Professional Documents
Culture Documents
net/publication/315743987
CITATIONS READS
0 316
1 author:
Mohanned Al Gharawi
University of Missouri
6 PUBLICATIONS 2 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Mohanned Al Gharawi on 02 April 2017.
A THESIS
SUBMITTED TO THE COLLEGE OF ENGINEERING
OF THE UNIVERSITY OF BAGHDAD IN PARTIAL
FULFILLMENT OF THE REQUIREMENTS FOR
THE DEGREE OF MASTER OF SCIENCE
IN CIVIL ENGINEERING
(STRUCTURES)
BY
MOHANNED IBRAHIM MOHAMMED HUSSAIN
AL-GHARRAWI
2004
(SUPERVISOR CERTIFICATE)
I certify that the preparation of this thesis titled “Finite Element Analysis of Post-
Tensioned Concrete Box Girders” accomplished by “Mohanned Ibrahim
Mohammed” was made under my supervision at the Department of Civil
Engineering, University of Baghdad as a partial fulfillment of the requirement for
degree of Master of Science in Civil Engineering (Structure).
Signature:
Name: Prof. Dr. Husain Mohammed Husain
Date:
In view of the available recommendations, I forward this thesis for debate by the
examination committee.
Signature: Signature:
Name: Prof. Dr. Husain M. Husain Name: Dr. Ihsan A. S. Al-Shaarbaf
(Supervisor) (Member)
Date: / / 2004 Date: / / 2004
Signature: Signature:
Name: Dr. Nazar K. Al-Oukaili Name: Prof. Dr. Khalid S. Mahmoud
(Member) (Chairman)
Date: / / 2004 Date: / / 2004
Signature:
Name: Prof. Dr. Ali Al-Kiliddar
Dean, College of Engineering
University of Baghdad
Date: / / 2004
ACKNOWLEDGEMENT
Firstly, all thanks and praise be to GOD who enabled me to complete this
work.
I would like to express my deepest gratitude and gratefulness to
Prof. Dr. Hussain M. Hussain the supervisor for his valuable guidance,
encouragement, constructive suggestions, and his assistance throughout
the final preparation of this work.
Sincere thanks to my colleague the Ph. D. student Wamidh A'Jel
for his help.
Special thanks to my colleagues and friends Mohammed Najim,
Nassir, Ghaidaq, Ahmed Adnan, Ali Ahmed Salman and Sabah for
their encouragements.
Special thanks to my real friends Amjad, Najah, Adil, Dhirgham
and Ahmed Khudher for their helps to complete this thesis.
Cordial thanks are also to my colleagues Miss Ruwayda, Miss
Hind and Mirs. Rana for their encouragements throughout all stage of
this study.
Finally, I would like to express my deepest gratitude to all who
gave helpful hand throughout this study, especially to my family for their
encouragement and assistant during this work.
MOHANNED
2004
I
ABSTRACT
II
CONTENTS
SUBJECT PAGE
Acknowledgement I
Abstract II
Content III
List of tables VI
Notations. XI
III
3.3.2 Reinforcing Bar Representation 28
3.3.3 Bond and Bond Slip Representation 31
3.4 Numerical Integration Technique 33
3.5 Nonlinear Solution Techniques 35
3.6 General Nonlinear Solution Procedure 36
3.6.1 Incremental Technique 36
3.6.2 Iterative Technique 37
3.6.3 Combined Incremental-Iterative Technique 39
3.7 Convergence Criteria 41
IV
5.2.2 Anchoring Losses 63
5.3 Algorithm for Use with the Finite Element Analysis 65
5.3.1 Basic Assumption 65
5.3.2 Geometry of the Tendon and Variation of the Prestressing
Forces 65
5.3.3 Element Local Loads Due to Prestressing 67
5.3.4 Vector of Primary Nodded Forces of the Element 68
5.4 Algorithm for Use the Lin's Method 70
5.5 Computer Program 71
REFERENCES 135
APPENDIX A
APPENDIX B
V
LIST OF TABLES
TABLE
TITLE PAGE
NO.
6.2 Variation longitudinal stresses at mid span for Pu/P=0.250 and 0.875
81
VI
LIST OF FIGURES
FIG.
TITLE PAGE
NO.
1.1 Development of the box-girder cross-section 2
VII
4.7 Stress-strain relationship of steel bars 60
6.12 Variation of transverse stresses in the top surface of the top flange 84
Variation of longitudinal stresses in the bottom surface of the bottom
6.13
flange 85
6.14 Variation of transverse stresses in the bottom surface of the bottom flange
86
6.15 Variation of longitudinal stresses in the outer surface of the web 87
VIII
Simulated Ontario Highway Bridge Design trucks (OHBDC) for two-cell
6.21
box-girder bridge 95
Analytical and Experimental Load-Midspan Deflection Curves for two-
6.22
cell box-girder bridge 95
Deflected shape of the bridge at various load levels for two-cell box-girder
6.23
bridge 96
Longitudinal normal stress on the top slab at midspan for two-cell box-
6.24
girder bridge 96
Longitudinal normal stress on the top slab at quarterspan for two-cell
6.25
box-girder bridge 97
Variation of concrete longitudinal normal stress along the Top Slab for
6.26
two-cell box-girder bridge 97
Longitudinal normal stress variation across the section at midspan for
6.27
two-cell box-girder bridge 98
Longitudinal normal stress variation across the section at quarterspan for
6.28
two-cell box-girder bridge 98
6.29 Variation of longitudinal stresses in the top surface of the top flange 101
6.30 Variation of transverse stresses in the top surface of the top flange 102
Variation of longitudinal stresses in the bottom surface of the bottom
6.31
flange 103
6.32 Variation of transverse stresses in the bottom surface of the bottom flange 104
6.33 Variation of longitudinal stresses in the outer surface of the right web 105
6.34 Variation of transverse stresses in the outer surface of the right web 106
6.35 Variation of longitudinal stresses in the inner surface of the right web 107
6.36 Variation of transverse stresses in the inner surface of the right web 108
6.37 Variation of longitudinal stresses in the outer surface of the left web 109
6.38 Variation of transverse stresses in the outer surface of the left web 110
6.39 Variation of longitudinal stresses in the inner surface of the left web 111
6.40 Variation of transverse stresses in the inner surface of the left web 112
Structural details and cable profile of prestressing box-girder bridge with
6.41
inclined webs 114
6.42 Finite element mesh for the bridge model before and after loading 115
6.45 Variation of longitudinal stresses in the top surface of the top flange 120
IX
6.46 Variation of transverse stresses in the top surface of the top flange 121
Variation of longitudinal stresses in the bottom surface of the bottom
6.47
flange 122
6.48 Variation of transverse stresses in the bottom surface of the bottom flange
123
6.49 Variation of longitudinal stresses in the outer surface of the left web 124
6.50 Variation of transverse stresses in the outer surface of the left web 125
6.51 Variation of longitudinal stresses in the inner surface of the left web 126
6.52 Variation of transverse stresses in the inner surface of the left web 127
6.53 Variation of longitudinal stresses in the outer surface of the right web 128
6.54 Variation of transverse stresses in the outer surface of the right web 129
6.55 Variation of longitudinal stresses in the inner surface of the right web 130
6.56 Variation of transverse stresses in the inner surface of the right web 131
X
NOTATIONS
The major part of the symbols used in the text is listed below. Others are defined they
first appear.
General Symbols
d, Differential symbols
, det. Determinate of matrix or absolute value
{} Vector
[] Matrix
Scalar
Cp Plasticity coefficient
dλ Plastic multiplier
f Yield function
XI
I1' First strain invariant
J Jacobian
u, v, w Displacement components
V Volume
γ Shear strain
XII
1 , 2 , 3 Shear retention parameters
ε Strain
εe Elastic strain
Strain corresponding to peak Uniaxial concrete compressive
εo
stress
o Total strain corresponding to the parabolic part of uniaxial
compressive strength stress-strain curve
εp Plastic strain
Natural coordinates for tendon
λ Reduction factor
υ Poisson's ratio
σ Stress
o Effective stress at onset of plastic deformation
Effective stress
τ Shear stress
12 , 13 , 23 Shear stresses in principal planes
12 , 13 , 23 Shear strains in principal planes
Matrices
XIII
[K] Stiffness matrix of concrete element
Vectors
XIV
Chapter One Introduction
Chapter One
INTRODUCTION
1.1 General
Box-girder bridges have been used extensively over the past decades, especially
in urban areas where aesthetics is a major principle in construction. The box-
girder cross section evolved structurally from the hollow cell deck bridge or the
T-beam bridge. The first box-girder cross-sections possessed deck slabs that
cantilevered out only slightly from the main box portion Fig. (1.1 a-e)[57]. With
prestressed concrete, the length of the cantilever could be increased. The high
formwork costs caused a reduction in the number of cells Fig. (1.1 f, g)[57]. In
order to reduce the construction loads to the minimum possible or to require
only one longitudinal girder in the working state even with multiple traffic lanes,
the one-cell built-up cross-section constructed in modular fashion emerged as
the last development, Fig. (1.1 h)[57].
