Mekanik KTH

MSc Thesis
2003:08

Finite Element Modelling

of Composite Bridge Stability

Martin Ålenius

MSc Thesis
Stockholm, 2003

Royal Institute of Technology

Department of Mechanics
 Finite Element Modelling
of Composite Bridge Stability

by

Martin Ålenius

August 2003
Technical Reports from
Royal Institute of Technology
Department of Mechanics
SE-100 44 Stockholm, Sweden
Kungliga Tekniska Högskolan i Stockholm, Valhallavägen 79, Stockholm.


c Ålenius 2003
Preface

The work presented in this thesis was initiated by the structural engineering com-
pany Tyréns AB and the Department of Mechanics at the Royal Institute of Tech-
nology in Stockholm, where it was carried out between January 2003 and July 2003.
I would like to thank Adjunct Prof. Dr. Per-Olof Thomasson for introducing and
letting me take part in this field of research, and also for his valuable advice and
guidance throughout the course of this thesis.
I also express my deepest gratitude and admiration to Prof. Anders Eriksson for his
encouragement during his courses at KTH and particularly the courses in the finite
element methods that paved the way for this thesis. I also would like to thank him
for his shrewd scientific guidance throughout the course of this thesis.
Furthermore, I thank Johan Kölfors at Scanscot Technology for helping me solve
modelling problems in BRIGADE/Plus.
A special thank goes to the staff at the Department of Mechanics, particularly to
Dr. Gunnar Tibert for his helpfulness and interesting discussions on finite element
analysis.
Finally, I would like to thank my family and friends for their love, patience and
support in accomplishment of this thesis.

Stockholm, August 2003

Martin Ålenius

iii
Abstract

This thesis deals with stability problems for a thin-walled box girder steel-concrete
composite bridge. A case study of bridge Y288 over river Ljungan was performed.
The study was based on finite element modelling made in BRIGADE/Plus.
Initially, the background to this thesis is presented together with a description of
the studied bridge. An introduction to the used theories, such as buckling of thin
plates and torsion of thin-walled open sections is given. An extensive description of
the modelling procedures in BRIGADE/Plus is presented alongside a comparison
with ABAQUS.
Three different finite element models were analysed: (i) a simply supported rectan-
gular plate uniformly compressed in one direction, (ii) profiled sheeting subjected
to shear forces and (iii) the lateral torsional stability of the bridge.
The first analysis aimed to use a simple model for which analytical results were
available and compare them with finite element modelling results. The finite element
analysis provided results well in accordance with the analytical results for the critical
buckling load by Timoshenko.
The analysis of the profiled sheeting in shear aimed to study the importance of the
attachment techniques of the profiled sheeting. This analysis clearly illustrated that
a substantial reduction of the stresses in the profiled sheeting is obtained with an
all around attachment between the profiled sheeting and box girder, compared to a
two-sided attachment. Furthermore, it distinctly demonstrated that the large axial
forces that arose at the free edge of the profiled sheeting, when it is attached along
two sides, were considerably diminished when an all around attachment is used.
The main aim of the lateral torsional stability analysis was to study the overall
behaviour of the bridge and how it is influenced by initial imperfections and different
attachment of the profiled sheeting. The profiled sheeting creates a closed cross-
section, providing a structure that is more rigid than an open cross-section. In
order to obtain a closed cross-section, the profiled sheeting must be given an all
around attachment and also be strong enough to withstand arisen stresses, mainly
at the connections between the profiled sheeting and the box girder. If the profiled
sheeting fails to do this, the cross-section of the bridge behaves like an open, thin-
walled cross-section.
Keywords: buckling mode, connection, finite element analysis, initial imperfec-
tions, profiled sheeting, thin-walled structure, lateral torsional stability.

v
Contents

Preface iii

Abstract v

1 Introduction 1
1.1 Aims of the Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Structure of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Composite Bridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.4 Finite Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Studied Bridge and Used Computational Tools 9

2.1 Bridge Y288 over Ljungan . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Computational Tools . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.1 BRIGADE/Plus . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2.2 Comparison of BRIGADE/Plus and ABAQUS . . . . . . . . . 11

3 Loads and Engineering Analysis 13

3.1 Loads Acting on the Bridge . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.1 Dead Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.1.2 Live Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3.2 Engineer Calculations of Stability . . . . . . . . . . . . . . . . . . . . 16

4 Thin-Walled Structures 19
4.1 Brief Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.2 Bending of Thin Plates . . . . . . . . . . . . . . . . . . . . . . . . . . 21

vii
 4.3 Buckling of Thin Plates . . . . . . . . . . . . . . . . . . . . . . . . . 22
4.4 Plate and Box Girder Analysis . . . . . . . . . . . . . . . . . . . . . . 24
4.5 Torsion of Thin-Walled Open Sections . . . . . . . . . . . . . . . . . 24
4.5.1 Uniform Torsion of Thin-Walled Open Sections . . . . . . . . 25
4.5.2 Non-Uniform Torsion of Thin-Walled, Open Sections . . . . . 26
4.6 Buckling of Simply Supported Rectangular Plate . . . . . . . . . . . 28
4.6.1 Comparison of Theory and BRIGADE/Plus . . . . . . . . . . 31

5 Creating Models in BRIGADE/Plus 35

5.1 Modelling Procedures in BRIGADE/Plus . . . . . . . . . . . . . . . . 35
5.2 Buckling in General . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
5.3 General Static Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 38
5.4 Linear Eigenvalue Buckling Prediction . . . . . . . . . . . . . . . . . 38
5.4.1 The BUCKLE Command . . . . . . . . . . . . . . . . . . . . 39
5.4.2 The IMPERFECTION Command . . . . . . . . . . . . . . . . 39
5.5 Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.6 Improving Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.7 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41

6 Profiled Sheeting in Shear 43

6.1 Profiled Sheet Attachment . . . . . . . . . . . . . . . . . . . . . . . . 43
6.2 Profiled Sheet Thickness . . . . . . . . . . . . . . . . . . . . . . . . . 45
6.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

7 Lateral Torsional Stability 51

7.1 Implementation of Geometric Imperfections . . . . . . . . . . . . . . 53
7.1.1 Linear Eigenvalue Prediction . . . . . . . . . . . . . . . . . . . 53
7.2 Postbuckling Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . 55
7.3 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
7.3.1 Box Girder with Open Cross-Section . . . . . . . . . . . . . . 56
7.3.2 Box Girder with Closed Cross-Section . . . . . . . . . . . . . . 70

viii
 7.3.3 Box Girder with Partly Closed Cross-Section . . . . . . . . . . 75
7.3.4 Comparison of Results . . . . . . . . . . . . . . . . . . . . . . 80

