicable to cracks only.

65
The Tassion and Vintzèleou (1987) model is an example of an empirical
model. It covered two types of interfaces, the rough interface and the
smooth interface, for normal stresses 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 (μb) and roughness friction
coefficient (μi):

λ dμ b + χiμi
μ= .......(4.7)
1 − λ d χiμ bμi

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

μb = 0.865-0.050 σn for 0.5 ≤ σ n ≤ 2.0MPa

μi is defined by the equations in Table (4.1). Two correction factors were

introduced: (1) λd, the dynamic reduction factor equal to 1.00 for static
loading and 0.85 for dynamic loading; and (2) χi, 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.

Table (4.1) Concrete interface model roughness coefficient

(Fronteddu et al., 1998)
Interface type σn (MPa) Peak μip
σ n ≤ 0.4 0.90-1.367 σn
homogeneous 0.4 ≤ σ n ≤ 1.5 0.40-0.1167 σn
1.5 ≤ σ n ≤ 2 0.30-0.050 σn
σ n ≤ 0.275 0.875-1.75 σn
Water-blast 0.275 ≤ σ n ≤ 1.2 0.44-0.185 σn
1.2 ≤ σ n ≤ 2 0.25-0.0375 σn
σ n ≤ 1. 0 0.15-0.15 σn
Untreated
1.0 ≤ σ n ≤ 2.0 0.05-0.005 σn

66
 Based on the experimental results presented by Fronteddu et al.
(1998), a bilinear relationship between shearing stress and slip is adopted,
Fig. (4.13).

Shearing stress (MPa)

0.02
Slip (mm)

Fig. (4.13) Adopted shearing stress-slip relationship.

67
 CHAPTER FIVE
FINITE ELEMENT IDEALIZATION

5.1 Introduction:
During recent years, interest in nonlinear analysis of concrete structures
has increased steadily, because of the wide use of plain, reinforced, and
prestressed concrete as a structural material, and because of the
development of relatively powerful finite element procedures (Bathe,
1996). If a realistic nonlinear analysis of a concrete structure can be
carried out, the safety of the structure is increased and the cost can
frequently be reduced.
There are number of factors that have prevented the wider
acceptability of nonlinear finite element analysis procedures in the
analysis of concrete structures. A first important consideration is that the
constitutive properties of concrete have not as yet been identified
completely, and there is still no generally accepted material law available
to model concrete behavior. A second important factor is that nonlinear
finite element analysis of concrete structures can be very costly and may
require considerable user expertise. The high cost of nonlinear analysis of
concrete structures is largely due to the difficulties encountered in the
stability and accuracy of the solutions. These difficulties, however, are a
direct consequence of the specific numerical implementation of the
concrete nonlinearities.
A completely rational analysis of reinforced concrete structures for
internal stresses and displacements is made difficult by several factors.
These include (a) the nonhomogeneus nature of the construction, (b) the
nonlinear response of the material to load, (c) relative slip which occurs
between concrete and reinforcement, (d) progressive destruction of bond
in local areas, and (e) the profound influence of progressive cracking. The
member to be studied frequently is defined by a highly irregular boundary
as a result of cracking. The applied load may be concentrated, distributed
uniformly or nonuniformly, or may be applied in the form of bond stress
distributed in a complex way along the steel-concrete interface. In spite of
these complications, a refined analysis is necessary to understand such
phenomena as shear, diagonal tension, and other complex failure modes.
The analysis of reinforced concrete members may be approached on
a one, two, or three-dimensional basis. For members which are long
relative to their depth and width, and which remain uncracked, the
68
significant stresses are mostly in the direction parallel to the axis except
in the vicinity of concentrated loads or reactions. The usual one-
dimensional analysis is satisfactory. After cracking, however, even
members of ordinary proportion are composed of sub-elements, which
may have length and depth of the same order. A realistic analysis must
recognize that the state of stress is at least two-dimensional. Although in
many cases the width of the element to be analyzed is of the same order
as the length and depth, the reinforcement is often more or less uniformly
distributed across that width. Loads and reactions often act only in the
vertical plane. In such circumstances, stresses in the third principal
direction are nearly zero, and analysis for two-dimensional stress is
satisfactory.
Classical methods of analysis cannot cope with the complications
described. Application of the methods of two-dimensional elasticity and
plasticity is limited to cases so idealized as to bear little relation to the
member under study. Simplified representations such as the truss analogy
and the tooth analogy, sometimes used for studying concrete beam
behavior, cannot possibly represent true conditions.
The development of the finite element method for analysis of
structural continua offers a powerful tool to be used in understanding the
behavior of reinforced concrete. Originally developed in the aircraft
industry, it has been applied to the analysis of civil engineering structures
for the last four decades. Details of the finite element method can be
found in a number of textbooks (Bathe 1996, Chandrupatla and
Belegundu 1997).
Ngo and Scordelis (1967) suggested finite element analysis of
reinforced concrete structures. They studied simple beams using two-
dimensional triangular finite elements along with special bond-link
elements to give steel-concrete interaction. Linear analysis was made on
beams with predefined crack patterns.

5.2 Finite element idealization of reinforced concrete

members:
The finite element idealization of reinforced concrete members should be
able to represent concrete cracking, the interaction between concrete and
reinforcement, and the capability of concrete to transfer shear after
cracking by aggregate interlock.
To create such an idealization, the following element types are used:

69
1- Plane stress elements to represent concrete.
2- Line elements to represent reinforcement.
3- Linkage elements to represent the bond between concrete and
reinforcement and the dowel action.
4- Interface elements to represent shear-transfer.

5.3 Representation of concrete:

Since the finite element analysis of three-dimensional structural problems
is costly and time consuming, it is preferable to approximate the original
three-dimensional problems with equivalent two-dimensional once when
possible.
Many important classes of structures, such as panels, walls, beams,
slabs and shells, can be approximated as being in a state of plane stress.
A structural problem in which the member thickness is much smaller
than its length or span, and is loaded in its plane is called plane stress
problem. In such problems, the stress state is defined in terms of two
normal stress components (σx and σy) and one shearing component (τxy).
These stress components are averaged over the thickness and assumed to
be independent of the member’s depth, while all other stress components
are assumed to be zero.
There are many types of plane stress elements. One of the widely
used elements is the three-noded plane constant strain triangular element.
The widespread use of this element may be attributed to its formulation
simplicity, its capability to fit structures with irregular geometry, as well
as its early use in the finite element method. This element, however, has
many limitations. The most important drawback is that its strain, and
hence stress, is assumed to be constant in the whole element.
Better results can be obtained when using four-noded isoparametric
quadrilateral elements even with less nodal degrees of freedom; hence
less computational effort that when using the constant strain triangular
elements. This is due to the capability of its interpolation function to
describe a linear variation of strain. However, an important drawback of
this element is its poor behavior under pure bending.
Generally, there are two approaches to improve the behavior of such
elements; either by improving the interpolation function of the four-
noded isoparametric elements, or by using higher order elements with
more nodes. The first approach is simple and result in elements with very
good bending characteristics (Cook 1974, Cook 1974, Cook 1977), while

70
the second approach is costly, for it requires a great deal of computational
effort.
Some investigators (Ngo and Scordelis 1967, Nilson 1968, Cedolin
and Nilson 1978) used the constant strain triangular element in their
analysis of two-dimensional reinforced concrete problems by finite
element method. Many of them (Bergan and Holand 1979, Bathe et al
1989, Vecchio 1989, Ayoub and Filippou 1998) used the four-noded
isoparametric element. Others preferred the improved four-noded
isoparametric element (Al-Sabah, 1983) or the higher order elements
(Bathe et al. 1989, Gajer and Dux 1989).

5.3.1 Four-node isoparametric quadrilateral element:

For the four-node quadrilateral element (Fig. (5.1)) in the context of plane
elasticity, the element stiffness matrix can be calculated by integrals of
the form:

[ K ] = ∫ [ B]T [ D][ B]d ( vol.) .......(5.1)

vol .

where [B] and [D] represent the strain-displacement and stress-strain

matrices, respectively.

Y y

y
2
1

x θ

θ
X x

(a) (b)

Fig. (5.1) Four-node quadrilateral element (a) element in

global coordinates, (b) cracked element.

If the element is rectangular with its sides parallel to the x and y

axes, the term under integral consists of simple polynomial terms which
can be easily integrated in closed form by separation of variables
resulting in compact term. In general, however, quadrilateral elements
71
will lead to very complicated expressions under the integral sign, which
can only be tackled numerically.
Noting that “two-point” Gaussian quadrature (total number of
integration points in one element equal four) leads to an exact solution for
the stiffness matrix of a four-node quadrilateral, a compromise approach
is to evaluate the contribution to stiffness for each of the “Gauss points”
algebraically and add them together (Smith, 1998), thus:

2 2
[ K ] = ∑∑ w i w j (det J ) ij [ B]Tij [ D]ij [ B]ij .......(5.2)
i =1 j=1

where J is the Jacobian matrix relating the natural coordinate derivatives

to the local coordinate derivatives, i and j specify the location of the
integration points, while wi and wj are the corresponding weight factors
(wi = wj = 2).
One advantage of the numerical integration scheme is that the
material stress-strain matrix [D] is evaluated at each integration point;
hence, different material properties can be described within an element.
This feature is very helpful in defining partial cracking or crushing in an
element.

5.3.2 Modeling of cracked concrete:

For nonlinear finite element analysis of concrete structures two main
approaches, namely the discrete crack model and the smeared crack
model, can be used for modeling cracking, coupled either with a strength
criterion or with linear elastic fracture mechanics.
In the discrete crack model, cracking is assumed to occur as soon as
the nodal force normal to the element boundaries exceeds the maximum
tensile force that can be sustained and continuous remeshing is required.
In the smeared crack model, a cracked solid is imagined to be a
continuum where the notions of stress and strain remain valid. The
behavior of cracked concrete can then be described in terms of stress-
strain relations and upon cracking it is sufficient to replace the initial
isotropic stress-strain relation by an orthotropic stress-strain relation. This
implies that the topology of the original finite element mesh remains
preserved. This approach leads to a straightforward computer
implementation in two- as well as three-dimensional situations, and it is
for this reason that the method has come into widespread use.
With nonlinear fracture mechanics, taking the process zone at the
crack tip into account, the range of validity of both approaches can be

72
significantly extended, leading to the fictitious crack model and to the
crack band model. Based on the crack band model, two crack models can
be distinguished, the fixed crack model and the rotating crack model. In
both models a crack is initiated when the maximum principal stress
violates the tensile strength and the initial orientation of the crack is
normal to the maximum principal strain. In the fixed crack model the
crack plane is fixed during the total analysis process. On the other hand,
the rotating crack model allows the crack plane to rotate. If the axes of
principal stress do not change during the total analysis process, there will
be no difference between the two models, however if the axes change
after the crack is initiated, the responses of both models will be different.
From the physical point of view the rotating crack model seems
more reasonable than the fixed crack model. First, in the fixed crack
model a shear retention factor has to be chosen to decide the shear
stiffness, which sometimes is quite delicate. Second, the shear strain may
arise along the crack plane, which will lead to the shear stress over the
crack plane, consequently, the residual normal stress may exceed the
tensile strength in a direction inclined to the existing crack plane. These
two inconveniences do not exist in the rotating crack model, in which the
tangential shear stiffness automatically arises from the requirement of
coaxiality between principal stress and principal strain, and then the
possibility that normal stress violates the crack criterion in a direction
inclined to the crack plane is removed. Among existing concrete crack
simulation models the rotating crack model is certainly one which lends
itself to a relatively simple implementation; it is therefore of interest to
evaluate its simulation capabilities. In this work the rotating crack model
is adopting in representing cracked concrete.

5.4 Modeling of the reinforcement:

The major nonlinearities in the reinforced concrete structure are from the
effect of cracking. Cracks are initiated even at a low service load and the
redistribution of the stresses in the concrete to the surrounding
reinforcement necessitates that the reinforcement be modeled correctly. In
two-dimensional problems, three methods used to represent the
reinforcements. These methods are the discrete, the embedded, and the
distributed models.

5.4.1 Discrete reinforcement:

The reinforcement is represented by discrete elements which may be
either one-dimensional (bar or beam) or plane stress elements (Fig. (5.2)).

73
 Herein, the use of discrete element model permits to represent bond
slip between concrete and reinforcement. This model is proved to be
helpful to study the effect of lumped or discrete reinforcement with a
certain layout on the behavior of reinforced concrete structures. Also it is
suitable for discrete crack representing.
In this method, the stiffness of the steel element is calculated
according to the type of the element, and then added to the global
stiffness of the structure.

(a) (b)

Fig. (5.2) Discrete representing of reinforcement (a) one dimensional (bar

or beam) elements, (b) plane stress (constant strain triangle)
elements.