The relative economy of the box-girder bridges contributed greatly to its
popularity, as it has its relatively slender and unencumbered appearance. Some
proponents of the box-girder bridge have claimed its smooth soffit to be very
desirable in urban areas for reasons of esthetics. The structural simplicity of the
box-girder bridges, particularly in continuous structures of medium to long
spans, has been well demonstrated. The efficiency of the cross-section for
positive and negative longitudinal bending moments, as well as torsional
moments is apparent even to casual observer. Special advantages for this mode
of concrete bridge construction include the low depth-to-span ratio that can be
economically achieved together with the ease with which variable conditions of
bridge width, superelevation and curvature, both vertical and horizontal can be
1
Chapter One Introduction
a) c)
b) d)
e)
f)
g)
h)
2
Chapter One Introduction
3
Chapter One Introduction
4
Chapter One Introduction
conditions of the structure and any significant interaction among them. The
finite element method employs an assemblage of discrete one, two and three-
dimensional members to represent the structure. The structure is divided into
elements that are only connected at their nodes, Fig. (1.2), which possess an
appropriate number of degrees of freedom[57].
Recently the development of two and three-dimensional finite elements
capable of simulating the actual nonlinear behavior of prestressed and reinforced
concrete after the formation of cracks has been a considerable progress.
Thus, the finite element method may be seen to be very general in
application and, indeed, for difficult bridge deck problems it is sometimes the
only valid form of analysis[30].
applied loading
top flange
finite element
nodal points
web
fine mesh
bottom flange
coarse mesh
5
Chapter One Introduction
6
Chapter Two Review of Literature
Chapter Two
REVIEW OF LITERATURE
2.1 Introduction
During the past three decades, interest in the nonlinear analysis of concrete
structures has increased steadily. The nonlinear finite element method of
structural analysis has been ensured the wide use of both reinforced and
prestressed concrete structures. This is because of the development of a
relatively powerful analysis technique implemented on electronic computers.
In this chapter, a brief review of literature of the main contributions in the
field of the analysis of box-girder bridges is presented. However, a literature
review for more specific area will be included in the following sections related
to this topic. This is believed to give a clearer picture to the several topics which
this study deals with.
7
Chapter Two Review of Literature
8
Chapter Two Review of Literature
9
Chapter Two Review of Literature
warping and shear lag are calculated in the second and third stages, respectively.
These stresses arising in each stage of analysis are then superposed to determine
the overall response of the box section to the applied loading. The results of only
one example were presented for illustrating the application of the computer
programs. They agree precisely with Maisel’s detailed hand-calculation results.
In 1990, Abdullah M. A. and Abdul-Razzaq A. A.[1] used the same
examples that were analyzed by Scordelis et al[59] but they used the finite strip
method for the analysis of a prestressed concrete box-girder bridge using higher
order bending and in-plane strips. The auxiliary nodal line technique was
adopted for both bending and in-plane strips. A flexibility procedure was
incorporated for the solution of continuous box-girder bridges when redundant
forces are induced by intermediate supports. Extension of the finite strip method
to the analysis of prestressed forces was described. The higher order finite strip
gives good results for the analysis of both simply supported and continuous
prestressed box-girder bridges compared with the elasticity solution.
In 1993, Abdullah M. A. and Abdul-Razzaq A. A.[2] applied a refined
finite strip method for the analysis of continuous box-girder bridges using higher
order bending and inplane strips. The auxiliary nodal line technique has been
adopted for both bending and inplane actions. A flexibility approach was used
for the solution of continuous box-girder bridges. Sixth order bending strip was
combined with third order inplane strip to form a flat shell strips. At strip
boundaries, displacement parameters were similar to those of lower order finite
strip. Static condensation technique was adopted to eliminate the auxiliary nodal
line parameters for both bending and inplane actions at strip level. Therefore, the
solution is marginally higher than the lower finite strip although extra manual
effort is necessary in developing the condensed stiffness and force matrices at
the strip level. Afterwards Abdul-Razzaq A. A. developed and used a higher
conical frustum shell strip to analyze curved box girders[4].
10
Chapter Two Review of Literature
Later, a finite strip method for the elastic analysis of simply supported and
continuous prismatic cellular box-girder bridges was presented by Abdul-
Razzak[3]. Higher order finite strip method with auxiliary nodal line technique
has been used for the analysis. A flexibility approach is incorporated.
11
Chapter Two Review of Literature
12
Chapter Two Review of Literature
rectangular plate element have been used for webs and the 12-degrees of
freedom annular plate bending element for flanges. Good agreement had been
obtained with both curved folded plate theory and finite strip method.
The same authors made a model study in 1976[25] to investigate the effect
of different parameters in the linear elastic region. A three-dimensional finite
element program was developed. Plexiglass was chosen as the model material.
Three kinds of elements were used; for flanges the annular element with corner
nodes, cylindrical elements for webs, and rectangular elements for diaphragms.
The results of three-dimensional finite element program developed for this
purpose were found to be in good agreement with the experimental results for
deflections, radial stresses, and tangential stresses.
In 1977, Scordelis A. C. and Larsen P. K[61][62] made a model study on
multi-cell reinforced concrete box-girder. The model construction was similar to
that used for the prototype in the field. In that study theoretical and experimental
results were compared for reactions, deflections, and moments due to point
loads. Three different analytical methods were used. The computer programs
associated with these three analytical solutions are titled SAP, CURDI, and
CELL. The SAP program was used to analyze the bridge as an idealized three-
dimensional frame using one-dimensional members. A total of 24 straight beam
elements were used to model the curved bridge. In each beam member’s section
properties represent the entire four-cell bridge cross section. In CURDI analysis
the bridge structure was modeled as an assemblage of flat and curved plate
elements spanning between the end supports. The solution was based on the
harmonic finite strip analysis. The CELL program is a general finite element
program for cellular bridges, in which the structure is analyzed as a three-
dimensional assemblage of plane plate elements representing top and bottom
flanges and vertical webs. Both CELL and CURDI accounted for both the
membrane and plate bending action in the elements comprising the structure.
From this study it was obvious that SAP program can only give an indication of
13
Chapter Two Review of Literature
14
Chapter Two Review of Literature
spacing and curvature affect the moment distribution and peak longitudinal in-
plane stresses even under symmetric loading. Qualitatively, an increase in
curvature or increase in web spacing increases the difference in flexural stiffness
of the inner and outer halves of the cross section. A slight reduction of the
maximum stress occurred with introduction of intermediate diaphragms.
15
Chapter Two Review of Literature
16
Chapter Two Review of Literature
17
Chapter Two Review of Literature
approach, based on the tangent stiffness method, to solve the nonlinear problem
in a series of incrementally linear analysis. For comparison, the load-deflection
curves show that the predicted and measured strength of the bridges are in
excellent agreement for the single-cell box while for the two-cell box, the
difference being less than 4.3 %. However, the overall analytical response is
slightly stiffer than the experimental response.
18
Chapter Two Review of Literature
θ y , My
Z, w, Pz
σy
θz , M z τxy
σx
τxy midsurface
X, u, Px
shell layers θ x , Mx
Fig. (2.6) Typical 8-noded, 6 D.O.F. layered isoparametric flat shell element.
19
Chapter Two Review of Literature
top surface
η
ζ midsurface
In the present study, the 20-noded isoparametric brick element has been
used to model the concrete, each node has three transition degrees of freedom.
The reinforcing bars have been idealized as axial members within the concrete
elements. Bond slip between the steel and concrete is considered. The equivalent
load method is used to compute the force applied by the tendon upon structure.
The program is extended to include the prestressing effects. The program is
applied to several prestressed box-girder bridge.
20
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
Chapter Three
FINITE ELEMENT FORMULATION
3.1 Introduction
The finite element method is a popular numerical technique that has been widely
used by engineers and scientists, besides it is a very powerful and modern
computational tool. This technique has the potential to play an increasingly
important role in all areas of reinforced and prestressed concrete research,
design, and analysis.