8 Conclusions 83
8.1 Finite Element Calculations . . . . . . . . . . . . . . . . . . . . . . . 83
8.2 Suggested Direction for Further Research . . . . . . . . . . . . . . . . 85

Bibliography 87

A ABAQUS Template 89
A.1 Template . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89

B ABAQUS Input Files 95

B.1 Profiled Sheets in Shear . . . . . . . . . . . . . . . . . . . . . . . . . 95
B.1.1 Long side Attachment Input File . . . . . . . . . . . . . . . . 95
B.1.2 All Around Attachment Input File . . . . . . . . . . . . . . . 97
B.2 Torsional Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
B.2.1 Box Girder with Open Cross-Section . . . . . . . . . . . . . . 99
B.2.2 Box Girder with Closed Cross-Section . . . . . . . . . . . . . . 103
B.2.3 Box Girder with Partly Closed Cross-Section . . . . . . . . . . 106
B.2.4 Example of Assembly . . . . . . . . . . . . . . . . . . . . . . . 109

ix
Chapter 1

Introduction

In June 2002 bridge Y15041 over Gide älv collapsed during construction. The result
of this collapse is shown in Figure 1.1. The bridge was a simply supported girder
bridge with a span of 65 m and a width of 7 m. The bridge was a conventional
composite construction containing a flanged box whose lower part was made of a
steel box on which a concrete carriageway is casted. The construction demands that
the load-bearing capacity of the steel box is sufficient for carrying the self-weight
during launching and for carrying the loads from casting. Other loads, such as
traffic load, are jointly carried by the steel and the concrete carriageway when the
construction is complete. The part of the carriageway that was situated between the
flanges was intended to be cast on so called lost formwork made of profiled sheeting.
These profiled sheets were at the same time designated to act as stabilising elements
for the box girder during construction. The casting of the carriageway was divided
into two steps where in the first step a volume of 66 m3 concrete was to be casted,
covering 33 m of the central part of the bridge.

Figure 1.1: Bridge Y1504 over Gide älv after collapse (Courtesy of NCC AB).
1
working name

1
CHAPTER 1. INTRODUCTION

The casting work was initiated at the centre of the bridge and was thereafter intended
to be carried on towards the abutments. When about 17 m3 concrete had been cast
the bridge twisted and the midsection turned almost 90◦ instantaneously.
At the moment, a similar bridge, bridge Y288 over the river Ljungan, is under
construction. This bridge will be closely studied in this thesis. The fundamental
differences between bridge Y1504 and Y288 are that the latter has a span of 57 m
and that the profiled sheeting has not been used as stabilising element in the design
calculations. The casting procedures have also been altered. Instead of starting
at the centre of the bridge, casting began by casting about 2 m of the ends of the
carriageway, thus creating an increased torsional warping restraint for the structure.
In this chapter the aims of the thesis and an outline of its structure are presented.
Furthermore, an introduction to composite bridges and the finite element method
are given.

1.1 Aims of the Study

This thesis aims to study the influence of different design parameters on the lateral
torsional stability of a certain type of composite bridge. For this purpose a close
study of bridge Y288 over the river Ljungan is performed. A detailed description of
this bridge is given in Chapter 2.
The studied parameters are: variations of thickness and attachment of the profiled
sheeting to the flanges, influence of different attachment techniques to the lateral
torsional stability and influence of initial horizontal rotations.
Each of these factors is assumed to have a considerable impact on the torsional
rigidity of the bridge. The study is performed in order to determine their effect and
whether the bridge will have sufficient lateral torsional stability during construction.
This thesis also includes an investigation of the possibilities and shortcomings of
BRIGADE/Plus, a commercial program for finite element analyses.
This thesis also intends to demonstrate the potential of computational software
today that can handle these types of analyses in a reasonably simple manner.

1.2 Structure of the Thesis

An overview of the general structure of this thesis is presented below.

In Chapter 2, the studied composite bridge is presented. Information about the ge-
ometry and material properties is given. A description of the program used for mod-
elling, BRIGADE/Plus, is presented along with a comparison between BRIGADE/
Plus and ABAQUS.

2
 1.3. COMPOSITE BRIDGES

In Chapter 3, the loads used in the bridge design process are presented. A summary
of performed engineering calculations for the lateral torsional stability is given.
In Chapter 4, a selection of common theories are discussed. The subjects are:
theory of thin-walled structures, bending of thin plates, buckling of thin plates,
box girder analysis and torsion of thin-walled open sections. A comparison between
Timoshenko’s analysis of buckling of a simply supported rectangular plate, uniformly
compressed in one direction, and results from a finite element analysis of the same
problem is presented.
In Chapter 5, the procedure of creating models in BRIGADE/Plus is illustrated.
Different analysis procedures are described, along with a discussion about element
choice and ways to improve convergence. Additionally, some of the assumptions for
the model are discussed.
In Chapter 6, the design and analysis of profiled sheeting in shear is presented,
containing the studies of influence of different attachment choices of the profiled
sheeting as well as variations in thickness of them.
In Chapter 7, the design and analysis of lateral torsional stability is presented, con-
taining studies of critical loads for the structure with various initial imperfections
and different attachment choices of the profiled sheeting.
In Chapter 8, conclusions are stated and some recommendations for further research
are suggested.
In Appendix A, some basic commands in ABAQUS are presented.
In Appendix B, the keywords from the models are given.

1.3 Composite Bridges

Steel beams supporting concrete slabs have been used to form the basic super-
structure of large numbers of deck bridges for many years. Since 1945 the number
of composite bridges being built has significantly increased. The pressure of steel
shortage in Germany after the Second World War forced engineers to adopt the
most economical design method available to be able to cope with the large amount
of reconstruction of bridges and buildings destroyed. New codes of practice in other
countries, the publications of papers describing the results of experimental work and
eventually the publication of text books have all helped to make engineers familiar
with the composite construction [14].
Composite bridges are structures that combine materials like steel, concrete, tim-
ber or masonry in some combination. The behaviour of the composite structure is
heavily influenced by the properties of its component materials. For example, the
use of a concrete slab on a steel girder uses the strength of concrete in compression
and the high tensile strength of steel. Looking at the basic behaviour of a com-
posite structure there are two fundamental effects that need to be considered: the