5.4.2 Embedded reinforcement:

In this case, discrete reinforcing bars are assumed to be embedded within
the concrete element, see Fig. (5.3a). Perfect bond is assumed to exist
between the reinforcement and the surrounding concrete. For straight
reinforcement, the coordinates of the two ends of the reinforcement have
to be specified. As the reinforcement is curved in plane (for two-
dimensional problems), more coordinate points along the reinforcement
have to be defined.
Complete compatibility between concrete and reinforcement is
assumed, so that the deformation is continuous over the adjoining
boundaries of concrete and steel. The stiffness and internal forces
associated with the reinforcement are integrated and added to those of the
concrete to get the total stiffness and internal forces of the element.
The computation of the added stiffness of a single reinforcement is
about the same as the computation of the concrete element stiffness itself.
As the number of reinforcing bars within an element is increased, the

74
computational effort increases rapidly until the solution is
computationally prohibitive.

5.4.3 Distributed reinforcement:

In this case, the amount of reinforcement and their directions are lumped
at the integration points, Fig. (5.3b). Perfect bond between the concrete
and the steel is assumed.
This approach is computationally efficient and is best in modeling
closely spaced reinforcing bars. It does not model the discrete
reinforcement exactly, however. On the other hand, when the distributed
crack method is used to represent cracking in the structure, this method is
a more consistent approach.

(a) (b)

Fig. (5.3) Modeling of reinforcement (a) embedded reinforcement,

(b) distributed reinforcement.

5.4.4 Modeling of the reinforcement in the present study:

The discrete and the distributed methods of modeling reinforcement are
used in this study:
1- A two-noded bar (truss) element with two degrees of freedom per node
is used to represent the reinforcement (Fig. (5.4)).
The element stiffness matrix in global coordinates [Kb]G is:

⎡ c2 sc − c 2 − sc⎤
⎢ ⎥
s2 − sc − s2 ⎥
[K ] =
EbAb ⎢
.......(5.3)
Lb ⎢ sc⎥
b G
c2
⎢ ⎥
⎣ s2 ⎦

75
 where
Eb is the secant modulus for the steel reinforcement; Ab and Lb are the
area and length of the steel bar element, respectively; c = cosθ, s =
sinθ, θ is the angle between global and local coordinate of the bar.
2- The integration points are chosen to be the controlling point.
Uniaxial stress-state is assumed in each reinforced direction.
Fig. (5.5) shows a typical plane stress element. At the integration
point P, the element is reinforced in both x’ and y’ directions.
The incremental constitutive matrix for the reinforcement [Ds]L in
the material coordinate x’ and y’, can be written as:

⎡ρ x E sx 0 0 ⎤
⎢ ⎥
'

[D ]
s L =⎢ 0 ρ y E sy ' 0 ⎥ .......(5.4)
⎢ 0 0 0 ⎥⎦
⎣

where ρx and ρy are the reinforcement ratio in x’ and y’ directions,

respectively; Esx’ and Esy’ are the secant modulus for the reinforcements in
x’ and y’ directions, respectively. The modulus Esx’ is determined for a
particular stress-strain state using the relationship:
f
E sx = sx
'
.......(5.5)
εsx
'

'

where fsx’ and εsx’ is the average stress and strain in the reinforcement,
respectively. Esy’ is determined in similar manner.
Transformation must be applied to convert the reinforcement
stiffness to the global system:

[D ]
s G
= [Ts ] [D s ]L [Ts ]
T
.......(5.6)

where [Ds]G is the reinforcement stiffness matrix in global coordinate;

[Ts] is the transformation matrix with θ, θ = θs + θc , θ is the angle
between global and local coordinate of the element; θs is the angle
between the x’ direction of the steel reinforcement and the local x-axis of
the concrete element; and θc is the angle between global and local
coordinate of concrete element.

76
 vj
x′

uj
y′
j
vi

i ui

Fig. (5.4) A two-dimensional truss element.

x’
y’
θ

Fig. (5.5) Plane stress element with distributed reinforcement

lumped at the integration points.

5.5 Linkage element:

In order to represent the bond slip phenomenon, linkage elements are
used to link the concrete and the reinforcement nodes that occupy the
same location.
Linkage element was first developed in a two-dimensional form by
Bresler and Scordelis (1964), used first in finite element analysis of
reinforced concrete by Ngo and Scordelis (1967), and then extended to
three-dimensional form by Ahmad and Bangash (1987). It is currently
used not only to represent the bond slip but also the aggregate interlock

77
and the dowel action in discrete crack representation, and will be used, in
this research, to connect between concrete and concrete elements. A
linkage element may be thought as being composed of two orthogonal
springs each with a given stiffness depending on the phenomenon they
describe. The linkage element has no physical dimensions at all, and only
its mechanical properties are of importance. Fig. (5.6) shows such an
element oriented at an arbitrary angle (θ) relative to the global coordinate
system.

y’ x’

θ
j

Fig. (5.6) Linkage element.

To incorporate the linkage element into the finite element computer

program, it is necessary to develop the stiffness matrix of the linkage
element. Let the springs in the local horizontal and vertical directions
have stiffness Kh and Kv, respectively. The stress-strain relationship is
given by Ngo and Scordelis (1967):

⎧σ h ⎫ ⎡ K h 0 ⎤ ⎧ε h ⎫
⎨ ⎬=⎢ ⎨ ⎬ .......(5.7)
⎩σ v ⎭ ⎣ 0 K v ⎥⎦ ⎩ε v ⎭
or

{σ} = [C]{ε}

78
where σh and σv are relative forces; εh and εv are relative displacements
between points (i) and (j) in the local horizontal and vertical directions
and are positive when they are tension. The strains and the global
displacements are related by the displacement transformation matrix [A]:

{ε} = [A]{u} .......(5.8)

or
⎧ui ⎫
⎧ε h ⎫ ⎡ − c − s c s⎤ ⎪⎪ v i ⎪⎪
⎨ ⎬ ⎢ = ⎥⎨ ⎬ .......(5.9)
⎩ε v ⎭ ⎣ s − c − s c ⎦ ⎪ u j ⎪
⎪⎩ v j ⎪⎭
where c = cosθ and s = sinθ.
By noting that the force transformation matrix [B] is equal to the
transpose of the displacement transformation matrix [A], the stiffness of
the linkage element can be obtained from:

[K] = [A]T [C] [A] .......(5.10)

⎡ k11 k12 − k11 − k12 ⎤
⎢ k 22 − k12 − k 22 ⎥
=⎢ ⎥ .......(5.11)
⎢ k11 k12 ⎥
⎢ k 22 ⎥⎦
⎣
where k11 = Kh cos2θ + Kv sin2θ
k12 = ( Kh – Kv ) cosθ sinθ
k22 = Kh sin2θ + Kv cos2θ

In the present analysis, the stiffness of the horizontal spring (Kh) is

given in Chapter (3) to represent the bond-slip relationship, while the
vertical spring has a very high stiffness value to represent the vertical
movement of the reinforcement relative to the concrete.

5.6 Interface element:

The behavior of a composite beam depends upon the interaction between
the precast beam and the cast-in-situ slab. There can be separation,
closing of gap, and slipping between the beam and slab. A four-noded
interface element as shown in Fig. (5.7) has been used to model this
behavior between concrete elements of the two parts. Two in-plane

79
translational degrees of freedom per node have been considered. The
displacement vector is:
{δ} = [u v]T .......(5.12)
The strains are the relative displacements at the top and bottom of
the element. The strain vector is defined as:
{ε} = [Δu Δv]T .......(5.13)
The relevant stress vector is:
{σ} = [σ u σ v ]T .......(5.14)
The material modulus matrix is defined as:

[D] = ⎡⎢ s
k 0⎤
⎥ .......(5.15)
⎣ 0 k n⎦

Where ks and kn are the shear and the normal stiffness coefficients,
respectively. The values of stiffness coefficients can be obtained from
Chapter (4).
The strain matrix is defined as:
[B] = [− [I]N1 − [I]N 2 [I]N1 [I]N 2 ] .......(5.16)
where [I] is identity matrix of order (2x2), and Ni are the shape functions.
The stiffness matrix is calculated as:
[K L ] = ∫ [B] ⋅ [D] ⋅ [B] ⋅ dx
T
.......(5.17)
Two Gauss points have been used to calculate the stiffness matrix.
The stiffness matrix in the global coordinate system has been calculated
as:
[K G ] = [T ]T ⋅ [K L ] ⋅ [T ] .......(5.18)
where [T] is the transformation matrix.
Y

y x
l
ζ
j
ζ = −1 ζ=1
k

i θ
X

Fig. (5.7) Interface element.

80
5.7 Nonlinear analysis:
In general, structures exhibit nonlinear behavior if either of the following
conditions are violated: (i) linear stress-strain relationship, and (ii) linear
strain-displacement relationship. The nonlinear behavior caused by the
violation of the first condition or the second condition is called material
nonlinearity or geometrical nonlinearity respectively.
Both behaviors of nonlinearity can exist in structural problems.
However, the effect of the geometrical nonlinearity is of minimal effect in
reinforced concrete structures, which are relatively bulky. Hence, only the
effect of the material nonlinearity is considered in the present analysis.
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.
In the stiffness formulation, the nonlinear stress-strain dependent
equilibrium equation can be written as:

[K(σ,ε)] {U} = {F} .......(5.19)

where [K(σ,ε)] is the stress-strain dependent global stiffness matrix, {U}

is the unknown displacement vector, while {F} is the external load
vector.
Numerous iteration methods for solving systems of nonlinear
equations had been developed. Three commonly used methods are initial
stiffness method, tangent stiffness method, and secant stiffness method.
All three methods use the same initial stiffness for the first iteration in the
first load increment. From the second iteration on, the initial stiffness,
tangent stiffness, secant stiffness is used respectively by the three
methods. The initial stiffness method has a distinct advantage of using the
same stiffness for all iterations, and thus requires less computational
efforts, while the tangent stiffness method has the fastest convergence and
the secant stiffness method can reproduce the softening behavior.

81
 For reinforced concrete, the behavior is path dependent or tracing the
loading path is desired. Thus, an incremental-iterative technique and the
secant stiffness approach are utilized as a solution algorithm for the
present study. The system of the secant equilibrium equations obtained at
each iteration is solved for nodal displacements using the Gauss
elimination solver for symmetric banded equations.
The predominant stress-state of many reinforced concrete structures,
such as panels, beams, and shear walls, is that of plane stress. Many
constitutive models have been proposed for the nonlinear finite element
analysis of reinforced concrete structures under plane stress conditions.
These can be classified into orthotropic models, nonlinear elastic models,
plasticity models, endochronic models, fracture mechanics models, and
micromodels such as the microplane and nonlocal continuum model. Few
of the models, however, have been used to simulate the nonlinear load-
displacement behavior of different types of structural elements under
various stress combinations to permit an assessment of their general
applicability.
Among concrete constitutive relations orthotropic models strike a
balance between accuracy and economy. These models are generally
based on the concept of “equivalent uniaxial strain” by Darwin and
Pecknold (1977). The models proposed to date differ in the description of
the biaxial failure envelope the uniaxial equivalent stress-strain relation,
Poisson’s ratio, and the tension-compression behavior. From an extensive
experiment study by Vecchio and Collins (1986), Vecchio (1990, 1992)
developed an orthotropic concrete model with compressive strength
degradation as a function of tensile strain after cracking. In another model
by Balakrishnan and Murray (1988 a,b,c) the degradation in compressive
strength is a function of tensile stress at cracking.
Ayoub and Filippou (1998) proposed a model that is an extension of
the orthotropic models by Vecchio, and Balakrishnan and Murray.
Herein, a brief description of the salient features of the model (which has
been adopted in this thesis) and its numerical implementation.

5.8 Model description:

The described concrete material model is an orthotropic model with
equivalent uniaxial stress-strain relation in the axes of orthotropy. These
axes coincide with the principal axes of total strain and are continuously
updated during the analysis. This approach, known as the rotating crack

82
model, yields better agreement between analytical and experimental
failure loads for shear panels and deep beams. In the alternative approach
of the “fixed crack“ concept, the axes of orthotropy are fixed once a crack
has formed in one direction. This approach yields analytical failure loads
that overestimate experimental data by a large margin (Crisfield and
Wills, 1989).
In the proposed model, the unloading of elements in softening
regions of the finite element mesh takes place elastically along the
original loading path. While this is grossly inaccurate from the material
response standpoint, it does not affect significantly the monotonic
behavior of well-reinforced concrete structures for the following reasons:
(1) in reinforced concrete structures with commonly used amount of
reinforcement, the softening of concrete is balanced by a stress increase
in the reinforcement; and (2) tension softening involves small strains and
thus does not have significant influence on global deformations.
Under a biaxial stress state the strength in one direction of
orthotropy is a function of the magnitude and sign of the stress in the
orthogonal direction. A complete description of the model hinges on the
definition of a biaxial strength envelope, which relates the concrete
strength in each principal direction to the current principal stress ratio.
Concrete cracking is an important factor in the behavior of
reinforced concrete members. In plain or lightly reinforced concrete
members, a single crack propagates and dominates the global response. In
heavily reinforced members, stresses can be transmitted across cracks
through bonded reinforcing steel and several cracks form in the specimen
before failure. As cracking progresses, the load-displacement response of
the member softens until it attains the ultimate strength.
In summary, the ingredients of a complete orthotropic concrete
material model are: (1) a biaxial strength envelope; (2) a uniaxial stress
strain relation in the direction of orthotropy; (3) a crack model; (4) the
definition of the material stiffness matrix; and (5) a consistent numerical
determination strategy.