The finite element method consists of replacing the actual structural
system being analyzed by an assemblage of finite elements. These elements may
be one, two or three-dimensional, each of them is bounded by a defined number
of nodes. These nodes represent the points of connection between adjoining
elements.
21
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
e [ x y z xy yz zx ] T
…(3.6)
D is the constitutive matrix and e , is the strain vector given by:
u
x x 0 0
v
x 0 0
y y
y w u
z z 0
0
z
e u v v …(3.7)
xy
0 w
yz y x y x
v w
zx 0
z y z y
u w
0
z x z x
Substitution of Eq. (3.3) into Eq. (3.5) gives the stress-displacement
relationship:
e DB ae …(3.8)
The principle of virtual displacements of a deformable body is used to
establish the governing equations of static equilibrium. It states that,” if a
general structure in equilibrium is subjected to a system of small virtual
22
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
By making use of Eqs. (3.10) and (3.12), Eq. (3.9) may be expressed as:
T
a B D B dVeae N be dVe N t e ds e 0
T T T
n Ve n Ve se
…(3.14)
where n is the total number of the elements of the discrete system.
Since the vector of the virtual nodal displacements, a is arbitrary, the
T
23
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
where K is the stiffness matrix of the element assemblage and it is given by:
24
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
where Ni , , is the shape function at the i-th node and ui , v i and w i are
the corresponding nodal displacements. The shape functions for the 20-noded
brick element are given in Table (3.1).
0.125(1+ξξ i)(1+ηηi)
Corners ±1 ±1 ±1
(1+ζζi)( ξξ i+ ηηi+ ζζi-2)
Mid-Side
0 ±1 ±1 0.25(1-ξ2)(1+ηηi)(1+ζζi)
Nodes
Mid-Side
±1 0 ±1 0.25(1+ξξ i)(1-η2)(1+ζζi)
Nodes
Mid-Side
±1 ±1 0 0.25(1+ξξ i)(1+ηηI)(1-ζ2)
Nodes
ξi, ηi, ζi denote the coordinates of i-th node.
25
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
N i N i
x
N i 1 N i
J …(3.22)
y
N i N i
z
26
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
…(3.25)
Numerical integration (Guass quadrature) is used to evaluate the element
stiffness matrix K e given by Eq. (3.25).
27
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
Elements with three degree of freedom at each nodal point such as beam
elements subjected to axial, shear and bending forces can also be used. The
major advantages of this type of representation are its simplicity and ability to
represent the bond slip and dowel action phenomena. A major disadvantage of
this type is the extremely expensive computation time since a large number of
elements are implemented.
3- Embedded Representation:
This approach is often used in connection with high-order isoparametric
elements. In this approach perfect bond (or specified bond-slip relation) is
assumed to occur between the reinforcing bars and the surrounding concrete.
This assumption of compatibility of displacements and strains between the steel
and concrete allows the reinforcing bars to be treated as integral parts of the
element[68]. The stiffness of steel bars is added to that of the concrete to obtain
the global stiffness matrix of the element. Reinforcing bars are assumed to be
capable of transmitting axial forces only. It is further assumed that such bars are
restricted to be parallel to the local coordinate lines , and of the brick
element.
In the current study, the embedded representation is adopted. The
derivation of the stiffness matrix for a bar parallel to the local coordinate axis
(for example) is represented in this section. A similar derivation can be used for
bars parallel to and axes.
A bar is considered lying parallel to the local coordinate axis with
c and c , Fig. (3.2.c). The shape functions of the brick element can be
used to represent the displacements of the bar[9]:
20
u Ni ( ).ui
i 1
20
v Ni ( ).vi …(3.26)
i 1
20
w Ni ( ).wi
i 1
28
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
where,
h C 12 C 42 C 62 …(3.28)
and
x 2 x y x z
C1 ( ) ,C 2 ( ), C 3 ( )
…(3.29)
y 2 x z z 2
C4 ( ) ,C 5 ( ) ,C6 ( )
In Eq. (3.16), dV can be written as: dV As .dx As .h .d . Finally the
stiffness matrix of the embedded bar K e can be expressed as:
1
K e As B T D B .h. d …(3.30)
1
where, As is the cross sectional area of the bar, and D is the constitutive
matrix that represents the modulus of elasticity of the steel bar. Similar
derivation can be used for bars parallel to and axes.
29
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
a-Distributed representation
c-Embedded representation
30
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
31
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
The present bond and bond slip formulation is based on the experimental
results of Nilson[48]. In the study of bond slip, Nilson found that the bond slip
relationship depends on the compressive strength of concrete as well as the
distance from loaded face (or face of crack), C. The results of Nilson are
illustrated in Fig. (3.3).
In this study, the bond-slip curve with C=152 mm (6 in.) is used. This
curve represents the upper bound of the peak bond stress of experimental results
of Nilson. Two polynomials are used to describe this curve, one for ascending
portion, and the other for the descending part, as proposed in Refs.[7] and [8]:
u
0.083 ( 7.5 3 25 2 27.5 ) for 0 1 …(3.31)
f c
u
0.083 ( 2.5 3 15 2 22.5 ) for 1 …(3.32)
f c
where u = bond stress, MPa.
, normalized slip.
p
= slip, mm, and p = slip at peak bond stress, mm.
To obtain the bond stiffness Kb, Eq. (3.31) and Eq. (3.32) are
differentiated with respect to the slip :
f c
Kb
u
0.083
p
22.5 2 50 27.5 for 0 1 …(3.33)
f c
Kb
u
0.083
p
7.5 2 30 22.5 for 1 …(3.34)
32
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
Slip (mm)
direction respectively, w i , w j and w k are the weight factors of the i-th, j-th and
33
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
34
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
Table (3.2) Weight and Location of Sampling Points in the 27-Point Rule.
Sampling Location Coordinates
Weight
Integration Rule Points ξ η ξ
1,3,5,7,19,21,23,
The 27-Point
35
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
Ultimate Load
Load
A
Deformation
Fig. (3.5) Typical structure response.
36
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
where K is the stiffness matrix at the beginning of the increment, a and
P represent increments of displacements and loads respectively.
The size of error resulting from this piecewise linearization process
mainly depends on the size of the load increment. The accumulation of this error
over several increments results in an incorrect solution. The incremental process
is repeated until the failure load has been reached. Fig. (3.6.a) shows how such a
solution can deviate from the true one.
37
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
where P is the external applied load vector, fi is the internal nodal load
vector at iteration i, and i are the current stresses at iteration i. The
procedure is repeated until the increments of displacements or the unbalanced
forces become sufficiently close to a preselected criterion, Fig. (3.6.c).
The iterative type of technique is not suitable for tracing the nonlinear
equilibrium path because it fails to produce information about intermediate stage
of loading.
f f
3
2
a a
(a) (b)
f
1 2 3 …. n
a
(c)
Fig. (3.6) Basic techniques for solving nonlinear equations.
a- Incremental technique. b- Direct iteration.
c- Newton-Raphson method (conventional and modified).
38
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
39
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
second iteration, (the KT2 method), so that the nonlinear effects are more
accurately represented in the stiffness matrix[68].
These methods are generally more powerful than the standard
(or conventional) Newton-Raphson method since they involve fewer stiffness
matrix reformulation and inversion. However, convergence is slower and a large
number of iterations are required to achieve a converged solution. This is
particularly true for an increment of loading at which a sudden softening may
occur due to cracking, yielding or substantial nonlinear behavior of concrete in
compression. In order to make the modified methods more effective at loading
stages in which slow convergence occurs, stiffness matrix may be updated more
than once within such an increment. The developed program incorporates a
modified Newton-Raphson method (KT2a), in which the stiffness matrix is
updated at the 2nd, 12th, 22th, …etc. iterations of each increment of loading.
40
Chapter Three Formulation of Finite Element & Non-Linear Solution Technique
r a T
.r a
toler …(3.43)
P P T
41
Chapter Four Modeling of Material Properties
Chapter Four
MODELING OF MATERIAL PROPERTIES
4.1 Introduction
The most powerful and general solution tool for the nonlinear analysis of
reinforced concrete and prestressed structures is undoubtedly the finite element
method. It is a valuable solution tool, which can be used to predict the load-
deflection response, failure load, stresses and strains at different stages of
loading and crack patterns. Thus, by using the finite element modeling of
reinforced and prestressed concrete members, effects of different parameters on
the structural behavior can be studied. Accurate modeling for reinforced
concrete material is often one of the major factors in enhancing the analysis
results. This is because reinforced concrete is a complicated composite material
and its behavior cannot be accurately predicted without considering the
constitutive relations of its constituents (concrete and reinforcement).