3
CHAPTER 1. INTRODUCTION

differences between the materials and the connection of the two materials. Stronger,
stiffer materials like steel attract proportionally more load than materials such as
concrete. If there is no connection then the materials will behave independently,
omitting the positive effects, but if adequately connected the materials act as one
whole structure.
Most common composite structures are either precast, prestressed concrete beams
with an cast concrete slab or steel girders with a concrete slab. Composite struc-
tures can be used for a wide range of structures such as foundations, substructures,
superstructures and for a diverse range of bridge structures like tunnels, viaducts,
footbridges and cable stayed bridges.
Steel-concrete composite box girders may advantageously be used for bridges with
long spans, for bridges with significant horizontal curvature or simply for aesthetic
reasons. The boxes may be complete steel boxes with an overlay slab or an open
box where the concrete slab closes the top of the box.
The open top form of box girders, consisting of steel webs and a bottom flange, has
only small top flanges sufficient for stability during concreting. The advantages of
this form are that access to all parts of the section is available, which, e.g., facilitates
welding, and that the web can be inclined which allows a larger span in the transverse
direction of the bridge. A disadvantage of the open box is that the high torsional
stiffness of a closed section is not present during construction until the concrete
slab has gained strength, which makes it more sensitive to lateral instability during
construction.
The stresses induced by the loads will depend upon the magnitude of the load and
its eccentricity, the box geometry and the number and stiffness of diaphragms. The
use of a box form will aid the distribution of eccentric loads. Vertical loads that act
eccentrically with respect to the centre line in a box girder results in twisting of the
box section. Twisting moment is resisted by pure shear stresses in the walls of a
box. Longitudinal normal stresses arising from the relative warping of the section
under torsion are not considered in theory of pure torsion. However, these stresses
can attain very large values when the closed cross-sections are flexible. For example,
considering a general loading on a box section, as shown in Figure 1.2, in which a
single vertical eccentric load is replaced by sets of forces representing vertical, tor-
sional and distortional loading. The general loading in Figure 1.2 can be represented
as two different components of loading, one causing bending and the other causing
torsion as shown in Figures 1.2(b) and 1.2(c), respectively. The torsional loading
component can be subdivided further into a pure torsional component and a distor-
tional component as shown in Figures 1.2(d) and 1.2(e), respectively. Although the
pure torsional component will normally result in negligible longitudinal stresses, the
distortional component will always tend to deform the cross-section, thus creating
distortional stresses in the transverse direction and warping stresses in the longitu-
dinal direction. The distortion of the cross-section will be resisted by cross frames
and diaphragms and hence an accurate analysis involves evaluating the distortional
warping and shear stresses and the associated distortional bending stresses in the
transverse frames [14].

4
 1.4. FINITE ELEMENT METHOD

Figure 1.2: Idealisation of eccentric loading in box girder (from [14]).

Where composite action can be achieved, certain structural advantages appear. In

comparison with the non-composite case the advantages may be summarised as: a
reduction in steel area required, an increase in the overload capacity and a reduction
in construction depth. The first two factors will lead to a reduction in the steel weight
required to support a given load and the composite structures will show economy
over their non-composite counterparts [13].

1.4 Finite Element Method

The finite element method as we know it today seems to have originated with
Courant in 1943 [9]. Courant determined the torsional rigidity of a hollow shaft
by dividing the cross-section into triangles and interpolating a stress function φ
linearly over each triangle from the values of φ at nodes.
The name finite element was coined by Clough in 1960. Many new elements for stress
analysis were soon developed. In 1963, finite element analysis acquired respectability
in academia when it was recognised as a form of the Rayleigh-Ritz method. Thus
finite element analysis was seen not just as a special trick for stress analysis but as
a widely applicable method having a sound mathematical basis. The first textbook
about finite element analysis appeared in 1967 and today there exists an enormous
quantity of literature about finite element analysis [19].
General-purpose computer programs for finite element analysis emerged in the late
1960’s and early 1970’s. Since the late 1970’s, computer graphics of increasing
power have been attached to finite element software, making finite element analysis

5
CHAPTER 1. INTRODUCTION

attractive enough to be used in actual design. Previously it was so tedious that

is was used mainly to verify a design already completed or to study a structure
that had failed. Computational demands of practical finite element analysis are so
extensive that computer implementation is mandatory. Analyses that involve more
than 100 000 degrees of freedom are not uncommon [9].
Finite element analysis, also called the finite element method, is a method for nu-
merical solution of field problems. A field problem requires determination of the
spatial distribution of one or more dependent variables. Mathematically, a field
problem is described by differential equations or by an integral expression. Either
description may be used to formulate finite elements.
Individual finite elements can be visualised as small pieces of a structure. In each
finite element a field quantity is allowed to have only a simple spatial variation, e.g.
described by polynomial terms up to x2 , xy and y 2 . The actual variation in the
region spanned by an element is almost certainly more complicated, hence a finite
element analysis provides an approximate solution.
In more and more engineering situations today, we find that it is necessary to ob-
tain approximate numerical solutions to problems, rather than exact closed-form
solutions.
Elements are connected at points called nodes and the assemblage of elements is
called a finite element structure. The particular arrangement of elements is called
a mesh. How the finite element method works can be summarised in the following
general terms [11]:

1. Discretise the continuum. The first step is to divide the continuum or solution
into elements. A variety of element shapes may be used and different element
shapes may be employed in the same solution region.

2. Select interpolation functions. The next step is to assign nodes to each element
and then choose the type of interpolation function to represent the variation
of field variable over the element.

3. Find the element properties. Once the finite element model has been estab-
lished the matrix equation expressing the properties of the individual elements
is ready to be determined.

4. Assemble the element properties to obtain the system equations. The matrix
equations expressing the behaviour of the elements must be combined to form
the matrix equations expressing the behaviour of the entire solution region or
system.

5. Solve the system equations. The assembly process of the preceding step gives a
set of simultaneous equations that can be solved to obtain the unknown nodal
values of the field variable.

Finite element analysis has advantages over most other numerical analysis methods,

6
 1.4. FINITE ELEMENT METHOD

including versatility and physical appeal. The major advantages of finite element
analysis can be summarised as [9]:

• Finite element analysis is applicable to any field problem.

• There is no geometric restriction. The body analysed may have any shape.

• Boundary conditions and loading are not restricted.

• Material properties are not restricted to isotropy and may change from one
element to another or even within an element.

• Components that have different behaviours, and different mathematical de-

scriptions, can be combined.

• A finite element analysis closely resembles the actual body or region.

• The approximation is easily improved by grading the mesh.

Some disadvantages may be mentioned as well:

• It is fairly complicated, making it time-consuming and expensive to use.

• It is possible to use finite element analysis programs while having little knowl-
edge of the analysis method or the problem to which it is applied. Finite
element analyses carried out without sufficient knowledge may lead to results
that are worthless and some critics say that most finite element analysis results
are worthless [9].