5.8.1 Biaxial strength envelope:

5.8.1.1 Biaxial compression region:
The proposed model uses the biaxial strength envelope of Vecchio (1992)
in Fig. (5.8) to describe failure under biaxial compression. The advantage
of this formulation is computational simplicity without sacrifice of

83
accuracy relative to the experimental data of Kupfer et al. (1969). The
principal stresses in two orthogonal directions are denoted by σ1 and σ2
with σ1 ≥ σ 2 . The following equations define the failure surface:
− σ2 − σ2 2
K c1 = 1 + 0.92( ) − 0.76( )
f c′ f c′
−σ −σ
K c 2 = 1 + 0.92( 1 ) − 0.76( 1 ) 2 .......(5.20)
f c′ f c′
σ1p = K c1ε0
σ2 p = K c 2 ε0 .......(5.21)

Where fc’= uniaxial concrete compressive strength.

The strains corresponding to the ultimate strengths on the orthogonal
directions are:

ε1p = K c1ε 0
ε2 p = K c 2 ε0 .......(5.22)
where εo= uniaxial strain corresponding to the ultimate strength. It
follows readily that the above relations result in a symmetric state of
stress and strain under equal biaxial compression.

5.8.1.2 Compression-tension region:

A simple bilinear strength envelope of describing the tension-
compression behavior has been used in this study. The initial failure
(crack initiation) envelope is defined by the following bilinear relation in
reference to the original proposal by Balakrishnan and Murray (1988a),
and used by Ayoub and Filippou (1998):
⎛ σ ⎞
σ1 = ⎜⎜1 + 0.5 2 ⎟⎟f t
⎝ f c′ ⎠
σ 2 = −f c′ .......(5.23)
For 0 ≤ σ 2 ≤ f c' , the stress σ1 varies between the limits of 0.5f t ≤ σ t ≤ f t in
Fig. (5.9). The proposed model assumes that a tensile stress less than 0.5ft
do not affect the compressive strength. The crack initiation condition is a
modification of the smooth transition curve between a straight line and
the biaxial compression envelope of Kupfer et al. (1969). The cracking
strength (feq) of the stress-strain relation relative to the direction of
orthotropy (i) in Fig. (5.10) is the ordinate of the intersection point of the

84
failure curve in Eq. (5.23) with the straight line passing through the origin
and the current stress state (Fig. (5.9)).
After cracking takes place in one direction, concrete can still carry
compressive stresses in the direction orthogonal to the crack. There is
experimental evidence, however, that the concrete compressive strength
is reduced by lateral tensile strains (Vecchio and Collins, 1982). In the
proposed model, the reduction of concrete compressive strength depends
on the tensile stress at the instant of initial cracking and not on the lateral
tensile strain. This proposal, which dates back to the work of
Balakrishnan and Murray (1988a), assumes that concrete is permanently
damaged at the instant of crack formation. The reduction in concrete
compressive strength after cracking in the orthogonal direction is defined
by the following relations suggested by Belarbi and Hsu (1995):
⎡ ⎛ ε ⎞ ⎛ ε ⎞2 ⎤ ε2
σ 2 = ζ σ f c ⎢2⎜⎜ 2 ⎟⎟ − ⎜⎜ 2 ⎟⎟ ⎥
'
if ≤1
⎢⎣ ⎝ ζ ε ε o ⎠ ⎝ ζ ε ε o ⎠ ⎥⎦ ζ εεo
⎡ ⎛ ε ζ ε − 1 ⎞2 ⎤ ε2
σ 2 = ζ σ f c ⎢1 − ⎜⎜ 2 ε o ⎟⎟ ⎥
'
if 〉1 .......(5.24)
⎣⎢ ⎝ 2 ζ ε − 1 ⎠ ⎦⎥ ζ ε
ε o

where εo is the concrete cylinder strain corresponding to peak cylinder

strength fc’; ζσ and ζε are stress and strain softening coefficient of
concrete, respectively, and can be defined as:
0.9 1
ζσ = ; ζε = .......(5.25)
1 + 400ε1 1 + 500ε1

σ2 is concrete compressive strength after cracking; ε2 is corresponding

strain; and fc’ is uniaxial compressive strength of concrete. Fig. (5.10)
depicts the relation between uniaxial compressive strength envelope
(curve a) and reduced stress-strain relation in the principal compressive
stress direction (curve b).

5.8.1.3 Biaxial tension region:

A constant tensile strength equal to the uniaxial tensile strength of
concrete is used in both principal strain directions:
f eq = f t .......(5.26)
where feq is the tensile stress at the instant of cracking; and ft is the
uniaxial tensile strength of concrete.

85
 0.2

0.0

-1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2

-0.2

-0.4

Kupfer 1969
Ayoub and Filippou 1998 -0.6

-0.8

-1.0

-1.2

-1.4

Fig. (5.8) adopted biaxial strength envelope.

⎛ σ ⎞ σ1
σ1 = f t ⎜⎜1 + 0.5 2' ⎟⎟
⎝ fc ⎠

tensile cracking
ft
feq
current state
0.5ft

σ2
σ2 = -fc’
compression crushing

Fig. (5.9) Tensile-compression failure surface (Balakrishnan

and Murray 1988a)

86
 εi

feq

εο ζεεo
εcr σi

curve b ζσfc’

curve a
fc’

Fig. (5.10) Stress-strain relation adopted.

5.8.2 Material stiffness matrix:

The direction of principal strains in the concrete deviated somewhat from
the directions of principal stresses in the concrete (Fig. (5.11)). However,
it remains a reasonable simplification to assume that the principal strain
and the principal stress axes for the concrete coincided (Vecchio and
Collins, 1986).
Concrete is treated as an orthotropic material whose axes of
orthotropy coincide with the principal directions of total strain. The
stress-strain relations are first defined in the axes of orthotropy and are
then transformed to the global reference system by a rotation
transformation. The local stress-strain relations are:

{σ} = [D]L{ε} .......(5.27)

where {σ} and {ε} are stress and strain vectors, respectively; [D]L is a
local material stiffness with the form:

87
 Inclination of principal compressive srressn,θ1)strain
50

θ1
40

30
(degree)

20

θ1)stress=θ1)strain+10 0

10 θ1)stress=θ1)strain

θ1)stress=θ1)strain-10 0

0
0 10 20 30 40 50
Inclination of principal compressive strain,θ1)strain
(degree)

Fig. (5.11) Comparison of principal compressive stress direction with principal

compressive strain direction (after Vecchio and Collins, 1986).

⎡ E1 ν E1 E 2 0⎤
1 ⎢ ⎥
[D]L = ν E1 E 2 E2 0⎥ .......(5.28)
1 − ν2 ⎢
⎢ 0 0 G ⎥⎦
⎣

where E1 and E2 are secant moduli of elasticity in directions 1 and 2,

respectively; ν is Poisson’s ratio; and G is secant shear modulus. The
determination of the Poisson’s ratio is addressed in the following section.
The secant modulus G is determined from the formula:

(
G = E1 + E 2 − 2ν E1 E 2 K/ 4 ) .......(5.29)

The global material stiffness matrix is derived from the local material
stiffness matrix by rotation transformation matrix according to:

[D]G = [T]T [D]L [T] .......(5.30)

88
where [T] is a rotation transformation matrix between principal strain
axes and global axes:

⎡ c2 s2 cs ⎤
[T ] = ⎢⎢ s 2 c 2 − cs ⎥⎥ .......(5.31)
⎢⎣ − 2cs 2cs c 2 − s 2 ⎥⎦

5.8.2.1 Poisson’s ratio:

The Poisson’s ratio in the stress-strain relations is given by:
ν = ν1ν 2 ............(5.32)
where νi denotes Poisson’s ratio in the equivalent uniaxial strain direction
i. Eq. (5.32) was originally proposed by Darwin and Pecknold (1977) for
biaxial stress states.
If the equivalent strain in direction i is tensile, the Poisson’s ratio νi
is assumed constant and equal to its initial value (νo) before cracking. In
the post-cracking range the Poisson’s ratio is reduced to zero. If the
equivalent strain in direction i is compressive, Poisson’s ratio (νi) is a
function of the equivalent uniaxial compressive strain (Kupfer et al.,
1969). The following simple expression for νi agrees well with
experimental evidence:
⎡ ⎛ εi ⎞ ⎤
2

ν i = ν o ⎢1 + 1.5⎜⎜ ⎟⎟ ⎥ .......(5.33)
⎢⎣ ε
⎝ ⎠ ⎦
ip ⎥
where νo is the initial value; εi is the equivalent uniaxial strain in direction
i; and εip is the strain corresponding to peak stress in direction i according
to Fig. (5.10). An upper bound value of 0.5 is imposed on νi, which
corresponds to incompressible behavior.

5.8.3 State determination:

The process of state determination requires the calculation of stresses (σx,
σy, and τxy) for the given strain state (εx, εy, and γxy). Subscript “ t-1 ”
denotes the value of a variable at the last converged point of the load-
deformation response. Subscript “ t ” denotes the current load step and
superscript “ j ”, the iteration counter. The process consists of the
following steps:
1- Determine the new crack direction (θ) relative to the global reference
axis from the total strains (εx, εy, and γxy):

89
 j
γ xy
tan 2θ = .......(5.34)
εx − εy t

2- Determine the principal concrete strains (ε1 and ε2):

j
⎧ εx ⎫
⎧ ε1 ⎫ cos 2 θ sin 2 θ sin θ cos θ⎤ ⎪ ⎪
j
⎡
⎨ ⎬ =⎢ ⎥ ⋅ ⎨ εy ⎬ .......(5.35)
⎩ε 2 ⎭ t +1 ⎣ sin 2 θ cos 2 θ − sin θ cos θ⎦ ⎪ ⎪
⎩γ xy ⎭ t
3- With the principal strains determine the equivalent uniaxial strains (ε1f
and ε2f):
⎧ ε1f ⎫ ⎧ ε1f ⎫ 1 ⎡1 ν ⎤ ⎛⎜ ⎧ ε1 ⎫ ⎧ ε1 ⎫ ⎞⎟
j j

⎨ ⎬ =⎨ ⎬ + 2 ⎢ ⎥ ⋅ ⎜ ⎨ε ⎬ − ⎨ε ⎬ ⎟ .......(5.36)
ε ε
⎩ 2 f ⎭ t ⎩ 2 f ⎭ t −1 1 − ν ⎣ ν 1 ⎦ ⎝ ⎩ 2 ⎭ t ⎩ 2 ⎭ t −1 ⎠
4- With the approximate stress state from the previous iteration determine
the uniaxial stress-strain relation in the principal strain direction. This
determination depends on the quadrant of the biaxial strength
envelope in which the current stress-state falls.
5- With the current equivalent uniaxial strains and stiffness values (E1 and
E2), determine the current stresses in the principal strain directions (σ1
and σ2). These stresses are not exact and will be subsequently
corrected. At this stage they are supposed to yield the approximate
location of the current state in stress space.
j j
(
σ1 t = σ1 t −1 + E1 ⋅ ε1f t − ε1f t −1
j
)
( )
.......(5.37)
σ 2 t = σ 2 t −1 + E 2 ⋅ ε 2 f t − ε 2 f t −1
j j j

6- The stresses in the principal strain directions are transformed to the

stresses in the global reference system (σx, σy, and τxy):
j
⎧σ x ⎫ ⎡ cos 2 θ sin 2 θ ⎤
⎥ ⎧ σ1 ⎫
j
⎪ ⎪ ⎢
⎨ σ y ⎬ = ⎢ sin θ cos θ ⎥ ⋅ ⎨ ⎬
2 2
.......(5.38)
⎪τ ⎪ ⎢ sin θ cos θ − sin θ cos θ⎥ ⎩σ 2 ⎭ t
⎩ xy ⎭ t ⎣ ⎦
The local stiffness matrix is updated and transformed to the global
reference system for use in the next iteration.

5.9 Solution algorithm:

The basic steps of the analysis procedure are summarized below with “ j ”
and “ t ” denoting the iteration and load increment number, respectively:
1. For the t th loading increment, evaluate the incremental nodal force
vector {ΔF}t .