42
Chapter Four Modeling of Material Properties
I- Elastic Models
Elastic behavior of a material is assured when after unloading, no residual
strains are retained, and the material returns to its original shape. The stress-
strain relations of elastic models may be either linear or nonlinear elastic
behavior.
I-1 Linear Elastic Models
In the linear elastic models, stress-strain relations for uncracked and cracked
concrete are based on the theory of linear elasticity. The linear elastic models
can be quite accurate for structures where the tensile cracking is the major cause
of the nonlinear behavior. However, these models fail to identify inelastic
deformation. This phenomenon becomes apparent when the material experience
unloading. This can be improved by introducing the nonlinear elasticity models.
I-2 Nonlinear Elastic Models
Nonlinear elasticity models are based on two different approaches. These are the
total “hyperelastic” and incremental “hypoelastic” stress-strain formulation. In
the total stress-strain models, the current state of stress is assumed to be
uniquely expressed as a function of the curved state of strain. This type of
formulation is reversible and path-independent which is not true for concrete in
general. When concrete exhibits unloading these models fail to predict the
inelastic deformation. The incremental stress-strain models are used to describe
the behavior of materials in which the state of stress depends on the current state
of strain and on the stress path followed to reach this state. This type of
formulation is incrementally reversible and path dependent and therefore, it
43
Chapter Four Modeling of Material Properties
44
Chapter Four Modeling of Material Properties
C p 1.0
f c
E
C p f c 1
o c
cu
o
Fig. (4.1) Uniaxial stress-strain curve for concrete [9].
where and are material parameters, I1 is the first stress invariant given by:
I1 x y z …(4.2)
and:
1
J 2 x2 2y z2 x y y z z x xy
2
2yz zx
2
…(4.3)
3
and o is the equivalent effective stress at the onset of plastic deformation and
this can be determined from uniaxial compression test as:
45
Chapter Four Modeling of Material Properties
o C p f c …(4.4)
46
Chapter Four Modeling of Material Properties
F C .I 1 C .I 1 2 3 J 2 …(4.10)
where represents the stress level at which further plastic deformation will
occur and it is termed as the effective stress or the equivalent uniaxial stress at
that level.
In the present model, a parabolic stress-strain curve is used for the
equivalent uniaxial stress-strain relationship beyond the limit of elasticity,
to when the peak compressive stress is reached, where then a perfectly plastic
response is assumed to occur. Returning to Fig. (4.1), the equivalent uniaxial
stress-strain curve is shown at the various stages of behavior.
The equivalent stress-strain relationships at various stages are:
1- Elastic stage:
E . c for C p . f c …(4.11)
47
Chapter Four Modeling of Material Properties
f c f
o o C p . or 0 2( 1 C p ) c …(4.14)
E E
Based on experimental data, the value of the plasticity coefficient ( C p )
was set to equal to 0.3 for normal strength concrete and 0.5 for high strength
concrete.
In order to derive the relationship between the effective stress and the
effective plastic strain, the total strain c , is decomposed into its elastic and
plastic components as:
c e p …(4.15)
e …(4.16)
E
By substituting Eqs. (4.15) and (4.16) into Eq. (4.12), the effective stress-plastic
strain relation can be expressed as:
C p f c E p 2 E 2 0 p …(4.17)
d 0
H E 1.0 …(4.18)
d p 2 p
48
Chapter Four Modeling of Material Properties
that the plastic strain increment can be determined for a given stress increment.
The associated flow rule has been widely used for concrete models mainly
because of its simple formulation. This approach is adopted in the current work.
The plastic strain increment can be expressed as[15][50]:
f ( )
d p d …(4.19)
where, d 0 is a scalar hardening parameter which can vary throughout the
f ( )
straining process. The gradient of the yield loading surface , defines the
direction of the plastic-strain increment vector d p while the length is
f
x
C 2 ( C 2 ) x ( 2 C 2 ) ( y z ) / Q
f
y
C 2 ( C 2 ) y ( 2 C 2 ) ( x z ) / Q
f
z
C 2 ( C 2 ) z ( 2 C 2 ) ( x y ) / Q …(4.21)
f xy
6
xy Q
f yz
6
yz Q
f zx
6
zx Q
49
Chapter Four Modeling of Material Properties
where:
Q 2 [( C 2 )( x2 y2 z2 ) ( 2 C 2 )( x y y z z x )
3 ( xy
2
yz
2
zx
2
)] 0.5 …(4.22)
1 0 0 0
1 0 0 0
1 0 0 0
1 2
D E
0 0 0 0 0 …(4.26)
( 1 ) ( 1 2 ) 2
1 2
0 0 0 0 0
2
1 2
0 0 0 0 0
2
Substituting of Eq. (4.25) into Eq. (4.24) yields:
d D d d a
1
…(4.27)
50
Chapter Four Modeling of Material Properties
Pre-multiplying both sides of Eq. (4.27) by, a D , and eliminating, ad ,
T
By substituting of Eq. (4.28) into Eq. (4.24) and multiplying both sides by D ,
the complete elastic-plastic incremental stress-strain relation becomes:
d De p d …(4.29)
where:
De p
D aaT D
D …(4.30)
H a D a
T
and where the second term in the bracket represents the stiffness degradation
due to plastic deformation.
C . I 1 ( C . I 1 ) 2 3 . J 2 cu …(4.31)
where I 1 and J 2 are the strain invariants, cu is the ultimate crushing strain of
concrete, extrapolated from uniaxial test.
I 1 x y z …(4.32)
51
Chapter Four Modeling of Material Properties
1
J 2 x2 y2 z2 x y y z z x ( xy ) 2
3
( yz ) 2 ( zx ) 2 …(4.33)
52
Chapter Four Modeling of Material Properties
Y
Y'
X'
53
Chapter Four Modeling of Material Properties
54
Chapter Four Modeling of Material Properties
1 E1 0 0 0 0 0 1
0
2 E /( 1 2 ) E /( 1 2 ) 0 0 0 2
3 0 E /( 1 2 ) E /( 1 2 ) 0 0 0 3 …(4.37)
12 0 0 0 1 0 0 12
23 0 0 0 0 G 0 23
31 0 0 0 0 0 1G 31
z direction, respectively.
For the tension-tension-compression and the triaxial tension state of
stress, the cracking criterion may be validated by the major principal stress, 1 ,
and the second principal stress, 2 , simultaneously. Thus, two sets of
orthogonal failure planes develop. These planes are perpendicular to the
principal axes 1 and 2, respectively. In this case, Poisson’s ratio is set to zero in
all directions and the constitutive matrix in the local material axes become a
diagonal matrix:
55
Chapter Four Modeling of Material Properties
E1 0 0 0 0 0
0 E2 0 0 0 0
0 0 E 0 0 0
Dcr …(4.41)
0 0 0 1 G 0 0
0 0 0 0 2 G 0
0 0 0 0 0 G
In the fixed crack model, the rotation of the principal stresses after
formation of the first crack is ignored. However, if the second crack initiates at a
subsequent stage of loading, rotation of the principal stresses within the planes
perpendicular to the direction of 1 might be taken into account. Magnitudes
and direction of the new second and minor principal stresses, 2 and 3 , may
be calculated from the original principal stresses 2 and 3 , and the shear
stress 23 accumulated during loading stage beyond the formation of the first
crack. Thus by using Mohr circle[15]:
2 3 ( 2 3 )2
i 23
2
i=2,3 …(4.42)
2 2
For tension-tension-compression state, the onset of the second crack is
expected in terms of the new principal stresses as:
0.75 3
2 cr f t 1.0 …(4.43)
f c
56
Chapter Four Modeling of Material Properties
E2 0 0
Dcr 0 E 0
…(4.45)
0 0 2 G
This matrix may be transformed to the original directions (2,3) using the
in-plane coordinate transformation matrix. Therefore, the in-plane incremental
stress-strain relationship in the original principal directions can be written as:
2 E2 0 0 2
3 T p T 0
E
0 T p 3 …(4.46)
0 2 G
23 0 23
where, T p is the in-plane transformation matrix given by:
C2 S2 CS
T
p
S2 C2 CS
…(4.47)
2CS 2CS C 2 S 2
where C cos p , and S sin p .