7
Chapter 2

Studied Bridge and Used

Computational Tools

2.1 Bridge Y288 over Ljungan

The studied composite bridge, bridge Y288 over Ljungan, is situated at Erikslund
in Ånge municipality in central Sweden. The bridge is a simply supported box
girder bridge with a span of 57 m and a width of 7 m. The girder is a conventional
composite construction containing a flanged box whose lower part is made of a
steel channel on which a concrete carriageway is cast [17]. Representative cross-
sections are given in Figure 2.1. The steel box is provided with diaphragms c/c
7125 mm. It is made in two parts which are transported to the construction site
where they are welded together and launched. The part of the carriageway that is
situated between the flanges is cast on so called “lost form” of profiled sheeting. The
bridge, which is built by the construction company PEAB, has been designed by
the engineering consulting company Tyréns AB on request of the Swedish National
Road Administration, Vägverket.
The materials chosen in the design are specified as follows [1, 2, 6]:

• Steel S275J2G3:


275 MPa for 0 < t < 16 mm
fyk =
265 MPa for 16 < t < 40 mm
fuk = 410 MPa

• Steel S420M:


420 MPa for 0 < t < 16 mm
fyk =
400 MPa for 16 < t < 40 mm
fuk = 500 MPa

• Profiled steel sheeting, Rydab 45-950 F:

fyk = 320 MPa

9
CHAPTER 2. STUDIED BRIDGE AND USED COMPUTATIONAL TOOLS

(a) (b)

Figure 2.1: (a) Cross-section of the box girder midsection and, (b) cross-section of
the box girder ends.

• Concrete BTG K40:

fcck = 28.5 MPa
fctk = 1.95 MPa
fcc = 15.8 MPa
fct = 1.08 MPa
Ec = 32 GPa

The reinforcement and wooden formwork have negligible effect upon the mechanical
behaviour of the bridge during construction and only their addition to the dead load
is studied.

2.2 Computational Tools

A general description of BRIGADE/Plus is given, presenting an insight to how the

program works and the possibilities it provides. Beside this description a comparison
of BRIGADE/Plus and ABAQUS/Standard is given.

2.2.1 BRIGADE/Plus

BRIGADE/Plus is a general-purpose finite element program providing powerful

analysis features in an interactive and visual environment. The geometry defini-
tion of each part is parametric and feature based, which allows quick modifications.
Parts can be created in numerous ways and they are then assembled together to cre-
ate the analysis model. Section and material properties can be defined and assigned
to regions of the parts. The program offers a range of analysis procedures, such

10
 2.2. COMPUTATIONAL TOOLS

as: static stress analysis, eigenvalue buckling analysis and collapse and postbuckling
analysis.
There are also possibilities to take nonlinear behaviour into consideration, including
geometric, material and contact nonlinearity. The loading can be created combin-
ing concentrated, distributed and pressure loads and body forces. BRIGADE/Plus
contains advanced algorithms for automatic meshing regions and the density of the
mesh can be controlled by applying mesh seeds globally and locally. The program
also includes an extensive element library, including element families such as:

• Solid elements
• Shell elements
• Membrane elements
• Beam elements
• Truss elements
• Spring elements
• Rigid elements

Finally BRIGADE/Plus provides a suite of postprocessing features in order to enable

efficient interpretation of results [3].

2.2.2 Comparison of BRIGADE/Plus and ABAQUS

The differences between ABAQUS/Standard [10] and BRIGADE/Plus [3] can be

done by describing the way in which BRIGADE/Plus is created. BRIGADE/Plus
consists, in short, of three parts

• A solver based on ABAQUS/Standard

• GUI (Graphical User Interface) based on ABAQUS/CAE
• Technology developed at Scanscot Technology

The solver is a subset of ABAQUS/Standard where functionalities only interesting

to other industries than the construction industry have been removed. The corre-
sponding functionalities have also been removed from the GUI of BRIGADE/Plus.
Functionalities concerning e.g. moving loads and advanced load combinations that
is available in BRIGADE/Standard today will soon be available in BRIGADE/Plus.
The advantages of BRIGADE/Plus are:

• It is particularly developed for the construction industry.

11
CHAPTER 2. STUDIED BRIDGE AND USED COMPUTATIONAL TOOLS

• Possibilities to apply moving loads and advanced load combinations will be

provided.

The major disadvantage concerns the removed functionalities which might be needed
to create the desired models. Then, of course, the full ABAQUS version is a more
suitable choice [4].

12
Chapter 3

Loads and Engineering Analysis

3.1 Loads Acting on the Bridge

The loads and load combinations that are considered in the engineering design pro-
cess for a bridge are defined in accordance with the Swedish design code BRO94, [18],
as: the dead load, the surfacing, the earth pressure, the traffic loads, the braking
forces, the load on the road embankment and the additional earth pressure.
These loads are all, of course, important to consider while designing the structure
but the calculations of these loads are beyond the scope of this thesis. Instead, due
to the low torsional warping restraint of the bridge during construction, the loads
acting on the structure during construction are of greater interest. These loads can
be divided into:

• dead loads

• live loads

where the dead loads consist of the steel box girder self-weight, the formwork and
the reinforcement, while the live loads are wind and concrete during casting. The
contributions of these loads are described in subsequent sections.

3.1.1 Dead Loads

Construction design calculations have been used for calculations of the bridge self-
weight.

Box Girder

The box girder self-weight is given by its cross-sectional area multiplied by its length.
The cross-sectional area is divided into three sections where the steel area, AS (x)

13
CHAPTER 3. LOADS AND ENGINEERING ANALYSIS

(a) (b)

Figure 3.1: (a) Cross-section of the box girder midsection and, (b) cross-section of
the box girder ends.

is [15]: 
 0.15233986 m2 if 0 < x ≤ 10 m
AS (x) = 0.20633676 m2 if 10 < x < 47 m

0.15233986 m2 if 47 ≤ x < 57 m

Representative cross-sections are shown in Figure 3.1, displaying their dimensions.

The total self-weight for the box girder is given as:
 L
qs (x)
Qsteel = dx = 92.4 · 103 kg
0 gγ
where γ is a load factor, g is the gravity and qs (x) is the load-intensity.
γ = 1.00
g = 9.81 m/s2

qs (x) = ρAS (x)gγ + 150gγ = 7850 · AS (x) · 9.81 · 10−6 · 1.00 + 150 · 9.81 · 10−6 · 1.00
with ρ = 7850 kg/m3 and the contribution from diaphragms = 150 kg/m.