90
2. State the iterative loop (j=1).
3. Solve the linear system of equations for the incremental displacement
{ΔU}tj as follows:

{ΔU} = [K ] {ΔF}
j
t
−1 j
t .......(5.39)
Add incremental displacement to total displacement vector {U}.
4. For each Gauss point, evaluate the corresponding strain.
5. Calculate the stresses using the process of state determination.
6. Repeat steps (4-5) for each Gauss point in an element. Calculate the
equilibrium nodal force vector for that element by evaluating the
volume integral (numerically) of the computed total stresses over that
element:

{R } = [B] {σ}dvol.
∫
T
e .......(5.40)
vol

7. Repeat steps (4-6) for each element in the structure. Assemble the total
available force vector {R}, where

∑{
E

{R} t
j
= Re} .......(5.41)
e =1

Finally evaluate the residual force vector {ΔF} as follows:

{ΔF}tj+1 = {F} − {R}tj .......(5.42)
where {F} is the total externally applied force vector until this load
increment.
8. If iteration has not converged, start a new iteration (j+1), replacing the
external applied force vector by the residual force vector {ΔF} and go
to step 3.
The local stiffness matrices are updated on the second iteration of
every increment, rather than the first. This is very helpful in modeling
very nonlinear materials such as concrete 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 bases of the
actual stress present at that increment, rather than on the bases of the
previously reached stress state.

To implement this numerical model of concrete, suitable nonlinear

solution techniques are required to trace the equilibrium path of a
structure. Because of numerical problems associated with fracture

91
processes of concrete, the use of the true incremental tangent stiffness
matrix is often undesirable. The global stiffness matrix is assembled
using the secant (positive) material moduli and thus solution schemes
have to rely heavily on iterative procedures (Fig. (5.12)).
Load

Deflection

Fig. (5.12) Solution scheme.

5.10 Load increment:

In the finite element analysis of reinforced concrete structures, which
feature cracking, yielding, and crushing, the selection of load increment
size is generally regarded as a matter of experience and judgement. It has
been shown (Gajer and Dux, 1989), however, that the prediction of the
structural response can be sensitive to this choice. Herein, in addition to
the predefined load increments, an automatic load increment selection
scheme, which removes the need for guesswork, can be adopted. The
scheme employs a formula used to govern the size of the load increment.
It is:
⎛I ⎞
ΔFt = ⎜⎜ d ⎟⎟ ⋅ ΔFt −1 KK ≤ ΔFt −1 .......(5.43)
⎝ I t −1 ⎠

where Id is the desired number of iterations (Id = 10 in the present study);

It-1 is the number of iterations required at the last (t-1)th increment.

92
 The failure load is predicted quite accurately by the sudden increase
in the unbalanced forced in the last stage of loading.

5.11 Convergence criterion:

The convergence criterion is an indication for the control of the level of
accuracy of the solution. It terminates the equilibrium iteration as soon as
the desired accuracy is achieved. Essentially, four quantities may be
monitored during the iteration process in order to establish a criterion:
displacement, residual forces, incremental strain energy, and secant
stiffness moduli.
If displacements are chosen, then one possible method is to compare
each individual nodal value with the corresponding value obtained during
the previous iteration. Then, provided that this change is negligible small
for all nodal points, convergence can be deemed to have occurred as:

∑
N

({ΔU } )
i
j 2
t

i =1
* 100 ≤ Tolerance .......(5.44)

∑({
N

U i })
2

i =1

where N denotes the total number of nodal points in the problem.

For the convergence based on the residual force values, the criterion
employed is:

∑
N

({ΔFi }t ) 2
j

i =1
* 100 ≤ Tolerance .......(5.45)

∑{ }
N

( Fi ) 2
i =1

Criteria based only on displacements or residual forces can be

misled by slow convergence rates. The energy criterion has proven to
perform well in most cases. This criterion provides some indication of the
amount of work done by the out-of-balance forces on the displacement
increments. When both the displacements and residual forces are near
their equilibrium values, the increment in internal energy during each
iteration can be compared to the initial energy increment as:

93
 φ t ΔU t
≤ Tolerance ......(5.46)
R U t −1
A tolerance of 1/100 is found satisfactory throughout this study.
In addition to the convergence tolerances described above, a ceiling
is provided to limit the number of iterations (equal 30) performed for
each load step in case convergence tolerances provided are too stringent.

5.12 Verification of the program:

To achieve the requirements of the present study, a computer program
(MHND) has been written. It is coded in Microsoft Visual Basic 6.0 and
utilized a Pentium II at the computer laboratory of the Civil Engineering
Department, Engineering College of the University of Baghdad. The
equation solver used is based upon symmetric banded Gaussion
Elimination. The flow chart of the main steps of this program are shown
in Fig. (5.13).
The purpose of this section is the application of the computer
program for the analysis of different cases and to compare the analytical
results with available experimental data in order to inspect the
applicability of the computer program in the analysis of such cases.

5.12.1 Shallow beam OA1:

The prototype, simply supported, reinforced concrete beam shown in Fig.
(5.14) was tested by Bresler and Scordelis (1963). This problem has
become nearly a classical case for testing of different finite element
formulations and programs, see e.g. (Al-Sabah, 1983), and (Cedolin and
Nilson, 1978). Fig. (5.15) shows the finite element idealization of half the
beam used in the present study, the dashed lines indicating reinforcement
bars.
The concrete is idealized by using 133 four-noded rectangular
elements as shown in Fig. (5.15). 38 two-noded bar elements represent
the reinforcement. Bond-slip between concrete and reinforcement is
modeled by defining two nodes occupying the same location, one in a
concrete element and the other in steel element, linked together by 38
fictitious linkage elements. As a result of the above idealization, 196
nodes are needed to define the structure.
An incremental concentrated load is applied at node 196. The load is
applied in previously defined twenty increments up to the failure load.
Fig. (5.16) illustrates the experimental and analytical load-midspan

94
 START

INPUT DATA

COMPUTE STRUCTURAL STIFFNESS WITH

INITIAL MATERIAL PROPERTIES

LOAD INCREMENT

ITR = ITR + 1

COMPUTE STRUCTURAL
YES
ITR = 2 STIFFNESS WITH SECANT
MATERIAL PROPERTIES
NO

SOLVE FOR INCREMENTAL DISPLACEMENT

CALCULATE STRAINS, STRESSES

&
CHECK FOR FAILURE

CALCULATE UNBALANCED
LOAD VECTOR

CHECK YES
MAX
ITR

NO

NO CHECK
CONVER.

YES

PRINT RESULTS
(DISPLACEMENTS, STRESSES, FORCES)

FIG. (5.13) Flow chart of the program MHND.

95
 deflection curves, where the analytical results approach the experimental
curve very closely.

425.45

95.25
95.25
310 228.6 3656 mm 228.6

4 φ 28.5 mm
(φ 9 bars) fy = 551.6 Mpa (80 ksi) fc’ = 22.55 Mpa (3.27 ksi)
Es = 206850 Mpa (30000 ksi) ft = 3.96 Mpa (0.575 ksi)
Ec = 20000 Mpa (assumed)
εc = 0.0022 (assumed)

Fig. (5.14) Experimental beam OA1 (Bresler and Scordelis 1968).

196
133

5
4
4
3
3
2
2 1
1
i : Node no.
i : Element no.

Fig. (5.15) Finite element discretization of beam OA1.

96
 350

300

250
Load (kN)

200

150

100 Experimental
Present study
50

0
0 1 2 3 4 5 6 7
Mid-span deflection (mm)

Fig. (5.16) Experimental and present study load-midspan

deflection response of beam OA1.

5.12.2 Tension specimen UC-15:

Abrishami and Mitchell (1996) reported a test series involving 10 tension
specimens loaded under tension force. Tension specimen UC-15, shown
in Fig. (5.17.a), had a length of 1500 mm and cross-section dimensions of
(95x170 mm). A single reinforcing bar, 16 mm diameter, was provided.
Due to symmetry, only a quarter of the specimen has been idealized
by finite element analysis, using 20 rectangular Q-4 elements for
concrete, 10 truss bar for reinforcing bar, and 11 linkage elements for
bond-slip, see Fig. (5.17.b).
The predicted response for the specimen obtained from a finite
element analysis using the program MHND is compared to the observed
behavior in Fig. (5.17.c). Due to shrinkage of concrete, the elongation of
the specimen is negative prior to load application. As can seen, the test
results included elastic uncracked, post-cracking, and post-yielding
regions. Good agreement is seen both in terms of load-deflection
response and ultimate strength.

97
 N

100

80

60

N (kN)
1500 mm

40
Test result

95 Present study
20
170 Bare bar

0
-1 0 1 2 3 4 5
Elongation (mm)
N
(a) (b) (c)
Fig. (5.17) Tension specimen UC-15, (a) Dimensions, (b) Finite element
mesh, (c) Comparison of test results and predicted response.

5.12.3 Shear transfer specimen CB2M:

The composite shear transfer specimen CB2M, tested by Mattock (1981),
was cast in two stages. The interface between the two concretes lying in
the shear plane (Fig. (5.18)). The first part of this specimen was three
days old when the second part was cast. The interface was deliberately
roughened to amplitude of 6 mm (¼ in.) to conform to the requirements
of the ACI Code. An effort was made to obtain good bond between the
two concretes by cleaning and wetting the first cast concrete surface.
Details of the specimen is shown in Fig. (2.3). It is designed to be gripped
by friction on faces “A” as shown in Fig. (5.18). The 32260 mm2 (50 in2.)
shear plane is subjected to shear without moment. The shear transfer
reinforcement was in the form of closed stirrups, which wrapped around
longitudinal reinforcement so as to ensure positive anchorage on both
sides of the shear plane. Additional reinforcement was provided to

98
prevent failure of the specimen away from the shear plane, details of the
specimen geometry and reinforcement are shown in Fig. (2.3). The
specimen was cracked in the shear plane before being subjected to shear
loading.
The specimen was tested in the specially constructed two-part frame
shown in Fig. (5.18). Opposite sides of the specimen were attached to the
parts of the frame by gripping plates. The shearing forces were provided
by the diagonally opposed pairs of 266880 N (60 kips) capacity hydraulic
center hole rams X and Y in Fig. (5.18). Failure was considered to have
occurred when the shear could not be increased further and both slip and
separation increased rapidly.
The specimen model in Fig. (5.19) consists of 144 four-node
isoparametric quadrilateral concrete elements. The reinforcement is
represented with 108 truss bar elements. 18 linkage elements are used to
link stirrup reinforcement bars to concrete elements, while other bars are
perfectly bonded to concrete elements. The shear plane is represented
with 9 shear-transfer interface elements.
The analytical response of the shear transfer specimen CB2M is
compared with the experimental measurements of Mattock (1981) in Fig.
(5.20). The results obtained by finite element analysis agree well with the
experimental response of the specimen.
Fig. (5.21) shows a comparison between the experimental response
of the composite initially cracked shear transfer specimen CB2M and the
response of the monolithic initially cracked shear transfer specimen
MN2M tested by Mattock too. It can be seen that under monotonic
loading the behavior and strength of the initially cracked composite
specimens with good bond at the interface, was very similar to that of the
initially cracked, monolithic concrete specimens.

99
 182
144

5
5
4
4
3
3
2
2
1
1

i : Node no. Interface element

i : Element no. Bar element

Fig. (5.19) Finite element modeling for specimen CB2M.

100
 250

200
Shear (kPa)

150

100

Experimental

50 Present study

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Slip (mm)
Fig. (5.20) Comparison between experimental and analytical
response of shear transfer specimen CB2M.

250

200
Shear (kN)

150

100

CB2M Specimen (composite)

50 MN2M Specimen (monolithic)

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Slip (mm)
Fig. (5.21) Comparison between experimental response of composite
specimen CB2M and monolithic specimen MN2M.

101
 CHAPTER SIX
ANALYSIS OF COMPOSITE BEAMS

6.1 Introduction:
Several experimental investigations of composite concrete beams were
done in literature. Unfortunately, some of these investigations are not
available in Iraq. Also some of the available tests concentrated only on
specific aspects of the structural problem and their experimental data is
usually incomplete. In this chapter, the computer program (MHND) is
used to analyze several composite concrete beams, which were tested by
others, and comparisons between the experimental and finite element
results are shown. Also, a parametric study deals with shear and moment
capacity of a composite concrete beam is presented.

6.2 Revesz tests:

Preliminary tests were made by Revesz (1953) at the Imperical Collage of
Science and Technology London University, London, England, on
composite T-beams of 4.267 m (14 ft) span to determine the behavior of
the beams under loads.
The section of the test beam is shown in Fig. (6.1). The particular
shape of the section was chosen so as to represent a strip of a floor
construction.

546.1 mm
38.1

in-situ
203.2 mm
152.4

76.2

Fig. (6.1) Section of test beam (Revesz, 1953).

102
 Yield point of the mild steel was found to be 268 MPa. No provision
had been made for stirrups, or serration at the top of the precast beam.
Age of concrete of cast-in-place flange and precast web at time of test
was 7 days and 29 days, respectively. The test load was applied at third-
points of the span.
Due to the symmetry in both geometry and loading, only half of the
beam is considered in the finite element idealization by introducing the
appropriate boundary conditions along the beam centerline.
The concrete is idealized by using 420 four-noded rectangular
elements as shown in Fig. (6.2). The reinforcement is represented by 70
bar elements, connecting with concrete elements by 71 bond-slip linkage
elements. Surface between the two concretes is idealized by 70 shear-
transfer interface elements.