1 ´
3 1
´ 1 1 ´
3
p
3
3 3
p p
3
3
3
3 3
3
3 c)
b) Doublecracks
b) Double cracks c) Triple cracks
Triple cracks
a)a)Single
Singlecracks
cracks
57
Chapter Four Modeling of Material Properties
cr
2 . cr ( 1 a cr )
a 2 . cr
( 1 1 )
E
1 ε
cr n 1 . cr
58
Chapter Four Modeling of Material Properties
I) For cr n 1 cr
n
( 1 )
cr
2 cr …(4.48)
1 1.0
II) For n 1 cr
n 0.0 …(4.49)
where n and n are the stress and strain normal to the crack plane.
cr is the cracking strain associated with the cracking stress cr .
59
Chapter Four Modeling of Material Properties
1.0 1 2
1 n 3
1 1.0 cr
2
3
ε
cr 1 cr
Fig. (4.6) Shear retention model for concrete[9].
y
Es Es s
su
Fig. (4.7) Stress-strain relationship of reinforcing and prestressing steel bars.
60
Chapter Five Formulation of Prestressing
Chapter Five
FORMULATION OF PRESTRESSING
5.1 Introduction
Prestressed concrete has become one of the major structural building materials
and it is widely used all over the world. In practice, concrete box-girder bridges
are most often prestressed and therefore the method of their analysis would be
incomplete without taking this important factor properly into account. Since this
study is devoted for the analysis of prestressed concrete structures, this chapter
considers in detail the derivation of an efficient prestress model capable of being
incorporated in a finite element solution[28].
There are different methods to take into analysis the prestressing tendon.
The simplest one is the method used by Lin and applied by Loo and Cusens[39].
It is also applied by Abdulla and Abdul-Razzak[1]. In there method the tendon is
assumed to be frictionless and acts in the neutral plane of the strip. The tendon
may be replaced by two types of in-plane forces: two end anchorage forces and a
uniform pressure along the span of the bridge structure.
In the second method used by Ghalib[28], the prestressing tendons are
considered as an integral part of the structure. A very sophisticated prestressing
tendon model, capable of modeling effectively the interaction between concrete
and the housed tendon throughout various prestressing processes is developed.
For bonded tendons, a generally curved tendon element embedded in a higher-
order isoparametric flat shell element is developed. For unbonded tendons, since
interaction between the prestressing tendon and the structure occurs mainly at
anchorages, the average extension method is considered very efficient in
modeling the stiffness contribution of such element.
61
Chapter Five Formulation of Prestressing
Another method was used by Kang et al[33], they analyzed the prestressing
tendon by using approximately a series of prestressing steel segments each of
which is straight, spans a frame element, and has a given profile, initial
tensioning force, and a constant cross-sectional area along its length. The stress
relaxation is computed utilizing the formula developed by Magura et al[42].
In the present study, the method used by Zyle and Scordelis[70] and
applied by Jirousek and Bouberguig[32] is considered. In this approach, the
equivalent load method is used to compute the force applied by the tendon upon
the structure where the prestressing tendon takes the form of a particular loading
case and as such it may be taken into consideration in the form of appropriate
local loads at the level of each element. Practically, this means that the effect of
prestressing manifests itself in the evaluation of the vectors of primary nodal
forces only[32]. The development of the necessary relations for practical
computation is briefly outlined hereafter.
Same approach was implemented by Roca and Mari[54][55] in 1993 with
including bonded and unbonded post-tensioned tendon with general formulation
for material and geometric nonlinearity.
62
Chapter Five Formulation of Prestressing
P Po . e ( x ) …(5.1)
where, p o : force in jack end (x=0).
p : force in tendon at distance x.
: curvature friction coefficient.
: angle change in prestressing tendon over distance x.
: wobble friction coefficient.
For parabolic profiles of constant curvature, Eq. (5.1) can be written as
follows:
p p o . e qx …(5.2)
where q a
q : constant profile curvature ( ax ).
The values of and are listed in Appendix A[49].
63
Chapter Five Formulation of Prestressing
p p o . e q ( 2 la x ) x la …(5.4)
p p o . e qx x la …(5.5)
Po
Po
P2
P1
La
Distance (x)
(b) Prestress force variation along the tendon both
before and after anchor release
Fig. (5.1) Prestress forces losses due to anchor slip[28].
64
Chapter Five Formulation of Prestressing
( 1 )...( i 1 )( i 1 )...( m )
Mi ( ) …(5.7)
( i 1 )...( i i 1 )( i i 1 )...( i m )
65
Chapter Five Formulation of Prestressing
where:
2 i1
1 1 , 2 1 , ... i 1 2 , ... m 1 …(5.8)
m1 m1
where Ti (i=1,2,…m) are given magnitudes of the tension at nodal points. They
will be specified on the basis of the customary prediction of loses of the
prestressing, such as those mentioned in section (5.2).
66
Chapter Five Formulation of Prestressing
acting on the element (Fig. (5.2)) in the direction of the main normal n of the
element axis (R is the radius of the principal curvature). Observing that the
global Cartesian components of the unit vectors t and n may be defined:
Vt dx
t with Vt …(5.12)
Vt d
and:
Vn 1 d 2 xc a dx c
n with V n ( ) …(5.13)
Vn Vt
2
d 2 Vt
2
d
d 2 x c dx c
where: a .
d 2 d
It is possible to combine the global components of the tangential and
normal loads into a unique global Cartesian local vector:
67
Chapter Five Formulation of Prestressing
Px
p Py Pt t Pn n …(5.14)
P
z
where:
dT 1 dT T
Pt and Pn V n T …(5.15)
ds V t d R
68
Chapter Five Formulation of Prestressing
Fs N T ( c , c , c ) P ( ) ds ( N T ( c 1 , c 1 , c 1 ) P1 or
l
N ( cm , cm , cm ) Pm )
T
Fs N T ( c , c , c ) P ( ) Vt ( ) d
l …(5.19)
( N T ( c 1 , c 1 , c 1 ) P1 or N T ( cm , cm , cm ) Pm )
The large brackets, containing the end anchoring forces, have been used
to indicate that the use of these terms is limited to cases where the point C 1 or
C m coincides with the end of the tendon. If more than one tendon transverse the
element considered, the effect of each one is computed separately, using Eq.
(5.19), the results being summed.
The integration in Eq. (5.19) is carried out numerically by using the
Gaussian quadrature rule. A three-point Gaussian rule is adopted in this study.
Since Gaussian points have to be distributed along the curved tendon
length, one is confronted with the problem of determining the particular values
c , c and c of the curvilinear coordinate , and at these points which have
given values of the parameter . Thus an inverse mapping procedure from the
global coordinates system to natural coordinates system is required in order to
locate the Gaussian points in the natural coordinate system. The problem is non-
linear and leads to the following system of three non-linear equations[32]:
x( c , c , c ) x c ( ) …(5.20)
where the left-hand term is represented by the coordinate definition Eq. (3.18)
for the brick element, whilst the second term is computed from the tendon
definition Eq. (5.6). Since the relationship (3.18) in Chapter three between the
Cartesian and curvilinear coordinates is one-to-one, the system (5.20) may easily
be numerically solved for any given value of the parameter by the use of the
Newton-Raphson method.
69
Chapter Five Formulation of Prestressing
(y3,z3)
-P (y2,z2)
(y1,z1)
θ Y
y
Fig. (5.3) Analysis of curved cable[1].
This treatment of curved tendon follows the procedure used by Lin and
applied by Loo and Cusens[39][40].
The span of the cable is assumed to be parabolic and the total change of
the slope is calculated as:
tan 1 ( B 2C z 3 ) tan 1 ( B 2C z 1 ) …(5.21)
where:
y1 y 2 ) y y3
( )( 2 )
z1 z2 z2 z3
C …(5.22)
( z1 z3 )
70
Chapter Five Formulation of Prestressing
y2 y3
B C ( z2 z3 ) …(5.23)
z2 z3
in which (y1,z1), (y2,z2) and (y3,z3) are the coordinates for any three points in the
parabolic curve spanning a brick element. The uniform pressure along the
tendon duct may be replaced by a uniformly distributed in-plane load P along a
line parallel to the z-axis, as shown in Fig. (5.3), and:
T
P …(5.24)
( z3 z1 )
where T is the prestressing force in the tendon. The load is assumed to act as
line y , where:
y ( y1 y 2 y 3 ) / 3 …(5.25)
In the present study, the uniform pressure is distributed equally upon the
nodes for the elements spanning with parabolic tendon.