Formwork

The bridge is assumed to be loaded by the wooden formwork over the full length
by:
qformwork = 5 kN/m

Reinforcement

The contribution from the reinforcement is included in the concrete weight. The
weight of the concrete, including the reinforcement, is 2400 kg/m3 .

14
 3.1. LOADS ACTING ON THE BRIDGE

3.1.2 Live Loads

The live loads that act upon the bridge during construction, that significantly affect
the mechanical behaviour, are the wind load and the cast concrete.

Wind Load

The wind load has been calculated according to the Swedish regulations [5]. Its
impact on the bridge is given by:
 L
Qwind = qwind (x)dx = 170 kN
0

where:
qwind (x) = ψγwind µtot qk,wind Hbridge = 2.99 kN/m
with µtot = 2.0, qk,wind = 0.5 kN/m2 , ψγwind = 1.3 and Hbridge = 2.3 m

Cast Concrete

The casting is performed in two or three stages, depending on whether warping

restraint casting of bridge ends is performed or not.
The casting of the midsection of the bridge is performed in several steps, where
about 10 m3 of concrete is cast per step. The cast concrete will contribute to the
loads with
qcasting = ρgAc (x) = 24 · 2.032 = 48.768 kN/m
where ρg = 24 kN/m3 and Ac (x) is the concrete carriageway cross-sectional area.
The concrete is cast starting from one end of the midsection and proceeds towards
the other end, as shown in Figure 3.2.

15
CHAPTER 3. LOADS AND ENGINEERING ANALYSIS

(2)
(1)

Figure 3.2: Casting of bridge Y288. (1) is the starting-point for the casting of the
central part and (2) is the end-point. The starting-point and end-point
are located 14.25 m from the ends of the bridge.

3.2 Engineer Calculations of Stability

Calculations concerning loads acting on the completed bridge are performed ac-
cording to existing codes and regulations. More relevant for this study though are
calculations concerning the bridge response to loads during construction. The risk
of lateral instability is imminent until the concrete has hardened and it is therefore
vital to check the stability during construction. These calculations for bridge Y288
over the river Ljungan were carried out using second-order differential equations
considering assumed imperfections. The imperfections are assumed to exist as an
initial rotation φ0 (x) and as an initial inclination in the horizontal plane v0 (x). The
initial rotation is assumed to be sinusoidal with a maximum of about 0.3%, or 1◦
and the initial inclination is also assumed sinusoidal with a maximum amplitude of
30 mm.
The response of the bridge is governed by the following differential equations:
  
d2 d2 φ d dφ
EKw 2 − (GKv + KwC )
dx2 dx dx dx
    
d d dv d dw
− My + Mz = mtwist (3.1a)
dx dx dx dx dx

d2 v
EIz + My (φ + φ0 ) = Mz (x) (3.1b)
dx2

where Kw is the sector moment of inertia (m6 ), GKv is the torsional rigidity (Nm2 )
and KwC is the Wagner coefficient (Nm2 ).

16
 3.2. ENGINEER CALCULATIONS OF STABILITY

From these equations the rotations, deflections and stresses are calculated, and
subsequently the ultimate load. For a complete understanding of the calculations
see [15].

17
Chapter 4

Thin-Walled Structures

In this chapter a brief literature review of thin-walled structures and an overview of

bending and buckling of thin plates is presented together with the theories of torsion
of thin-walled open sections. Finally an analysis by Timoshenko of buckling of a
simply supported rectangular plate is presented and compared with results from a
finite element analysis of the same problem made in BRIGADE/Plus.
A thin-walled structure is defined as a structure that is made from thin plates joined
along their edges. The plate thickness is small compared to other cross-sectional
dimensions which are in turn often small compared to the overall length [16].
There are several reasons why thin-walled structures must be given special consid-
eration in their analysis and design. In a thin-walled beam the shear stresses and
strains are much larger relative to those in a solid rectangular beam.
When certain thin-walled structures are twisted there is a so-called warping of the
cross-section and the Bernoulli hypothesis1 is violated. The term warping is defined
as the out-of-plane distortion of the cross-section of a beam in the direction of its
longitudinal axis. This is further discussed in section 4.5.
Thin-walled structures are also susceptible to local buckling if the in-plane stresses
reach their critical values. If this happens, the geometry of the cross-section changes,
in contrast to overall buckling where the cross-sectional form is retained, as in the
case of a pin-ended column. However, if a thin-walled column is made sufficiently
long it may suffer overall buckling before it buckles locally. This means that thin-
walled structures must be designed against both local and overall buckling. Theory
and experiments show that these two phenomena can interact and when this happens
the buckling load can decrease below the values of the individual loads.
1
The Bernoulli hypothesis states that plane sections remain plane over the entire cross-section,
from Jakob Bernoulli (1654–1705)

19
CHAPTER 4. THIN-WALLED STRUCTURES

4.1 Brief Literature Review

There is an extensive amount of literature dealing with the theory of thin-walled

structures and it is beyond the scope of this thesis to review this literature to any
larger extent. However, a brief review of some literature on the elastic buckling of
thin-walled structures is given below.
The membrane theory of plates was first studied by Euler (1766) and the flexural
theory by Bernoulli (1789) and Navier (1823). The theory for combined membrane
and flexural effects was developed by Kirchhoff (1877) and Saint-Venant (1883). At
this state, the governing equation for thin isotropic plates loaded laterally with q
per unit area and in-plane forces Nx , Nz and Nxz per unit length was
 4
∂ w ∂4w ∂ 4w ∂2w ∂ 2w ∂ 2w
D + 2 + = q + N x + 2N xz + N z (4.1)
∂x4 ∂x2 ∂z 2 ∂z 4 ∂x2 ∂x∂z ∂z 2
where w is the lateral deflection [12].
At the turn of the twentieth century the equation governing the buckling of flat
plates was available and it was known that it forms the basis of an eigenvalue
problem. At that time it was not recognised that as the plate buckles the values of
Nx , Nz and Nxz at a given point would vary because of the stretching of the plate.
The next development which overcame this deficiency was due to Föppl (1907) who
introduced the stress function Φ and paved the way for von Karman (1910) to derive
the governing equations for perfectly flat plates [12]:
 4
∂ w ∂4w ∂4w ∂ 2Φ ∂ 2w ∂ 2Φ ∂ 2w ∂ 2Φ ∂ 2w
D +2 2 2 + = + −2 +q (4.2)
∂x4 ∂x ∂z ∂z 4 ∂z 2 ∂x2 ∂x2 ∂z 2 ∂x∂z ∂x∂z