639

420

5
4
4
3
3
2
2
1
1

i : Node no. Interface element

i : Element no. Bar element

Fig. (6.2) Finite element discretization of the beam tested by Revesz.

Fig. (6.3) represents the load mid-span deflection curve of the beam
considered. Comparison with experimental results indicates a close
agreement till about 80% of the ultimate load. A stiffer behavior of the
theoretical model was observed during the next load increments. This
discrepancy of results may be reasoned due to the tension stiffening
adopted in this research which is not suitable for poor reinforced

103
concrete. However, since this phenomenon only produce secondary
effects, the analytical ultimate load level (382 kN) is detected quite well
compared with the experimentally observed of 379.4 kN, with an error of
only 0.7%.

400

300
Load (kN)

200

Experimental

100 Present study

0
0 40 80 120 160
Deflection at mid-span (mm)

FIG. (6.3) load mid-span deflection curve of the beam tested by Revesz.

6.3 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. (6.4) shows the beam cross section used with nominal
dimensions indicated. The reduced breadth of web produced high
horizontal shear 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. (6.5) shows the arrangement of stirrups (#4 bars) in 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

104
 for 0.20 but #3 bars were used instead of #4 bars. As the nominal steel
percentage decreased from 0.11 to 0.06 half of the #3 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 (6.1). Percentage of steel was
calculated by dividing the area of all stirrups crossing the joint by the
total joint area.

101.6 381

Neutral axis φ 12 bar

88.9

431.8
Cage stirrup

152.4
50.4 50.4

152.4 φ 25 bars

Fig. (6.4) Beam cross section (Saemann and Washa, 1964).

P P
2 304.8 2

10 stirrups at 127 mm 355.6

plus 9 cage stirrups 2438.4 mm 203.2

Fig. (6.5) Arrangement of stirrups (#4 bars) in beams having the maximum
percentage of stirrup steel across the joint (Saemann and Washa,
1964)

105
 Properties of the reinforcing steel are given in Table (6.2). Cage
stirrups were added to reduce shear stresses below the joint. All steel
members of the reinforcement unit were welded together.

Table (6.1) Stirrup data.

Nominal percentage of
Bar spacing
stirrup steel across Bar size
mm (in.)
The joint
1.02 #4 127 (5)
0.51 #4 254 (10)
0.20 #4 508 (20)
0.11 #3 508 (20)
0.06 #3 1016 (40)
0.00 - -

Table (6.2) Properties of reinforcing steel.

Ultimate tensile
Yield point
Size strength
MPa
MPa
#3 370 560
#4 294 404
#8 252 401

Seven days after the webs were made the slabs were cast. All beams
were tested 28 days after the slabs were cast, 35 days after the webs were
cast. Beams were supported and loaded as shown in Fig. (6.5).
The measured central deflection with respect to the applied load are
shown in Fig. (6.6) for beams have 2.438 m (8 ft) clear span, and with
different percentage 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 deflection) 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 does not broken, and its effectiveness begins after a
crack forms in the joint (dowel action).

106
 400

350
0.51% 1.02%
0.20%
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. (6.6) Measured central deflection with respect to the applied load
for beams with different percentage steel across the joint.

6.3.1 Finite element idealization:

Due to symmetry, both in geometry and loading in beams tested by
Saemann and Washa (1964), only half of the beams is idealized by
restraining the horizontal movement of the nodes at the beam centerline.
Fig. (6.7) illustrates the finite element mesh layout for 2.438 m (8 ft)
beam having the maximum percentage of stirrup steel across the joint.
Concrete is idealized by using 198 four-nodded 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.

107
 A

5
4
4
3
3
2
2
1
1

i Interface element
: Node no.
Bar element
i : Element no.

Fig. (6.7) Finite element discretization of a beam tested by Saemann and Washa.
 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.

6.3.2 Discussion of results:

Fig. (6.8) shows the experimental and finite element results of six beams
having different percentage of steel across the joint. The same trend of
behavior is seen for numerical and tests results, but the finite element
results show less value. This may be attributed to the selfweight effects
on the stresses and strains, which are neglected in the analysis.
As the shear caused slip to develop between flange and web, the
beam began to act as a partially composite member. This action is shown
in Figs. (6.9) and (6.10) where the computed normal strain distribution at
Section A-A (refer to Fig. (6.7)) 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
remaining in its position although the load increases. This behavior
results from the tendency of the beam and the slab to act separately.
Figs. (6.11) and (6.12) show the horizontal stress (in X-direction) at
section A-A, 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 steel) is shown in Fig. (6.13). The
overall depth and pattern of the cracks indicates that the failure is due to
shear. This results agrees with the type of failure observed experimentally
by Saemann and Washa (1964).
The cracking and failure mechanism of the joint was 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 toward
the load point as the loading progressed. In general, the cracks stopped

109
when they reach the horizontal joint. However, after additional loading
some of the cracks continued horizontally below the contact surface
toward 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.
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. (6.14) and (6.15),
respectively, for a series of increasing loads. The curves show that the
values increasing as the load increased, that maximum shearing stresses
and slips are usually located about 900 mm from the left end of the beam,
and that separations (loss of bond) are 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 notice that the shearing stresses and the normal
stresses between the two concretes decrease with the increase in the steel
percentage across the joint.

110
 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. (6.8) Comparison between experimental and present study results.
 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. (6.9) Normal strains distribution at section A-A in beam

with 0% 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

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

Fig. (6.10) Normal strains distribution at section A-A in

beam with 1.02% steel across the joint, showing
two-beam action.
112
 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. (6.11) Normal stresses distribution at section A-A in

beam with 0% steel across the joint, showing two-
beam action.

450

400
Distance above bottom (mm)

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. (6.12) Normal stresses distribution at section A-A in
beam with 1.02% steel across the joint, showing
two-beam action.

113
5
4
4
3
3
2
2
1
1

i Interface element
: Node no.
Bar element
i : Element no.

Fig. (6.13) Prediction of cracking by the finite element method at the ultimate load for beam 16C (zero percentage steel).
 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)
(a) (b)
1 0.01

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. (6.14) 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
Shearing stress (MPa)

-0.02
-1

Slip (mm)
-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)

(a) (b)
1 0.01

Normal separation (mm)

Normal stress (MPa)

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. (6.15) 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.
6.4 Parametric study:
The position of the concentrated load affects the distribution of the
shearing stresses (or slips) and the normal separations at the joint. The
maximum shearing stress occurs between the position of the
concentrated load, while the same distribution can be seen away from
the load, Fig. (6.16) and (6.17) respectively.
When an incremental uniformly distributed load (U.D.L.) is
applied on the composite concrete beams with different percentage of
steel across the joint between the precast web and the cast-in-place
slab, the differences in the responses of the beams (midspan
deflections) are less than 1%. Fig. (6.18) shows the calculated midspan
deflections of the beam 16C (zero steel percent) and the beam 10A
(1.02% steel percent).
Fig. (6.19) shows the shearing stresses, slips, normal stresses, and
normal separations at the interface between the precast beam and the
cast-in-place slab under the ultimate uniformly distributed load
(U.D.L. = 130 kN). The curves show that the maximum shearing stress
and slip are located about 600 mm from the left end of the beam, and
there is no reparation between the two elements.

1
1 2 3
Shear stress (MPa)

-1

-2 1 2 3

-3

-4
0 200 400 600 800 1000 1200 1400
Distance from left end of the beam (mm)

Fig. (6.16) Variation of shearing stress distribution at the

joint with location of concentrated load.

117
 1 2 3
0.01
Normal separation (mm)

0.00

-0.01

-0.02

-0.03

-0.04 1 2 3
-0.05

-0.06
0 200 400 600 800 1000 1200 1400
Distance from left end of the beam (mm)
Fig. (6.17) Variation of normal separation at the joint with
location of concentrated load.

140

120

100
U.D.L. (kN/m)

80

60
Steel percentage
40
0%
20 1.02%

0
0 1 2 3 4 5
Central deflection (mm)

Fig. (6.18) Midspan deflection of the composite concrete beam

under U.D.L.

118
 0 0.00
Shearing stress (MPa)

-0.02
-1

Slip (mm)
-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 the beam (mm)

(a) (b)

0 0.00

Normal separation (mm)

Normal stress (MPa)

-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 end of the beam (mm)

(c) (d)
Fig. (6.19) Distribution of (a) shearing stresses, (b) slips, (c) normal stresses, and (d) normal separations across the joint
for beam 16C (zero steel percentage) under ultimate uniformly distributed load (U.D.L.=130 kN/m).
 Conventionally, a reinforced concrete beam has an ultimate shear
and flexural strength (Vn and Mn) as follows (ACI Code):

Vn = Vc + Vs .......(6.1)

where
⎛ Vd⎞b d
Vc = ⎜⎜ f c' + 120ρ w u ⎟⎟ w ≤ 0.3 f c' b w d .......(6.2)
⎝ Mu ⎠ 7
Vu d
& ≤1
Mu
A v f yd 2 '
Vs = ≤ fc bwd .......(6.3)
s 3
⎛ f ⎞
M n = ρbd 2 f y ⎜⎜1 − 0.59ρ y' ⎟⎟ .......(6.4)
⎝ fc ⎠

where Vc and Vs are nominal shear strength provided by concrete and

shear reinforcement, respectively; ρw= As/bwd; Mu is the factored moment
at section; and ρ is the ratio of reinforcement.
Figs. (6.20) and (6.21) show a comparison between the ultimate
shear strength and moment strength, respectively, according to the
experimental, the finite element, and the ACI Code results when changing
the percentage of steel across the joint.

A summary of results (ultimate moment, ultimate shear force, and

maximum shearing stress) obtained from ACI Code, experimental tests,
and finite element analyses are listed in Table (6.3). Several important
observations may be made from Table (6.3), which shows clearly that the
ultimate shear and flexural capacities, of a composite concrete beam,
increase as the steel percentage across the joint (shear connectors),
between the precast web and the cast-in-place slab, increases. These
increases diminish as the steel percentage become more than 0.50% from
the contact area between the two elements. Also, the ACI Code failed to
predict the ultimate capacities in shear and bending for a composite beam,
this may be due to neglecting the effect of partial action. So, the statement
“......as for a monolithically cast member of the same cross-sectional
shape.” In the ACI Code (Chapter 17) needs to inspect.

120
 180

Ultimate shear capacity (kN)

160

140

Experimental
120
Present study
ACI Code
100
0.00 0.20 0.40 0.60 0.80 1.00 1.20
Steel percentage (%)

Fig. (6.20) Variation of ultimate shear capacity with steel percentage

across the joint.

180
Ultimate moment capacity (kN.m)

160

140
Experimental
Present study
120 ACI Code

100
0.00 0.20 0.40 0.60 0.80 1.00 1.20
Steel percentage (%)
Fig. (6.21) Variation of ultimate moment capacity with steel
percentage across the joint.

121
 Table (6.3) Ultimate capacity.
Ultimate moment
Steel fc’, Ultimate shear capacity, Ultimate shear stress,
Beam capacity,
across MPa kN Mpa
kN.m
joint,
ACI Present ACI Present ACI Present Shear
No. Series % Web Slab Exper. Exper. Exper.
Code study Code study Code study friction
10 A 1.02 21.1 19.8 160 163 164 131 178 179 3.500 6.442 6.703 3.299

5 C 0.51 20.8 22.5 162 163 162 131 178 177 2.700 6.442 6.695 1.649

5 D 0.20 23.4 24.7 164 155 155 138 170 170 2.153 6.146 6.378 0.825

12 C 0.11 20.6 23.9 164 140 140 132 154 154 2.044 5.552 5.721 0.584

14 C 0.06 21.6 19.8 160 126 126 132 138 138 0.600 4.987 5.113 0.292

16 C 0.00 20.9 21.1 161 106 107 130 116 117 0.600 4.180 4.364 0.000
 Table (6.4) shows a comparison in the ultimate shear force for the
beams 16C and 10A with different provided stirrups. It can be shown that
the ACI Code gives different values compared with the predicted in this
study.

Table (6.4) Comparison in ultimate shear force.