71
Chapter Six Applications and Results
Chapter Six
APPLICATIONS AND RESULTS
6.1 Introduction
The aim of this chapter is to verify the present nonlinear finite element model
developed to investigate the behavior and ultimate load capacity of prestressed
concrete box girders. A computer program has been developed to carry out the
analysis. It is also intended to check the accuracy and validity of the computer
program which deals with box-girder bridges subjected to nonproportional loads
and initial prestressing forces. Several examples are considered in this chapter.
The results of the analysis of these examples are discussed and the load-
deflection behaviour obtained by the finite element analysis is compared with
those obtained from the available experimental investigation.
72
Chapter Six Applications and Results
Since the one-cell box-girder was symmetrically loaded with respect to its
longitudinal axis, only one-half of the box-girder is modeled. The one-half
structure has been modeled by 252 brick elements with a total number of 1815
nodal points, as shown in Fig. (6.2 a).
Fig. (6.4) represents the load-midspan deflection curve of the prestressed
box-girder bridge. The obtained results show close agreement in comparison
with the experimental results. The box-girder bridge failed due to crushing in
concrete at failure load level of 56.425 kN. Comparison between the present
study 56.425 kN (4.41 trucks) and experimental value 55.0 kN (4.33 trucks)
shows good agreement with an error just exceeding 2.53%. This is in the same
comparison with the NONLACS study.
Fig. (6.5) shows the deflected shape of the bridge at various load levels.
The level of the load is indicated by the ratio P/Pu, where P is the load at which
the deflection is evaluated and Pu is the ultimate load for the bridge. These
deflections were measured along the longitudinal centerline of the bridge. The
obtained results are in good agreement with the experimental and NONLACS
studies.
The concrete longitudinal normal stresses at various locations in the
bridge are illustrated in Figs. (6.6), (6.7) and (6.8). Fig. (6.6) shows the
longitudinal normal stress in the centerline of the top slab at midspan versus the
level of load. In this figure, the initial longitudinal stress at zero loads in the
present study is less than the initial longitudinal stress in the experimental and
NONLACS results. But after the loading reaching to 40 kN, the longitudinal
stress were near to the experimental results. Generally, the obtained results were
nearly close to the experimental results. While, Fig. (6.7) shows the longitudinal
normal stress on the centerline of the top slab at quarterspan versus the level
load. The initial longitudinal stress at zero loads also gives a value less than the
initial stress in the experimental and NONLACS studies.
73
Chapter Six Applications and Results
25
35
a- cross section 56
12 56
12
17 17
C
L
Midspan Section End Section
225
129
73
41
29
17
17
70 3 X 130 8 X 145 70
1690
All dimensions in mm
74
Chapter Six Applications and Results
a- Finite element idealization with 252 brick elements b- Deflected shape due to prestressing and
for one-half for one-cell box-girder external loading
Fig. (6.2) Finite element idealization for half bridge model of one-cell box-
girder before and after loading
75
Chapter Six Applications and Results
Concrete
76
Chapter Six Applications and Results
1 2 3 4 No. of axles
85
85
TRAVEL
424
85
85
85
All dimension in mm
35
Fig. (6.3) Simulated Ontario Highway Bridge Design trucks (OHBDC)
for one-cell box-girder bridge
60.00
50.00
without losses
30.00
20.00
Experimental Ref.(53)
NONLACS Ref.(53)
Present study
10.00
0.00
77
Chapter Six Applications and Results
0.00
P/Pu=0.250
2.00
P/Pu=0.500
4.00
deflection (mm)
6.00 P/Pu=0.750
8.00
10.00
Experimental Ref.(53)
12.00
P/Pu=0.875
NONLACS Ref.(53)
14.00
Present study with losses
60.00
50.00
without losses
40.00
with losses
load (kN)
30.00
0.00
78
Chapter Six Applications and Results
60.00
50.00
without losses
with losses
40.00
load (kN)
30.00
20.00
Experimental Ref.(53)
NONLACS Ref.(53)
0.00
0.00
2.00
4.00 P/Pu=0.250
6.00
P/Pu=0.500
Stress (MPa)
8.00
10.00
P/Pu=0.750
12.00
Experimental Ref.(53)
14.00
16.00
NONLACS Ref.(53)
79
Chapter Six Applications and Results
Experimental
NONLACS
Present study
P/Pu=0.250 P/Pu=0.875
Experimental
NONLACS
Present study
P/Pu=0.250 P/Pu=0.875
Fig. (6.10) Longitudinal normal stress variation across the section of slab
at quarter span for one-cell box-girder bridge
80
Chapter Six Applications and Results
Table (6.2) Variation of longitudinal stresses at midspan for Pu/P=0.250 and 0.875.
P/Pu=0.250 P/Pu=0.875
Points Present Present
Exp. NONLACS Exp. NONLACS
study study
A B C D E
’
C
F
I J K
81
Chapter Six Applications and Results
Table (6.3) Variation of longitudinal stresses at quarter span for Pu/P=0.250 and
P/Pu=0.250 P/Pu=0.875
Points Present Present
Exp. NONLACS Exp. NONLACS
study study
0.875.
A B C D E
’
C
F
I J K
82
Chapter Six Applications and Results
2
0
-2
-4
-6
-8
-10
-12
-14
-16
-18
-20
-22
-24
-26
Fig. (6.11) Variation of longitudinal stresses in the top surface of the top flange
83
Chapter Six Applications and Results
1
0.5
0
-0.5
-1
-1.5
-2
-2.5
-3
-3.5
-4
Fig. (6.12) Variation of transverse stresses in the top surface of the top flange
84
Chapter Six Applications and Results
6.5
6
5.5
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
-0.5
Fig. (6.13) Variation of longitudinal stresses in the bottom surface of the bottom
flange
85
Chapter Six Applications and Results
1.5
1
0.5
0
-0.5
-1
-1.5
-2
-2.5
-3
-3.5
-4
Fig. (6.14) Variation of transverse stresses in the bottom surface of the bottom
flange
86
Chapter Six Applications and Results
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
-1
-2
Fig. (6.15) Variation of longitudinal stresses in the outer surface of the web
87
Chapter Six Applications and Results
2.2
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.4
Fig. (6.16) Variation of transverse stresses in the outer surface of the web
88
Chapter Six Applications and Results
12
11
10
9
8
7
6
5
4
3
2
1
0
-1
-2
Fig. (6.17) Variation of longitudinal stresses in the inner surface of the web
89
Chapter Six Applications and Results
1.20
1.00
0.80
0.60
0.40
0.20
0.00
-0.20
-0.40
-0.60
-0.80
-1.00
-1.20
Fig. (6.18) Variation of transverse stresses in the inner surface of the web
90
Chapter Six Applications and Results
91
Chapter Six Applications and Results
obtained results are in good agreement with respect to the experimental and
NONLACS results until the load is nearly equal to 135.0 kN. A good agreement
with respect to the experimental results at the top slab at quarter span of the
bridge is also shown in Fig. (6.25). It can be noted that, the obtained results in
the present study are more close to the experimental results than the NONLACS
results. The variation of concrete longitudinal normal stresses along the
centerline of the top slab is shown in Fig. (6.26). The comparison is fairly close
with respect to the experimental results. Figs. (6.27) and (6.28) show the
variation of longitudinal normal stress at the cross-section at midspan and
quarter span for the bridge at a ratio P/Pu equal to 0.182 and 0.727. These
results are listed in tables (6.5) and (6.6).
The deflected shape due to the loading is shown in Fig. (6.20 b). All
displacements in the structure are multiplied by 10 for clearness. The
distribution of the longitudinal and transverse normal stresses for the bridge is
shown in Figs. (6.29) to (6.40).