 2 
∂4Φ ∂4Φ ∂ 4Φ ∂ 2w ∂ 2w ∂ 2w
+ 2 + + = Et − (4.3)
∂x4 ∂x2 ∂z 2 ∂z 4 ∂x∂z ∂x2 ∂z 2

These equations enabled the post-buckling behaviour of perfectly flat plates to be

studied.
The von Karman large-deflection equations for flat isotropic plates with in-plane
loading were modified to account for anisotropy by Rostovtsev (1940) and later the
effect of initial imperfections were included resulting in the following simultaneous
equations:
∂4w ∂4w ∂4w ∂ 2 Φ ∂ 2 (y + w)
Dx + 2H + D z = +
∂x4 ∂x2 ∂z 2 ∂z 4 ∂z 2 ∂x2 (4.4)
∂ 2 Φ ∂ 2 (y + w) ∂ 2 Φ ∂ 2 (y + w)
− 2 +q
∂x2 ∂z 2 ∂x∂z ∂x∂z
 
1 ∂4Φ 1 νx νz ∂4Φ 1 ∂4Φ
+2 − − + =
tz Ez ∂x4 Kxz tx Ex tz Ez ∂x2 ∂z 2 tx Ex ∂z 4
 2 2
∂ 2y ∂ 2w ∂ 2y ∂ 2w ∂ 2y ∂ 2w ∂ 2w ∂ 2w ∂ w
2 2
+2 − 2 2 − 2 2
+ (4.5)
∂z ∂x ∂x∂z ∂x∂z ∂x ∂z ∂z ∂x ∂x∂z

20
 4.2. BENDING OF THIN PLATES

These appear to be the most general equations currently available for solving plate
buckling problems [12].
Bryan (1891) was the first to solve the problem of a simply supported rectangu-
lar plate with two opposite sides carrying uniform compressive loads. The same
problem was later solved by Timoshenko (1907). This is described in detail in sec-
tion 4.6 alongside a comparison of results from Timoshenko’s analysis and finite
element analysis in BRIGADE/Plus. Timoshenko also analysed many other cases
of flat plates with different boundary conditions. Following the advent of the finite
difference and the relaxation technique and later with the increasing use of comput-
ers and finite elements it has become relatively easy to solve this problem for a wide
variety of plate shapes and stress distributions.

4.2 Bending of Thin Plates

The structural analysis of a plate is carried out by considering the state of stresses
at the middle plane of the plate. All the stress component are expressed in terms of
the deflection w(x, y) of the plane. This deflection function has to satisfy a linear
partial differential equation which, together with its boundary condition, completely
defines w(x, y).
Figure 4.1 shows a plate element cut from a plate whose middle plane coincides with
the xy plane. The middle plane of the plate is subjected to a lateral load of intensity
q. It can be shown, by considering the equilibrium of the plate element, that the
stress resultants are given as:
 2 
∂ w ∂ 2w
Mx = −D +ν 2 (4.6)
∂x2 ∂y

 
∂2w ∂ 2w
My = −D + ν (4.7)
∂y 2 ∂x2

dx

Qy
My
dy
M yx
Qx x
Mx δMx
Mx + dx
δx
Mxy + δMxy dx
δx
Mxy z

M yx + δMyx dy δQ x
δy Qx +
δx
dx

My + δMy dy
δy
δQ y
Qy + dy
δy
y

Figure 4.1: Plate element (from [14]).

21
CHAPTER 4. THIN-WALLED STRUCTURES

∂2w
Mxy = −Myx = D(1 − ν) (4.8)
∂x∂y

∂3w ∂3w
Vx = + (2 − ν) (4.9)
∂x3 ∂x∂y 2

∂ 3w ∂ 3w
Vy = + (2 − ν) (4.10)
∂y 3 ∂y∂x2
 
∂ ∂2w ∂2w
Qx = −D + (4.11)
∂x ∂x2 ∂y 2
 
∂ ∂2w ∂2w
Qy = −D + (4.12)
∂y ∂x2 ∂y 2

∂2w
R = 2D(1 − ν) (4.13)
∂x∂y

where Mx and My are the bending moments per unit length in the x and y directions,
respectively. Mxy and Myx are the twisting moments per unit length. Qx and Qy
are the shearing forces per unit length in the x and y directions, respectively. Vx
and Vy are supplementary shear forces in the x and y directions, respectively and R
is the corner force. D = Eh3 /12(1 − ν 2 ) which is flexural rigidity of the plate per
unit length, E is the modulus of elasticity, h is the thickness of the plate and ν is
Poisson’s ratio.
The governing equation for the plate is obtained as:

∂4w ∂4w ∂4w q

4
+2 2 2 + 4
= (4.14)
∂x ∂x ∂y ∂y D

Any plate problem should satisfy the governing equation (4.14) and boundary con-
ditions of the plate [14].

4.3 Buckling of Thin Plates

Buckling of a plate involves out-of-plane movement of the plate and results in bend-
ing in two planes. A significant difference between axially compressed columns and
plates is apparent if their buckling characteristics are compared. For a column,
buckling terminates the ability of the member to resist axial load, and the critical
load is thus the failure load of the member. However, the same is not true for plates
due to the membrane action of the plate. Subsequent to the critical load, plates
under compression will continue to resist increasing axial force, and will not fail
until a load considerably in excess of the critical load is reached. The critical load

22
 4.3. BUCKLING OF THIN PLATES

of a plate is therefore not its failure load. Instead, the load-carrying capacity of a
plate must be determined by considering its postbuckling behaviour.
To determine the critical in-plane loading of a plate by the concept of neutral equi-
librium a governing equation in terms of biaxial compressive forces Nx and Ny and
constant shear force Nxy as shown in Figure 4.2 can be derived as:
 4 
∂ w ∂ 4w ∂ 4w ∂2w ∂2w ∂2w
D + 2 + + N x + N y + 2N xy =0 (4.15)
∂x4 ∂x2 ∂y 2 ∂y 4 ∂x2 ∂y 2 ∂x∂y

Numerical methods of analysis of plates, which include both geometric and material
non-linearities are available today and these analyses are capable of assessing the
ultimate strength and post-critical stiffness of plates with fabrication imperfections.
A square element, as shown in Figure 4.3, whose edges are oriented at 45◦ to the
edges of a plate subjected to pure shear, experiences tensile stresses on two opposing
edges and compressive ones on the other two. These compressive stresses induce a
form of local buckling with elongated bulges oriented at about 45◦ to the plate
edges. As with the compressive loading, a thin plate loaded in shear can support
an applied stress well in excess of the elastic critical one. This is due again to the
resistance to in-plane deformation. As the applied shear stress is increased beyond
τcr the plate buckles elastically and retains little stiffness in the direction in which
the compressive component acts. However, the inclined tensile component is still
resisted fully by the plate. The inclined buckles become progressively narrower and
the plate acts like a series of bars in the tension direction, developing a so-called
tension field. Further increase of applied stress causes plastic deformation in the part
of the tension field, which rotates to line up more closely with the plate diagonal.
Tension field action is particularly important in plate and box girders, in which the
function of the web plates is primarily to resist shear [14].