Ultimate shear force,
Provided KN
Beam
stirrups Present
ACI Exper.
study
Full stirrups 130 116 117
Without cage
16C 109 Not available 117
stirrups

Without stirrups 34 Not available 93

Full stirrups 131 178 179
10A Without cage
131 Not available 145
stirrups

123
 CHAPTER SEVEN
CONCLUSIONS AND RECOMMENDATIONS
7.1 Conclusions:
The following conclusions can be drawn from the present study:

1. A good estimated for analysis and ultimate loads of composite

concrete beams that might be made using the Program MHND that
may contribute in more practical designs.
2. The present algorithm for analysis of composite concrete beams
proves to predict their 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, respectively,
between two concretes cast in different times, is quit good.
4. The adopted material constitutive relationships are found to give
satisfactory results at both the service and ultimate load stages of
composite concrete beams.
5. In general, the ultimate shear strength and flexural strength, for
composite concrete beams, increased considerably with intermediate
contact surface as the amount of stirrup steel across the joint (shear
connectors) increased up from zero to 0.50% from the contact
surface.
6. Shear connectors have no observable effect on the composite
concrete beam behavior as the early load stages, and their
effectiveness begins after a considerable slip occurs between the
precast beam and the cast-in-place slab.
7. From 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.
8. The shearing stress distributions and normal separation distributions,
along the contact surface between precast concrete beam and cast-in-
place slab, are considerably affected by the location of concentrated
load, i. e. distance from the support.
9. The ACI Code (ACI 318M-95) failed to predict the ultimate
capacities in shear and bending for a composite concrete beam.
7.2 Recommendations:
1. The extension of the present work to prestressed concrete is
required to make the written program applicable to a wider range
of composite concrete structures.
2. More detailed experimental tests on composite concrete beams are
required. These tests are needed to verify the results of the
program in the present study.
3. Time dependent effects are urgently need to be incorporated in the
present study. These effects should include concrete shrinkage,
creep, and aging of concrete.
4. Modification and extended of the present work may be done to
investigate dynamic behavior of the composite concrete beams.
5. Experimental tests, on shear transfer between two concretes cast
with different ages, are desirable to obtain a constitutive model for
shear transfer for the two and three-dimensions finite element
analysis.
 REFERENCES
AASHTO, Standard Specifications for Highway Bridges, 14th Edition, American
Association of State Highway and Transportation Officials, Washington, D. C., 1989.

Abdel-Halim, M. A. H., and Alostaz, Y. M. (1998). “Bond Slippage in Reinforced

Concrete Flexural Members-Experimental Investigation.” J. Struct. Engrg., 24(4),
189-194.

Abrishami, H. H., and Mitchell, D. (1996). “Influence of Splitting Cracks on Tension

Stiffening.” ACI Struct. J., 93(6), 703-710.

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

ACI-ASCE Committee 333 (1960). “Tentative Recommendations for Design of

Composite Beams and Girders, for Buildings.” ACI J., 57(6), 609-628; J. Struct. Div.,
ASCE, 86(12), 73-92.

Ahmad, M., and Bangash, Y. (1987). “A Three-Dimensional Bond Analysis Using

Finite Element.” Comp. Struct., 25(2), 281-296.

Aldstedt, E., and Bergan, P. G. (1978). “Nonlinear Time-Dependent Concrete-Frame

Analysis.” J. Struct. Div., ASCE, 104(7), 1077-1092.

Al-Imam, Raid F. S. (1992). “Nonlinear Finite Element Analysis of Reinforced

Concrete Slab-Beam Systems.” Ph.D. Thesis, Dept. of Civil Eng., Univ. of Baghdad,
183.

Al-Rawi, R. S. (1986). “A New Method for Determination of Tensile Strain Capacity

and Related Concrete Properties.” A Study presented to the College of Engineering,
Univ. of Baghdad, Iraq,.

Al-Sabah, Abdul Salam I. M. (1983). “Non-Linear Time-Dependent Finite Element

Analysis of Plane Reinforced Concrete Members.” M.Sc. Thesis, Dept. of Civil Eng.,
Univ. of Baghdad, 254.

Al-Sabah, A. S., and Al-Ne’aimi, A. A. (1988). “Nonlinear Stiffness Analysis of R/C

Frames.” Proc. Of the 2nd. Iraqi Conference on Engineering Held at Univ. of Mosul,
Iraq, 46-67.

Anderson, A. R. (1960). “Composite Design in Precast and Cast-in-Place Concrete.”

Progressive Architecture, 41(9), 172-178, (Cited according to Hofbeck et al., 1969).

Ayoub, A., and Filippou, F. C. (1998). “Nonlinear Finite Element Analysis of RC

Shear Panels and Walls.” J. Struct. Engrg., ASCE, 124(3), 298-308.

Azizinamini, A., Stark, M., Roller, J. J., and Ghosh, S. K. (1993). “Bond Performance
of Reinforced Bars Embedded in High-Strength Concrete.” ACI Struct. J., 90(5), 554-
561.

126
Badoux, J. C., and Hulsbos, C. L. (1967). “Horizontal Shear Connection in Composite
Concrete Beams under Repeated Loads.” ACI J., 64(12), 811-819.

Balakrishnan, S., and Murray, D. W. (1988a). “Concrete Constitutive Model for

NLFE Analysis of Structures.” J. Struct. Engrg., ASCE, 114(7), 1449-1466.

Balakrishnan, S., and Murray, D. W. (1988b). “Effect of Modeling on NLFE Analysis

of Concrete Structures.” J. Struct. Engrg., ASCE, 114(7), 1467-1487.

Balakrishnan, S., and Murray, D. W. (1988c). “Prediction of R. C. Panel and Deep

Beam Behavior by NLFEA.” J. Struct. Engrg., ASCE, 114(10), 2323-2342.

Balázs, G. L. (1993). “Cracking Analysis Based on Slip and Bond Stresses.” ACI
Mat. J., 90(4), 340-348.

Basu, A. K. (1967). “Computation of Failure Loads of Composite Columns.” Proc.

Inst. Civ. Eng., London, 36, 557-578.

Bathe, K. J. (1996). “Finite Element Procedures.” Prentice-Hall International, Inc.,

New Jersey, 1037.

Bathe, K. J., Walczak, J., Welch, A., and Mistry, N. (1989). “Nonlinear Analysis of
Concrete Structures.” Comp. Struct., 32(3/4), 563-590.

Bazant, Z. P., and Cedolin, L. (1979). “Blunt Crack Band Propagation in Finite

Element Analysis.” J. Engrg. Mech., ASCE, 105(2), 297-315.

Bazant, Z. P., and Oh, B. H. (1983). “Crack Band Theory and Fracture of Concrete.”

Mat. & Struct.; Res. & Test. (RILEM, Paris), 16(93), 155-177.

Belarbi, A., and Hsu, T. T. C. (1994). “Constitutive Laws of Concrete in Tension and
Reinforcing Bars Stiffened by Concrete.” ACI Struct. J., 91(4), 465-474.

Belarbi, A., and Hsu, T. T. C. (1995). “Constitutive Laws of Softened Concrete in

Biaxial Tension-Compression.” ACI Struct. J. 92(5), 562-573.

Bergan, P. G., and Holand, I. (1979). “Nonlinear Finite Element Analysis of Concrete
Structures.” Comp. Meths. Appl. Mech. Eng., 17/18, 443-467.

Birkeland, P. W., and Birkeland, H. W. (1966). “Connections in Precast Concrete

Construction.” ACI J., 63(3), 345-344.

Bortolotti, L. (1991). “First Cracking Load of Concrete Subjected to Direct Tension.”

ACI Mat. J., 88(1), 70-73.

Bresler, B., and Scordelis, A. C. (1963). “Shear Strength of Reinforced Concrete

Beams.” ACI J., 60(1), 51-74.

127
Bresler, B., and Scordelis, A. C. (1964). “Shear Strength of Reinforced Concrete
Beams.” Series II. SESM Report No. 64-2, Univ. of California, Berkeley, CA, (Cited
according to Ahmad and Bangash, 1987).

British Standards Institution (BS 5400: Part 4: 1978), “Steel, Concrete and Composite
Bridges; Part 4. Code of Practice for Design of Concrete Bridges.” London, 48.

Carreira, D. J., and Chu, K. H. (1985). “Stress-Strain Relationship for Concrete in

Compression.” ACI J., 82(6), 797-804.

Cedolin, L., and Nilson, A. H. (1978). “A Convergence Study of Iterative Methods

Applied to Finite Element Analysis of Reinforced Concrete.” Int. J. Numer. Meths.
Eng., 12, 437-451.

Cervenka, V. (1985). “Constitutive Model for Cracked Reinforced Concrete.” ACI

Struct. J., 82(6), 877-889.

Cervera, M., Hinton, F., and Hassan, O. (1987). “Nonlinear Analysis of Reinforced
Concrete Plate and Shell Structures Using 20-Noded Isoparametric Brick Elements.”
Comp. Struct., 25(6), 845-869.

Chan, E. C. Y. (1983). “Nonlinear Geometric, Material and Time Dependent Analysis

of Reinforced Concrete Shells with Edge Beams.” Ph.D. Dissertation, Univ. of
California, Berkeley, 361.

Chandrupatla, T. R., and Belegundu, A. D. (1997). “Introduction to Finite Elements in

Engineering.” 2nd. Edition, Prentice-Hall of India Private Limited, New Delhi, 461.

Chang, W. F., and Ferguson, P. M. (1963). “Long Hinged RC Columns.” ACI J.,
60(1), 1-26.

Chatterjee, P. K. (1971). “Shear Transfer in Reinforced Concrete.” MSCE Thesis,

University of Washington, Seattle, (Cited according to Mattock and Hawkins, 1972).

Cheong, H. K., and MacAlerey, N. (2000). “Experimental Behavior of Jacketed

Reinforce Concrete Beams.” J. Struct. Engrg., ASCE, 126(6), 692-699.

Climaco, J. C. T. S. (1990). “Repair of Structural Concrete Involving the Addition of

New Concrete.” PhD thesis, Council for Nat. Academic Awards, Polytechnic of
Central London, U.K., (Cited according to Cheong and MacAlerey, 2000).

Cook, J. P. (1977). “Composite Construction Methods.” John Wiley & Sons, New
York, 330.

Cook, R. D. (1974). “Improved Two-Dimensional Finite Element.” J. Struct. Div.,

ASCE, 100(9), 1851-1863.

Cook, R. D. (1977). “Ways to Improve the Bending Response of Finite Elements.”

Int. J. Numer. Meths. Eng., 11(6), 1029-1039.

128
Cook, R. D. (1989). “Concepts and Applications of Finite Element Analysis.” 3rd.
Edition, John Wiley and Sons, London,.

Crisfield, M. A., and Wills, J. (1989). “Analysis of R. C. Panels Using Different

Concrete Models.” J. Engrg. Mech., ASCE, 115(3), 578-597.

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.

Darwin, D., McCabe, S. L., Idun, E. K., and Schoenekase, S. P. (1992).

“Development Length Criteria: Bars not Confined by Transverse Reinforcement.”
ACI Struct. J., 89(6), 709-720.

Darwin, D., and Pecknold, D. A. (1997). “Nonlinear Biaxial Stress Strain Law for
Concrete.” J. Engrg. Mech., ASCE, 103(2), 229-241.

De Larrard, F., Schaller, I., and Fuchs, J. (1993). “Effect of Bar Diameter on the Bond
Strength of Passive Reinforcement in High-Performance Concrete.” ACI. Mat. J.,
90(4), 333-339.

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.

Desayi, P., and Krishnan, S. (1964). “Equation for the Stress-Strain Curve of
Concrete.” ACI J., 61(3), 345-350.

Edwards, A. D., and Yannopoulos, P. J. (1978). “Local Bond-Stress-Slip Relationship

under Repeated Loading.” Mag. Concrete Res., 30(103), 62-72.

El-Metwally, S. E., and Chen, W. F. (1989). “Load-Deformation Relations for

Reinforced Concrete Sections.” ACI Struct. J., 86(2), 163-167.

Elmorsi, M., Kianoush, M. R., and Tso, W. K. (1998). “Nonlinear Analysis of

Cyclically Loaded Reinforced Concrete Structures.” ACI Struct. J., 95(6), 725-739.

Essa, Mahdi S. (1990). “The Effect of Hot Weather Concreting on Bond Strength
between Concrete and Steel.” Ph.D. Thesis, Dept. of Civil Eng., Univ. of Baghdad,
157.

Evans, R. H., and Marathe, M. S. (1968). “Microcracking and Stress-Strain Curves for
Concrete in Tension.” Mat. Struct.; Res. and Test., (RILEM, Paris), 1(1), 61-64.

Fardis, M. N., and Buyukozoturk, O. (1979). “Shear Transfer Model for Reinforced
Concrete.” J. Engrg. Mech., ASCE, 105(2), 255-275.

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.

129
Gajer, G., and Dux, P. (1989). “A Damage-Sensitive Loading Algorithm for the
Analysis of Reinforced Concrete Structures.” Comp. Struct., 32(2), 493-496.

Gangadharam, D., and Reddy, K. N. (1980). “Effect of Cover Upon the Stress-Strain
Properties of Concrete Confined in Steel Binders.” Mag. Concrete Res., 32(112), 147-
155.

Gilbert, R. I., and Warner, R. F. (1978). “Tension Stiffening in Reinforced Concrete

Slabs.” J. Struct. Div., ASCE, 104(12), 1885-1900.

Ghosh, S. K., Fanella, D. A., and Rabbat, B. G. (1996). “Notes on ACI 318-95
Building Code Requirements for Structural Concrete.” 6th. Edition, Portland Cement
Association, Skokie, Illinois, USA.