L
C 560
30
a- cross section 50
135
35 19 30
19
35
19
19
1766
All dimensions in mm
a- Finite element mesh for the structure b- Deflected shape for the two-cell box-girder
with 308 elements due to prestressed and external loads
Concrete
94
Chapter Six Applications and Results
1 2 3 4 No. of axles
86
86
TRAVEL
257
429
86
86
86
36
All dimension in mm
Fig. (6.21) Simulated Ontario Highway Bridge Design trucks (OHBDC)
for two-cell box-girder bridge
180.00
without losses
160.00
140.00
120.00
with losses
load (kN)
100.00
80.00
20.00
0.00
0.00
P/Pu=0.182
2.00 P/Pu=0.364
4.00 P/Pu=0.546
6.00
P/Pu=0.727
8.00
deflection (mm)
10.00
12.00 P/Pu=0.909
Experimental Ref.(53)
14.00
16.00
180.00
with losses
160.00
without losses
140.00
120.00
load (kN)
100.00
80.00
Experimental Ref.(53)
60.00
NONLACS Ref.(53)
20.00
0.00
-5.00 0.00 5.00 10.00 15.00 20.00 25.00 30.00 35.00 40.00
stress (MPa)
96
Chapter Six Applications and Results
180.00
with losses
160.00
without losses
140.00
120.00
load (kN)
100.00
80.00
20.00
0.00
0.00
2.00
P/Pu=0.182
4.00
stress (MPa)
P/Pu=0.364
6.00
P/Pu=0.546
8.00
Experimental Ref.(53)
10.00
NONLACS Ref.(53)
Present study with losses
P/Pu=0.727 Present study without losses
12.00
Experimental
NONLACS
Present study
P/Pu=0.727 P/Pu=0.182
Experimental
NONLACS
Present study
P/Pu=0.727 P/Pu=0.182
98
Chapter Six Applications and Results
Table (6.5) Variation of longitudinal stresses at midspan for P/Pu=0.182 and 0.727.
P/Pu=0.182 P/Pu=0.7275
Points Present Present
Exp. NONLACS Exp. NONLACS
study study
A B C D E F G H I
C' G'
J K
L M
N R'
N' O P Q R
99
Chapter Six Applications and Results
Table (6.6) Variation of longitudinal stresses at quarter span for P/Pu=0.182 and 0.727.
P/Pu=0.182 P/Pu=0.7275
Points Present Present
Exp. NONLACS Exp. NONLACS
study study
A B C D E F G H I
C' G'
J Q
K R
L P'
L' M N O P
100
Chapter Six Applications and Results
0
-2
-4
-6
-8
-10
-12
-14
-16
-18
-20
-22
-24
Fig. (6.29) Variation of longitudinal stresses in the top surface of the top flange
101
Chapter Six Applications and Results
1.5
0.5
-0.5
-1
-1.5
-2
-2.5
-3
Fig. (6.30) Variation of transverse stresses in the top surface of the top flange
102
Chapter Six Applications and Results
1.5
1
0.5
0
-0.5
-1
-1.5
-2
-2.5
-3
-3.5
-4
Fig. (6.31) Variation of longitudinal stresses in the bottom surface of the bottom
flange
103
Chapter Six Applications and Results
0.5
0
-0.5
-1
-1.5
-2
-2.5
-3
-3.5
-4
-4.5
-5
-5.5
Fig. (6.32) Variation of transverse stresses in the bottom surface of the bottom
flange
104
Chapter Six Applications and Results
1
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
Fig. (6.33) Variation of longitudinal stresses in the outer surface of the right web
105
Chapter Six Applications and Results
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-0.7
-0.8
-0.9
-1
-1.1
Fig. (6.34) Variation of transverse stresses in the outer surface of the right web
106
Chapter Six Applications and Results
0.5
-0.5
-1.5
-2.5
-3.5
-4.5
-5.5
-6.5
-7.5
-8.5
-9.5
Fig. (6.35) Variation of longitudinal stresses in the inner surface of the right web
107
Chapter Six Applications and Results
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
Fig. (6.36) Variation of transverse stresses in the inner surface of the right web
108
Chapter Six Applications and Results
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
-11
-12
Fig. (6.37) Variation of longitudinal stresses in the outer surface of the left web
109
Chapter Six Applications and Results
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.4
-1.6
Fig. (6.38) Variation of transverse stresses in the outer surface of the left web
110
Chapter Six Applications and Results
0
-1
-2
-3
-4
-5
-6
-7
-8
-9
-10
-11
-12
Fig. (6.39) Variation of longitudinal stresses in the inner surface of the left web
111
Chapter Six Applications and Results
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
-1
-1.2
-1.4
-1.6
Fig. (6.40) Variation of transverse stresses in the inner surface of the left web
112
Chapter Six Applications and Results
113
Chapter Six Applications and Results
25.00 cm
56.00 cm
28500 kN 150.00 cm
25.00 cm
10.00 cm
625.00 cm 625.00 cm 625.00 cm
a- profile
190.00 cm
35.00 cm 375.00 cm
190.00 cm
b- plan
190.00 cm 185.00 cm
35.00 cm
22.50 cm
150.00 cm
32.00 cm
c- cross-section 25.00 cm
200.00 cm
Fig. (6.41) Structural details and cable profile of prestressing for box-girder
bridge with inclined webs
114
Chapter Six Applications and Results
Fig. (6.42) Finite element mesh for the bridge model before and after loading
115
Chapter Six Applications and Results
Table (6.7) Material properties for the prestressed box-girder bridge with inclined
webs.
Concrete
*assumed
*E c 5000 f c
*f t 0 .54 f c
116
Chapter Six Applications and Results
Experimental
NONLACS
Present study
117
Chapter Six Applications and Results
118
Chapter Six Applications and Results
A 0.283 0.3049
B .126 0.3203
C .078 0.2998
D 0.138 0.3060
E 0.189 0.3497
F 0.138 0.3519
G 0.078 0.303
H 0.126 0.3213
I 0.283 0.3059
J -0.752 -1.008
K -0.752 -1.015
L -2.851 -2.836
M -2.851 -2.841
N -3.601 -5.242
O -3.601 -5.238
P -4.171 -5.242
Q -4.314 -5.238
R -4.2 -5.38
S -4.314 -5.251
T -4.171 -5.242
A B C D E F G H I
J K
L M
N L
P Q R S T
119
Chapter Six Applications and Results
-0.35
-0.45
-0.55
-0.65
-0.75
-0.85
-0.95
-1.05
-1.15
-1.25
-1.35
Fig. (6.45) Variation of longitudinal stresses in the top surface of the top flange
120
Chapter Six Applications and Results
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-0.7
Fig. (6.46) Variation of transverse stresses in the top surface of the top flange
121
Chapter Six Applications and Results
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
-1
Fig. (6.47) Variation of longitudinal stresses in the bottom surface of the bottom
flange
122
Chapter Six Applications and Results
3.5
2.5
1.5
0.5
-0.5
-1
Fig. (6.48) Variation of transverse stresses in the bottom surface of the bottom
flange
123
Chapter Six Applications and Results
19
17
15
13
11
9
7
5
3
1
-1
Fig. (6.49) Variation of longitudinal stresses in the outer surface of the left web
124
Chapter Six Applications and Results
2.4
2
1.6
1.2
0.8
0.4
0
-0.4
-0.8
-1.2
-1.6
Fig. (6.50) Variation of transverse stresses in the outer surface of the left web
125
Chapter Six Applications and Results
19
17
15
13
11
9
7
5
3
1
-1
Fig. (6.51) Variation of longitudinal stresses in the inner surface of the left web
126
Chapter Six Applications and Results
2.2
2
1.8
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
-0.2
-0.4
-0.6
-0.8
Fig. (6.52) Variation of transverse stresses in the inner surface of the left web
127
Chapter Six Applications and Results
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
Fig. (6.53) Variation of longitudinal stresses in the outer surface of the right web
128
Chapter Six Applications and Results
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
-0.1
-0.2
-0.3
-0.4
-0.5
-0.6
-0.7
-0.8
Fig. (6.54) Variation of transverse stresses in the outer surface of the right web
129
Chapter Six Applications and Results
17
16
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
-1
-2
Fig. (6.55) Variation of longitudinal stresses in the inner surface of the right web
130
Chapter Six Applications and Results
1.1
0.9
0.7
0.5
0.3
0.1
-0.1
-0.3
-0.5
-0.7
-0.9
Fig. (6.56) Variation of transverse stresses in the inner surface of the right web
131
Chapter Seven Conclusions and Recommendations
Chapter Seven
CONCLUSIONS AND RECOMMENDATIONS
7.1 Conclusions
The present study has been used concrete brick element with embedded steel bar
element for the analysis of prestressed concrete box-girder bridges. The
prestressing forces are dealt by using the equivalent nodal force method and by
Lin's method. The losses in prestressing forces had been considered in the
analysis. Several examples are analyzed by the present method and the results
are compared with the available experimental and other analytical results. From
these results, the following conclusions are drawn.
1. The nonlinear finite element method presented in this study has shown to
be capable of predicting the structural response of prestressed concrete
box-girder bridges. The isoparametric brick element proved to be suitable
for predicting the state of ultimate load, deflection and stress with good
accuracy. The difference for the single-cell box-girder with respect to
experimental results of Ref.(53), in ultimate load is 2.59%, and in stress at
quarterspan is 2.27%, and it is equal to 12.69% for the stress at midspan.