Figure 4.2: Plate subjected to in-plane forces.

23
CHAPTER 4. THIN-WALLED STRUCTURES

Figure 4.3: Plate subjected to pure shear; stresses on square element at 45o .

4.4 Plate and Box Girder Analysis

The high bending moments and shearing forces associated with the carrying of
large loads over long spans as in the case of bridges frequently necessitates the
use of fabricated plate and box girders. In their simplest form, plate and box
girders can be considered as an assemblage of webs and flanges. In order to reduce
the self-weight of these girders and thus achieve economy, slender plate sections
are employed. Hence local buckling and postbuckling reserve strength of plates
are important design criteria. Flanges in a box girder and webs in plate and box
girders are often reinforced with stiffeners to allow for efficient use of thin plates.
The designer has to find a combination of plate thickness and stiffener spacing
that will result in the most optimal section with reduced weight and fabrication
cost. There are some difficulties that are usually encountered by designers of plated
structures [14]:

• The engineer’s simple ‘plane sections remain plane’ theory of bending is no

longer adequate, even for linear elastic analysis.
• Non-linear elastic behaviour caused by the buckling of plates can be of great
importance and must be allowed for.
• Because of this complex non-linear elastic behaviour, and also because of stress
concentration problems, some yielding may occur at loads which are quite low
in relation to ultimate collapse loads. While such yielding may not be of great
significance as regards rigidity and strength, it means that simple maximum
stress criteria are no longer sufficient.
• Because of the buckling problem in plates and stiffened panels, complete plas-
tification is far from being realised at collapse. Hence simple plastic criteria
are also not sufficient.
• Complex interactions occur between flanges, webs and diaphragms and the
pattern of this interaction can change as the level of load increases.

4.5 Torsion of Thin-Walled Open Sections

Thin-walled open cross-sections composed of slender plates are particularly suscepti-

ble to lateral torsional buckling, because the torsional rigidity of such cross-sections

24
 4.5. TORSION OF THIN-WALLED OPEN SECTIONS

are low and so their resistance to torsional instability is limited.

The analysis of lateral torsional buckling behaviour of beams is more complex than
that of in-plane buckling behaviour of columns because the lateral buckling problem
is intrinsically three-dimensional. The problem is further complicated because the
lateral (out-of-plane) deflection and twisting are coupled, so this coupling effect must
be considered in the analysis

4.5.1 Uniform Torsion of Thin-Walled Open Sections

When an equal and opposite torque T is applied to the ends of a simply supported
beam with a thin-walled open section, the twisting moment along the length of the
members is constant and the beam is said to be under a uniform torsion. Under the
action of the torque, warping of the cross-section will occur and plane sections of
the cross-section no longer remain plane as a result of the uneven axial deformation
that takes place over the entire cross-section.
For the simply supported beam, in which warping of all the cross-sections is un-
restrained, the applied torque is resisted only by shear stresses developed in the
cross-section. These stresses act parallel to the edge of the component plates of the
cross-section, as shown in Figure 4.4. The distribution of these shear stresses is the
same for all thin-walled, open cross-sections. The magnitude of these shear stresses
will be proportional to the distance from the midline of the component plate. These
shear stresses are called Saint-Venant shear stresses and the associated torsion is
referred to as Saint-Venant torsion, Tsv . The angle of twist γ over the length L
caused by the Saint-Venant torsion is given by
γ Tsv
= (4.16)
L GJ

where γ/L is the angle of twist per unit length, G is the shear modulus and J is the
torsional constant of the cross-section.

Figure 4.4: Saint-Venant shear stress distribution in an I-section (from [7]).

25
CHAPTER 4. THIN-WALLED STRUCTURES

The rate of twist is expressed as

dγ Tsv
= (4.17)
dz GJ

where z is the coordinate axis along the length of the beam. Equation (4.17) can
be written as
dγ
Tsv = GJ (4.18)
dz

The Saint-Venant torsion expressed in Equation (4.18) is also referred to as uniform

or pure torsion.

4.5.2 Non-Uniform Torsion of Thin-Walled, Open Sections

Consider a cantilever beam subjected to a torque applied at the free end. At the
free end the cross-section is free to warp, so the applied torque is resisted only by
Saint-Venant torsion. At the fixed end, however, warping is prevented. As a result,
in addition to Saint-Venant torsion, there exists another type of torsion known
as warping restraint torsion in the cross-section. If the cross-section is prevented
from warping, axial strain and axial stresses must be induced in the cross-section,
in addition to the shear stresses. These induced axial stresses are in self-balance
because no external axial force is applied to the beam.
The resultant of these axial stresses in the two flanges constitutes a pair of equal
moments called the bi-moment, Mf , acting oppositely in each of the two planes of
the flanges.
The development of these bending moments, or bi-moments, in the flanges in the
cross-section is shown in Figure 4.5. Since warping is prevented at the fixed end, the
two flanges of the beam must bend in opposite directions as the cross-section rotates
under the action of the applied torque. The bending of the flanges will thus induce
bending moments Mf at the fixed end. The bending moment can be expressed in
terms of the lateral displacement uf as

d 2 uf
Mf = EIf (4.19)
dz 2

where E is the modulus of elasticity, If the moment of inertia of one flange about the
y axis of the cross-section, and uf the lateral displacement of the flange. Associated
with the bending moment in one flange is the shear force Vf given by

dMf d3 uf
Vf = − = −EIf (4.20)
dz dz 3

The shear forces are present in both flanges of the I-section. They are equal in
magnitude but act in opposite directions, as shown in Figure 4.6. This pair of shear

26
 4.5. TORSION OF THIN-WALLED OPEN SECTIONS

Top flange

T T y

Bending stresses
in top flange
Bottom flange

Figure 4.5: Bending of flanges due to warping restraint at the fixed end (from [7]).

forces constitute a couple acting on the cross-section. The resulting torsion, which
is referred to as the warping restraint torsion or non-uniform torsion, is given by

Tw = Vf h (4.21)

where h is the distance between the lines of action of the shear forces. Equa-
tion (4.21) can be expressed as

h2 d3 γ d3 γ
Tw = −EIf = −ECw (4.22)
2 dz 3 dz 3

where
If h2
Cw = (4.23)
2

is called the warping constant of the I-section. The warping constant is different for
different cross-sections.