Gopalaratnam, V. S., and Shah, S. P. (1985). “Softening Response of Plain Concrete

in Direct Tension.” ACI J., 82(3), 310-323.

Grossfield, B., and Birnstiel, C. (1962). “Tests of T-Beams with Precast Webs and
Cast-in-Place Flanges.” ACI J., 59(6), 843-851.

Hanson, N. W. (1960). “Precast-Prestressed Concrete Bridges 2. Horizontal Shear

Connections.” Journal, PCA Res. & Develop. Laboratories, 2(2), 38-58, Also, PCA
Develop. Dept. Bulletin D35, (Cited according to Hofbeck et al., 1969).

Hertz, K. (1982). “The Anchorage Capacity of Reinforcing Bars at Normal and High
Temperatures.” Mag. Concrete Res., 34(121), 213-220.

Hillerborg, A., Modéer, M., and Petersson, P. E. (1976). “Analysis of Crack

Formation and Crack Growth in Concrete by Means of Fracture Mechanics and Finite
Elements.” Cement & Concrete Res., 6(6), 773-782.

Hofbeck, J. A., Ibraham, I. O., and Mattock, A. H. (1969). “Shear Transfer in

Reinforced Concrete.” ACI J., 66(2), 119-128.

Hogenstad, E. (1951). “A Study of Combined Bending and Axial Load in Reinforced

Concrete Members.” Univ. of Illinois, Eng. Experiment Station, Bulletin No. 399,
Nov., 128.

Hognestad, E. (1957). “Confirmation of Inelastic Stress Distribution in Concrete.” J.

Struct. Div., ASCE, 83(2),.

Holzer, S. M., Somers, Jr. A. E., and Bradshaw, J. C. (1979). “Finite Response of
Inelastic RC Structures.” J. Struct. Div., ASCE, 105(1), 17-33.

Hu, H. T., and Schnobrich, W. C. (1990). “Nonlinear Analysis of Cracked Reinforced

Concrete.” ACI Struct. J., 87(2), 199-207.

Hughes, B. P., and Chapman, G. P. (1966). “The Deformation of Concrete and

Microconcrete in Compression and Tension with Particular Reference to Aggregate
Size.” Mag. Concrete Res., 18(54), 19-24.

130
Hwang, S. J., Yu, H. W., and Lee, H. J. (2000). “Theory of Interface Shear Capacity
of Reinforced Concrete.” J. Struct. Engrg., ASCE, 126(6), 700-707.

Iraqi Building Code Requirements for Reinforced Concrete (Code 1/1987), Building
Research Center, Scientific Research Council, Baghdad, Republic of Iraq, 68.

Jimenez, R., Gergely, P., and White, R. N. (1978). “Shear Transfer across Cracks in
Reinforced Concrete” Ithaca (N.Y.), Cornell University, 357, Report 78-4, (Cited
according to Millard and Johnson, 1984).

Johnson, R. P. (1975). “Composite Structures of Steel and Concrete, Vol. 1- Beams,

Column, Frames and Applications in Buildings.” Constrado Monographs, Halsted
Press, John Wiley & Sons, New York, 170.

Johnson, L. J., and Nichols, J. R. (1914). “Shearing Strength og Construction, Joints

in Stems of Reinforced Concrete T-Beams, as Shown by Tests.” Transactions, ASCE,
77, 1499-1522, (Cited according to Grossfield and Birnstiel, 2962).

Kabaila, A. (1964). Discussion of “Equation for the Stress-Strain Curve of Concrete.”

by Desayi and Krishnan, ACI J., 61(9), 1227-1229.

Karayannis, C. G. (2000). “Smeared Crack Analysis for Plain Concrete in Torsion.” J.

Struct. Engrg., ASCE, 126(6), 638-645.

Kayal, S. (1984). “Finite Element Analysis of R.C. Frames.” J. Struct. Engrg., ASCE,
110(12), 2891-2908.

Kolleger, J., and Mehlhorn, G. (1990). “Material Model for the Analysis of
Reinforced Concrete Surface Structures.” Computational Mech., 6, 341-357.

Krauthammer, T., and Hall, W. J. (1982). “Modified Analysis of Reinforced Concrete

Beams.” J. Struct. Div., ASCE, 108(2), 457-475.

Kroenke, W. C., Gutzwiller, M. J., and Lee, R. H. (1973). “Finite Element for
Reinforced Concrete Frame Study.” J. Struct. Div., ASCE, 99(7), 1371-1390.

Kupfer, H. B., and Gerstel, K. H. (1973). “Behavior of Concrete under Biaxial

Stresses.” J. Engrg. Mech., ASCE, 99(4).

Kupfer, H., Hilsdorf, H. K., and Rusch, H. (1969). “Behavior of Concrete under
Biaxial Stresses.” ACI Struct. J., 66(2), 656-666.

Laible, J. P., White, R. N., and Gergely, P. (1977). “Experimental Investigation of

Seismic Shear Transfer across Cracks in Concrete Nuclear Containment Vessels.”
Reinforced Concrete Structures in Seismic Zones, Detroit, ACI Special Publication
SP-53, 203-226, (Cited according to Millard and Johnson, 1984).

Lazaro, A. L., and Richards, Jr. R. (1973). “Full Range Analysis of Concrete Frames.”
J. Struct. Div., ASCE, 99(8), 1761-1783.

131
Lin, C. S. (1973). “Nonlinear Analysis of Reinforced Concrete Slabs and Shells.”
Ph.D. Dissertation, University of California, Berkeley, Calif., UC-SESM Report No.
37-7, (Cited according to Chan, 1983).

Lin, C. S., and Scordelis, A. C. (1975). “Nonlinear Analysis of RC Shells of General

Form.” J. Struct. Div., ASCE, 101(3), 523-538.

Lutz, L. A. (1970). “Analysis of Stresses in Concrete near a Reinforcing Bar due to

Bond and Transverse Cracking.” ACI J., 67(10), 778-787.
Lutz, L. A., and Gergely, P. (1967). “Mechanics of Bond and Slip of Deformed Bars
in Concrete.” ACI J., 64(11), 711-721.

Marzouk, H. M., and Chen, Z. (1993). “Finite Element Analysis of High-Strength

Concrete Slab.” ACI Struct. J., 90(5), 505-513.

Mast, R. F. (1968). “Auxiliary Reinforcement in Concrete Connections.” J. Struct.

Div., ASCE, 94(6), 1485-1504.

Mattock, A. H. (1981). “Cyclic Shear Transfer and Type of Interface.” J. Struct. Div.,
ASCE, 107(10), 1945-1964.

Mattock, A. H., and Hawkins, N. M. (1972). “Shear Transfer in Reinforced Concrete-

Recent Research.” PCI J., 17(2), 55-75.

Mattock, A. H., Johal, L., and Chow, H. C. (1975). “Shear Transfer in Reinforced
Concrete with Moment or Tension Acting Across the Shear Plane.” PCI J., 20(4), 76-
93.

Mattock, A. H., Li, W. K., and Wang, T. C. (1976). “Shear Transfer in Lightweight
Reinforced Concrete.” PCI J., 21(1), 20-39.

Medland, I. C., and Taylor, D. A. (1971). “Flexural Rigidity of Concrete Column

Sections.” J. Struct. Div., ASCE, 97(2), 573-586.

Millard, S. G. (1983). “Shear Transfer in Cracked Reinforced Concrete.” Ph.D.

Thesis, Univ. of Warwick, 294, (Cited according to Millard and Johnson, 1984).

Millard, S. G., and Johnson, R. P. (1984). “Shear Transfer Across Cracks in

Reinforced Concrete Due to Aggregate Interlock and to Dowel Action.” Mag.
Concrete Res., 36(126), 9-21.

Mills, G. M. (1975). “A Partial Kinking Criterion for Reinforced Concrete Slabs.”

Mag. Concrete Res., 27(90), 13-22.

Mirza, S. M., and Houde, J. (1979). “Study of Bond Stress-Slip Relationships in

Reinforced Concrete.” ACI J., 76(1), 18-45.

Nam, C. H., and Salmon, C. G. (1974). “Finite Element Analysis of Concrete Beams.”
J. Struct. Div., ASCE, 100(12), 2419-2432.

132
Ng, S. F., Cheung, M. S., and Zhao, J. Q. (1993). “A Materially Nonlinear Finite
Element Model for the Analysis of Curved Reinforced Concrete Box-Girder Bridges.”
Can. J. Civ. Eng., 20, 754-759.

Ngo, D., and Scordelis, A. C. (1967). “Finite Element Analysis of Reinforced

Concrete Beams.” ACI J., 64(3), 152-163.

Nilson, A. H. (1968). “Nonlinear Analysis of Reinforced Concrete by the Finite

Element Method.” ACI J., 65(9), 757-766.

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, (Cited according to
Millard and Johnson, 1984).

Polak, M. A., and Vecchio, F. J. (1993). “Nonlinear Analysis of Reinforced-Concrete

Shells.” J. Struct. Engrg., 119(12), 3439-3462.

Rasheed, Hayder A. S. (1990). “Inelastic Behavior of Reinforced Concrete Frame

Structures.” M.Sc. Thesis, Dept. of Civil Eng., Univ. of Baghdad, 126.

Reinhardt, H. W., Cornelissen, H. A., and Hordijk, D. A. (1986). “Tensile Tests and
Failure Analysis of Concrete.” J. Struct. Engrg., ASCE, 112(11), 2462-2477.

Revesz, S. (1953). “Behavior of Composite T-Beams with Prestressed and

Unprestressed Reinforcement.” ACI J., 49(2), 585-593.

Rex, C. O., and Easterling, W. S. (2000). “Behavior and Modeling of Reinforced

Composite Slab in Tension.” J. Struct. Engrg., ASCE, 126(7), 764-771.

Russo, G., Zingone, G., and Romano, F. (1990). “Analytical Solution for Bond-Slip
of Reinforcing Bars in R. C. Joints.” J. Struct. Engrg., ASCE, 116(2), 336-355.

Sabnis, G. M. (1979). “Handbook of Composite Construction Engineering.” Van

Nostrand Reinhold Company, New York, 380.

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.

Saenz, L. P. (1964). Discussion of “Equation for the Stress-Strain Curve of

Concrete.” by Desayi and Krishnan, ACI J., 61(9), 1229-1235.

Santhanam, T. K. (1979). “Model for Mild Steel in Inelastic Frame Analysis.” J.

Struct. Div., ASCE, 105(1), 199-220.

Scanlon, A. (1971). “ Time Dependent Deflections of Reinforced Concrete Slabs.”

Ph.D. Dissertation, Dept. Civ. Engrg., University of Alberta, Edmonton, Canada,
(Cited according to Chan, 1983).

133
Seniwongse, M. (1979). “The Deformation of Reinforced Concrete Beams and
Frames up to Failure.” Struct. Eng., 57B(4), 77-81.

Smith, I. M., and Griffiths, D. V. (1998). “Programming the Finite Element Method.”
3rd. Edition, John Wiley & Sons Ltd, Chichester, England, 534.

Smith, G. M., and Young, L. E. (1956). “Ultimate Flexural Analysis Based on Stress-
Strain Curves of Cylinders.” ACI J., 53(6), 597-609.

Smith, R. G., and Orangun, C. O. (1969). “Evaluation of the Stress-Strain Curve of

Concrete in Flexure Using Method of Least Squares.” ACI J., 66(7), 553-559.

Soroushian, P., Obaseki, K., Baiyasi, M. I., El-Sweidan, B., and Choi, K. B. (1988).
“Inelastic Cyclic Behavior of Dowel Bars.” ACI Struct. J., 85(1), 23-29.

Soroushian, P., Obaseki, K., and Marikunte, S. (1991). “Analytical Modeling of

Bonded Bars under Cyclic Loads.” J. Struct, Engrg., ASCE, 117(1), 48-60.

Soroushian, P., Obaseki, K., Rojas, M., and Najm, H. S. (1987). “Behavior of Bars in
Dowel Action Against Concrete Cover.” ACI Struc. J., 84(2), 170-176.

Soroushian, P., Obaseki, K., Rojas, M., and Sim, J. (1986). “Analysis of Dowel Bars
Acting Against Concrete Core.” ACI J., 83(4), 642-649.

Standards Association of Australia (SAA), Prestressed Concrete Code, AS 1481,

Sydney, 1974.

Sturman, G. M., Shah, S. P., and Winter, G. (1965). “Effect of Flexural Strain
Gradients on Microcracking and Stress-Strain Behavior of Concrete.” ACI J., 62(7),
805-822.

Tassion, T. P., and Vintzèleou, E. N. (1987). “Concrete-to-Concrete Friction.” J.

Struct. Engrg., ASCE, 113(4), 832-849.

Tsai, W. T. (1988). “Uniaxial Compressional Stress-Strain Relation of Concrete.” J.

Struct. Engrg., ASCE, 114(9), 2133-2136.