In the second example, the two-cell box-girder bridge, the difference is to
be less than 9.15% comparing with experimental results in Ref.(53) for
ultimate load, but it is less than 5.23% comparing with theoretical results
for the same authors. The tolerance is equal to 2.06% for the deflection at
midspan, and it is equal to 8.57% for the stress at quarterspan, and it is
less than 9.97% for the stress at midspan. For the prestressed concrete
box-girder with inclined webs, the third example, the maximum
difference with respect to the theoretical results in Ref.(32), in deflection
132
Chapter Seven Conclusions and Recommendations
is equal to 8.0% at top flange, and in stress is to be less than 14.0% at the
bottom flange.
2. The losses in prestressing forces used in the present study improved the
obtained results. For the first example, it was found that the accuracy
increased with 2.83% for the ultimate load, and it is increased with 4.67%
for the deflection at midspan. Besides, the accuracy is increased with
0.75% for the stress at quarterspan, and it is increased with 0.26% for the
stress at midspan. In the second example, the tolerance increased with
4.51% for the ultimate load, and it is increased with 2.6% for the
deflection at midspan. While, the accuracy is increased with 11.8% for the
stress at quarterspan, and it is increased with 15.0% for the stress at
midspan. For the third example, the tolerance increased with 12.0% for
the deflection at the top flange.
3. The concept of equivalent nodal forces used in the present study is
capable to simulate the loads exerted by the prestressing tendon upon the
structure with fair accuracy. Also, the Lin's method is proved to be
suitable to simulate the forces due to the parabolic tendon upon the
structure.
4. The contribution of the prestressing tendon stiffness to the element
stiffness is contributed, and it is affected on the analysis.
5. Boundary conditions caused most of the differences in the results. The
boundary conditions should be taken as exactly as possible it is in the
experimental work.
133
Chapter Seven Conclusions and Recommendations
7.2 Recommendations
The following suggestion for the possible extension and improvement of the
present study can be made:
1. The dependent effects are needed to be incorporated in this study. These
effects include concrete shrinkage, creep and aging of concrete as well as
relaxation in prestressing steel.
2. Dynamic response of prestressed concrete box-girder bridges due to blast
load and earthquakes may be studies.
3. Modified the available program to take thermal loading into account.
4. Modified the available program by treating the tendon as an integral part
of the structure.
134
References
REFERENCES
135
References
136
References
137
References
138
References
139
References
140
References
141
Appendix A
Appendix A
Unbonded Tendons
A
Appendix B
Appendix B
ـ ﻋﺔ ﻣـ ﺧرﺳـﺎﻧﺔ ﻣ ـ ﻘﺔ ـ ر ذات اﻟ ﻘـﺎ ﻊ اﻟ ـ دوﻗ ﺔ واﻟ ﺗ ﻓـﻲ ﻫـذﻩ اﻟدراﺳـﺔ ﺗ ﻠ ـﻞ اﻟ
ــر اﻟ ـ از اﻟ ﻬــد ﺗ ــت ﺗــﺄﺛ ر اﻷﺣ ــﺎل ﻗ ـ رة اﻷﻣــد .ﺗ ـ ﺗ ــﻞ اﻟ ﻘــﺎ ﻊ اﻟ ـ دوﻗ ﺔ ﺎﺳ ـ دام اﻟﻌ
ﻋﻘدة ذات درﺟﺎت اﻟ ﻼﻗﺔ اﻟ ﻼﺛﺔ )ﺛـﻼث ازاﺣـﺎت( .أﻣـﺎ دراﺳـﺔ ﺣدﯾـد اﻟ ـﻠ ﺢ ﻓﻘـد اﻟ ﺎﺑ ﻗﻲ ذو اﻟﻌ ر
ذج ﻣرن-ﻟدن ﯾ ﻌﻪ ﺟزء ﻟدن ﺗﺎم اﻟذ ﯾ ﻬﻲ ﻋ د ﺑدا ﺔ ﺗﻬ ـ اﻟ رﺳـﺎﻧﺔ .أﻣـﺎ ﺗ ـت اﻻﻧ ﻐﺎط ﻣ ﻞ ﺑ
اﻟ ــد ﻓﻘــد اﺳ ـ دم ﻧ ـ ذج ﺗ ز ــﻊ اﻟ ــﻘ ق اﻟﻌ د ــﺔ ﻋﻠــﻰ ﻣ ــﺎور اﻹﺟﻬــﺎد اﻟرﺋ ـ ﺔ ﻣــﻊ اﻷﺧــذ ﺑ ــر
ــﻞ اﻟ ﻬــد ــر .ﺗـ اﺳـ دام ر ﻘــﺔ ﺛﺎﻧ ــﺔ ﻟ ﻣ ــﻠ ﺔ ﻓــﻲ اﻟﻌﻘــد ﻟ ــﻞ ﻋ ﻣـ ﻗ ــﻞ ﺣ ــﻞ اﻟ ــد إﻟــﻰ ﻗـ
(Global ــﻔ ﻓﺔ اﻟ ﻠ ــﺔ ﻫــﻲ ر ﻘــﺔ ﻟـ ) . (Linﺗـ إدﺧــﺎل ﻣ ــﻔ ﻓﺔ ﺣ ــﻞ اﻟ ــد إﻟــﻰ اﻟ ـ اﻟ
ــﺎ درﺳــت ﺣﺎﻟــﺔ ــر اﻟ رﺳــﺎﻧﺔ. ــر أﺣــﺎد اﻟ ﻌــد )ﻣ ـ ر ( ﻣ ـ ر ﻓــﻲ ﻋ ) stiffnessﺎﻋ ــﺎرﻩ ﻋ
ـ .وﻗـد اﺧـذ ﻓـﻲ اﻟ ﻠ ـﻞ ـﺎﻫرة اﻧـزﻻق اﺳ ﺎق اﻟ ﻬد ﻋ د وﺟ د ﺧ ﺎﺋر ﻗ ـ رة اﻟ ـد ﻓـﻲ اﻟ ﻬـد اﻟ
ﺗ ﺣﻞ ﻣﻌﺎدﻻت اﻟ ازن اﻟﻼﺧ ﺔ ﺎﺳـ دام ر ﻘـﺔ ﺗزاﯾد ـﺔ-ﺗﻌدد ـﺔ ﺗﻌ ـﻞ ﺗ ـت أﺣ ـﺎل ﻣ ـ رة
ــﻬﺎ أو ﺗ ﻠ ﻠﻬــﺎ ﻣ ـ ﻘﺎً ﺎﺳـ دام ــﺎرة واﻟ ــﻲ ﺗـ ﻓ ﺗ ــت دراﺳــﺔ وﺗ ﻠ ــﻞ ﻋــدد ﻣـ اﻷﻣ ﻠــﺔ اﻟ
رق ﻋ ﻠ ﺔ وﻧ ر ﺔ ﻣﻌ دة أﺧر .أ ﻬـرت ﻧ ـﺎﺋﺞ اﻟ ﻘﺎرﻧـﺔ ـﺎن ر ﻘـﺔ اﻟﻌ ﺎﺻـر اﻟ ـددة أﻋ ـت ﻧ ـﺎﺋﺞ
دوﻗ ﺔ ﻘﺔ اﻹﺟﻬﺎد ذات اﻟ ﻘﺎ ﻊ اﻟ اﻟ ﻠ ﻞ اﻹﻧ ﺎﺋﻲ ﻟﻠﻌ ارض اﻟ رﺳﺎﻧ ﺔ اﻟ
ﺎﺳ دام ر ﻘﺔ اﻟﻌ ﺎﺻر اﻟ ددة
إ روﺣﺔ ﻣﻘدﻣﺔ
إﻟﻰ ﻠ ﺔ اﻟﻬ دﺳﺔ ﻓﻲ ﺟﺎﻣﻌﺔ ﻐداد
ر ﻋﻠ م ﻓﻲ وﻫﻲ ﺟزء ﻣ ﻣ ﻠ ﺎت ﻧ ﻞ درﺟﺔ اﻟ ﺎﺟ
اﻟﻬ دﺳﺔ اﻟ دﻧ ﺔ )إﻧ ﺎءات(
ﻣ ﻗﻞ
اﻟﻐراو ﻣﻬ د إﺑراﻫ ﻣ د ﺣ
) ﺎﻟ ر س (2000