Tulin, L. G., and Gerstle, K. H. (1964). Discussion of “Equation for the Stress-Strain
Curve of Concrete.” by Desayi and Krishnan, ACI J., 61(9), 1236-1238.

Vangsirirungruang, K. (1971). “Effect of Normal Compressive Stresses on Shear

Transfer in Reinforced Concrete” MSCE Thesis, University of Washington, Seattle,
(Cited according to Mattock and Hawkins, 1972).

Vecchio, F. J. (1989). “Nonlinear Finite Element Analysis of Reinforced Concrete

Membranes.” ACI Struct. J., 86(1), 26-35.

Vecchio, F. J. (1990). “Reinforced Concrete Membrane Element Formulation.” J.

Struct. Engrg., ASCE, 116(3), 730-750.

134
Vecchio, F. J. (1992). “Finite Element Modeling of Concrete Expansion and
Confinement.” J. Struct. Engrg., ASCE, 118(9), 2390-2405.

Vecchio, F. J., and Collins, M. P. (1982). “The Response of Reinforced Concrete to

in-Plane Shear and Normal Stress.” Publ. No. 82-03, Dept. Of Civ. Engrg., University
of Toronto, Toronto, Canada, (Cited according to Ayoub and Filippou, 1998).

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.

Walraven, J. C. (1981). “ Fundamental Analysis of Aggregate Interlock.” J. Struct.

Div., ASCE, 107(11), 2245-2270.

Walraven, J. C. (1994). “Rough Cracks Subjected to Earthquake Loading.” J. Struct.

Engrg., ASCE, 120(5), 1510-1524, (Cited according to Fronteddu et al., 1998).

Wang, P. T., Shah, S. P., and Naaman, A. E. (1978). “Stress-Strain Curves of Normal
and Lightweight Concrete in Compression.” ACI J., 75(5), 603-611.

Zeris, C. A., and Mahin, S. A. (1988). “Analysis of Reinforced Concrete Beam-

Columns under Uniaxial Excitation.” J. Struct. Engrg., ASCE, 114(4), 804-820.

Zia, P. (1961). “Torsional Strength of Prestressed Concrete Members.” ACI J.,

57(10), 1337-1360.

135
 APPENDIX A

200 kN

1600 mm

bf
E = 20000 MPa hf
ν = 0.15
H
bw = 40 mm
bf = 40 → 200 mm
bw
hf = 40 mm
H = 200 mm Cross-section

NASTRAN NASTRAN MHND

Exact Error
(3D) (2D) (Plane stress)
bf/bw 3−4
*100%
1 2 3 4 4

1 33.53834 33.50686 32.91420 33.38000 -1.395%

2 25.64483 25.60760 25.18390 25.47639 -1.148%
3 22.02729 21.96145 21.59330 21.83289 -1.097%
4 19.96280 19.85231 19.50610 19.72862 -1.128%
5 18.62590 18.47699 18.13280 18.35313 -1.200%

1 Analysis by NASTRAN using brick elements (3-dimensions)

2 Analysis by NASTRAN using membrane elements (2-dimensions)
3 Analysis by MHND using Q-4 plane stress elements (2-dimensions)
⎡ ⎤
3 ⎢
PL 12EI ⎥
4 Using the equation Δ = ⎢1 + A ⎥
48EI ⎢ G L2 ⎥
⎣ 1.2 ⎦
A-1
Y

3D-finite element idealization by NASTRAN

2D-finite element idealization by NASTRAN

A-1
 ‫ﺳﻠﻮآﻴﺔ اﻟﻘﺺ واﻟﻌﺰم ﻟﻸﻋﺘﺎب اﻟﺨﺮﺳﺎﻧﻴﺔ اﻟﻤﺮآﺒﺔ‬

‫اﻟﺨﻼﺻﺔ‪:‬‬
‫ﻟﻐ ﺮض ﻣﻌﺮﻓ ﺔ ﻣﻘ ﺪار ﺗﺤﻤ ﻞ ﻋﺘ ﺐ ﺧﺮﺳ ﺎﻧﻲ ﻣﺮآ ﺐ و اﺳ ﺘﺠﺎﺑﺘﻪ اﻟﻼﺧﻄﻴ ﺔ ﻟﻸﺣﻤ ﺎل‬
‫اﻟﺴﺘﺎﺗﻴﻜﻴﺔ اﻟﻤﺆﺛﺮة ﻋﻠﻴﻪ ‪ ،‬ﺗﻢ ﺗﺤﻠﻴﻠﻪ ﺑﻮاﺳ ﻄﺔ اﻟﺤﺎﺳ ﺒﺔ اﻹﻟﻜﺘﺮوﻧﻴ ﺔ و ﺑﺎﺳ ﺘﻌﻤﺎل ﻃﺮﻳﻘ ﺔ‬
‫اﻟﻌﻨﺎﺻﺮ اﻟﻤﺤﺪدة‪ .‬ﻋﺪة ﻋﻼﻗﺎت ﻻﺧﻄﻴﺔ اﻋﺘﻤﺪت ‪ ،‬آﺎﻟﻌﻼﻗﺔ اﻟﻼﺧﻄﻴ ﺔ ﺑ ﻴﻦ اﻹﺟﻬ ﺎد و‬
‫اﻻﻧﻔﻌﺎل ﻟﻤﺎدة اﻟﺨﺮﺳﺎﻧﺔ ‪ ،‬آﺬﻟﻚ ﻟﻤﺎدة اﻟﺤﺪﻳﺪ ‪ ،‬اﻟﻌﻼﻗﺔ ﺑﻴﻦ ﻗﻮة اﻟ ﺮﺑﻂ و اﻻﻧ ﺰﻻق ﺑ ﻴﻦ‬
‫ﺣﺪﻳ ﺪ اﻟﺘﺴ ﻠﻴﺢ و ﻣ ﺎدة اﻟﺨﺮﺳ ﺎﻧﺔ اﻟﻤﺤﻴﻄ ﺔ ﺑ ﻪ ‪ ،‬اﻟﻌﻼﻗ ﺔ ﺑ ﻴﻦ ﻗ ﻮى اﻟﻘ ﺺ و اﻻﻧ ﺰﻻق‬
‫اﻟﺤﺎﺻﻞ ﻋﻨﺪ ﺳﻄﺢ اﻟﺘﻤﺎس ﺑﻴﻦ اﻟﻌﺘﺐ اﻟﻤﺴﺒﻖ اﻟﺼﺐ و ﺧﺮﺳ ﺎﻧﺔ اﻷرﺿ ﻴﺔ اﻟﻤﺼ ﺒﻮﺑﺔ‬
‫ﻣﻮﻗﻌﻴًﺎ ‪ ،‬اﻟﺘﺄﺛﻴﺮ اﻟﻮﺗﺪي ﻟﺮاﺑﻄﺎت اﻟﻘﺺ‪.‬‬
‫ﺑﺎﻟﻨﺴ ﺒﺔ ﻟﻠﺨﺮﺳ ﺎﻧﺔ ﺗ ﻢ اﻋﺘﻤ ﺎد ﻧﻤ ﻮذج اﻟﻌﻼﻗ ﺎت اﻟﺘﻜﻮﻳﻨﻴ ﺔ ﻟﻠﻤ ﺎدة اﻟﻤﺨﺘﻠﻔ ﺔ‬
‫ﺑﺎﻻﺗﺠﺎهﻴﻦ اﻟﻤﺘﻌﺎﻣﺪﻳﻦ‪ .‬و ﺗﻢ ﺗﻤﺜﻴ ﻞ اﻟﺘﺸ ﻘﻖ ﻓ ﻲ اﻟﺨﺮﺳ ﺎﻧﺔ ﺑﻄﺮﻳﻘ ﺔ اﻟﺸ ﻘﻮق اﻟﻤﻮزﻋ ﺔ‬
‫اﻟ ﺪاﺋﺮة‪ .‬ﺗ ﻢ اﻋﺘﺒ ﺎر ﺣﺪﻳ ﺪ اﻟﺘﺴ ﻠﻴﺢ ذو ﻣﻮاﺻ ﻔﺎت أﺣﺎدﻳ ﺔ اﻻﺗﺠ ﺎﻩ ‪ ،‬و ﺑﺎﺗﺠ ﺎﻩ ﻗﻀ ﺒﺎن‬
‫اﻟﺘﺴﻠﻴﺢ‪.‬‬
‫ﻟﺘﻤﺜﻴ ﻞ اﻟﺨﺮﺳ ﺎﻧﺔ ﺗ ﻢ اﺳ ﺘﻌﻤﺎل ﻋﻨﺼ ﺮ اﻹﺟﻬ ﺎد اﻟﻤﺴ ﺘﻮي ﻋﻨ ﺪ اﻟﺘﺤﻠﻴ ﻞ اﻟﻌﺘ ﺐ‬
‫اﻟﻤﺮآ ﺐ ﺑﻄﺮﻳﻘ ﺔ اﻟﻌﻨﺎﺻ ﺮ اﻟﻤﺤ ﺪدة‪ .‬ﺣﺪﻳ ﺪ اﻟﺘﺴ ﻠﻴﺢ ﻣُﺜ ﻞ ﺑﻌﻨﺼ ﺮ أﺣ ﺎدي اﻻﺗﺠ ﺎﻩ‪ .‬و‬
‫ﻟﺘﻤﺜﻴ ﻞ اﻟ ﺮﺑﻂ واﻻﻧ ﺰﻻق و اﻟﺘ ﺎﺛﻴﺮ اﻟﻮﺗ ﺪي ﺗ ﻢ اﺳ ﺘﻌﻤﺎل ﻋﻨﺼ ﺮ وﺻ ﻠﻲ ذو ﻧﺎﺑﻀ ﻴﻦ‬
‫ﻣﺘﻌﺎﻣﺪﻳﻦ‪ .‬ﻗﻮى اﻟﻘﺺ و اﻻﻧﺰﻻق ﻣُﺜﻠﺖ ﺑﻌﻨﺼﺮ اﻟﺴﻄﺢ اﻟﺒﻴﻨﻲ‪.‬‬
‫و ﻹﺟ ﻞ اﻟﺘﺤﻘ ﻖ ﻣ ﻦ ﺻ ﺤﺔ و ﺻ ﻼﺣﻴﺔ اﻟﻌﻼﻗ ﺎت اﻟﺘﻜﻮﻳﻨﻴ ﺔ اﻟﻤﺴ ﺘﻌﻤﻠﺔ ﻟﺘﻤﺜﻴ ﻞ‬
‫اﻟﻤ ﻮاد و اﻟﺘ ﺄﺛﻴﺮات اﻟﻤﺨﺘﻠﻔ ﺔ ‪ ،‬و اﻟﺒﺮﻧ ﺎﻣﺞ اﻟﻤﻜﺘ ﻮب ﻟﺘﺤﻠﻴ ﻞ ﻋﺘ ﺐ ﺧﺮﺳ ﺎﻧﻲ ﻣﺴ ﻠﺢ‬
‫ﻣﺮآﺐ ‪ ،‬ﺗﻤﺖ اﻟﻤﻘﺎرﻧﺔ ﻣﻊ اﻟﻨﺘﺎﺋﺞ اﻟﻤﺨﺘﺒﺮﻳﺔ اﻟﻤﻨﺸﻮرة ﻟﺒﻌﺾ اﻟﻨﻤﺎذج اﻟﺨﺮﺳﺎﻧﻴﺔ‪.‬‬
‫ﺗﻢ ﻋﻤﻞ ﻣﻘﺎرﻧﺔ ﺑﻴﻦ اﻟﻨﺘﺎﺋﺞ اﻟﻤﺴﺘﺤﺼﻠﺔ ﻓﻲ هﺬﻩ اﻟﺪراﺳ ﺔ ﻣ ﻊ ﻣ ﺎ ه ﻮ ﻣﺘ ﻮﻓﺮ ﻣ ﻦ‬
‫ﻧﺘ ﺎﺋﺞ ﻟﻠﻔﺤ ﻮص اﻟﻤﺨﺘﺒﺮﻳ ﺔ ﻻﻋﺘ ﺎب ﺧﺮﺳ ﺎﻧﻴﺔ ﻣﺮآﺒ ﺔ‪ .‬اﻟﻨﺘ ﺎﺋﺞ اﻟﻤﺴﺘﺤﺼ ﻠﺔ آﺎﻧ ﺖ‬
‫ﻣﻘﺎرﺑﺔ ﻟﻠﻨﺘﺎﺋﺞ اﻟﻤﺨﺘﺒﺮﻳﺔ‪.‬‬
‫ﺗﻢ ﻋﺮض دراﺳﺔ ﻣﻘﺎرﻧﺔ ﺗﺘﻌﻠﻖ ﺑﺘﺤﻤﻞ اﻟﻌﺘﺐ اﻟﺨﺮﺳﺎﻧﻲ اﻟﻤﺮآﺐ ﻟﻠﻘﺺ وﻋ ﺰم‬
‫اﻻﻧﺤﻨﺎء‪.‬‬