Top flange z

Vf

Mf

h
x

Bottom flange

Vf Mf

Figure 4.6: Moment and shear developed at the fixed-end cross-section of an I-

section (from [7]).

27
CHAPTER 4. THIN-WALLED STRUCTURES

If warping is restrained, the applied twisting moment will be resisted by both Saint-
Venant torsion and warping restraint torsion.

T = Tsv + Tw (4.24)

or
dγ d3 γ
T = GJ − ECw 3 (4.25)
dz dz

Equation (4.25) represents the internal twisting moment that will develop in the
cross-section when the member is twisted. The first term represents the resistance
of the cross-section to twisting and the second term represents the resistance of the
cross-section to warping.
Saint-Venant torsion is always present when a member is subjected to twisting and
rotates. Warping restraint torsion will develop if a cross-section is prevented from
warping when it is being twisted [7].

4.6 Buckling of Simply Supported Rectangular

Plate

A comparison between Timoshenko’s theory and results obtained by BRIGADE/Plus

is presented in this section. The comparison has been made for buckling of a simply
supported rectangular plate that is uniformly compressed in one direction. First,
results, and how they are achieved, using Timoshenko’s theory [16] are presented.
Then, these results are compared with the results from the finite element analysis
using BRIGADE/Plus.
A rectangular plate is compressed in its middle plane by forces uniformly distributed
along the sides x = 0 and x = a, as shown in Figure 4.7.
The magnitude of the compressive force per unit edge length is denoted by Nx . By
gradually increasing Nx , one arrive at the condition where the flat form of equilib-
rium of the compressed plate is disturbed and buckling occurs. The corresponding
critical value of the compressive force can be found in this case from a consideration
of the energy of the system. The deflection surface of the buckled plate can be
represented by the double series

∞ 
∞
mπx nπy
w= amn sin sin (4.26)
m=1 n=1
a b

The strain energy of bending in this case is

 2 2
π 4 ab   2
∞ ∞
m n2
∆V = D amn 2
+ 2 (4.27)
8 m=1 m=1
a b

28
 4.6. BUCKLING OF SIMPLY SUPPORTED RECTANGULAR PLATE

a
x

b
Nx

Figure 4.7: Simply supported rectangular plate uniformly compressed in one direc-
tion.

The work done by compressive forces during buckling of the plate will be
 a  b 2 ∞  ∞
1 ∂w π2b
Nx dxdy = Nx m2 a2mn (4.28)
2 0 0 ∂x 8a m=1 n=1

Thus, the equation for determining the critical value of compressive forces, becomes
∞   2 2
π 4 ab   2
∞ ∞ ∞
π2b 2 2 m n2
Nx m amn = D amn 2
+ 2 (4.29)
8a m=1 n=1
8 m=1 m=1
a b

from which  2 2
∞ ∞ m n2
π 2 a2 D m=1 a 2
n=1 mn a2
+ b2
Nx = ∞ ∞ 2 2
(4.30)
m=1 n=1 m amn

It can be shown that Equation (4.30) becomes a minimum if all coefficients amn ,
except one, are taken equal to zero. Thus
 2
π 2 a2 D m 2 n2
Nx = + 2 (4.31)
m2 a2 b

The smallest value of Nx will be obtained by taking n equal to 1. The physical

meaning of this is that a plate buckles in such a way that there can be several half-
waves in the direction of compression but only one half-wave in the perpendicular
direction. Thus the expression for the critical value of the compressive force becomes
 2
π2D 1 a2
Nx,cr = 2 m+ (4.32)
a m b2

The first factor in this expression represents the Euler load for a strip of unit width
and of length a. The second factor indicates in what proportion the stability of the
continuous plate is greater than the stability of an isolated strip. The magnitude
of this factor depends on the magnitude of the ratio a/b and also on the number

29
CHAPTER 4. THIN-WALLED STRUCTURES

4
k

m=1 m=2 m=3 m=4

0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5
a/b

Figure 4.8: Buckling stress coefficients for uniaxially compressed plate [14].

m, which gives the number of half-waves into which the plate buckles. The critical
load, with m = 1, can be written as
 2
π2D b a
Nx,cr = 2 + (4.33)
b a b

Equation (4.33) can be represented in the form

π2D
Nx,cr = k (4.34)
b2

in which k is a numerical factor, the magnitude of which depends on the ratio a/b.
This factor is represented in Figure 4.8 by the curve marked m = 1. Assuming that
the plate buckles into two half-waves and that the deflection surface is represented
by
2πx πy
w = a21 sin sin (4.35)
a b

provides an inflection line dividing the plate in halves and each half is in exactly the
same condition as a simply supported plate of length a/2. Using Equation (4.33)
for calculating the critical load and substituting a with a/2 in it gives
 2
π2D 2b a
Nx,cr = 2 + (4.36)
b a 2b

The second factor in this expression, depending on the ratio a/b, is represented
in Figure 4.8 by the curve m = 2. The curve m = 2 is obtained from the curve

30
 4.6. BUCKLING OF SIMPLY SUPPORTED RECTANGULAR PLATE

m = 1 by keeping the ordinates unchanged and doubling the abscissas. Proceeding

further in the same way and assuming m = 3, m = 4 and so on, the series of curves
shown in Figure 4.8 are obtained. Having these curves, the critical load and the
number of half-waves for any value of the ratio a/b can be determined by taking
the corresponding point on the axis of abscissas and to choose the curve having the
smallest ordinate for that point. In Figure 4.8 the portions of the curves defining
the critical values of the load are shown by full lines. For very short plates the curve
m = 1 gives the smallest ordinate. Beginning with the point of intersection of the
curves m = 1 and m = 2, the second curve has the smallest ordinate, i.e., the plate
buckles into two half-waves. Beginning from the point of intersection of the curves
m = 2 and m = 3, the plate buckles into three half-waves, and so on. The transition
from m to m + 1 half-waves occurs when the two corresponding curves have equal
ordinates, i.e., when

mb a (m + 1)b a
+ = + (4.37)
a mb a (m + 1)b

or
a 
= m(m + 1) (4.38)
b

After the number of half-waves m in which the plate buckles has been determined,
the critical load is calculated from Equation (4.33).
The critical value of the compressive stress is then given by

Nx,cr kπ 2 E h2
σcr = = (4.39)
h 12(1 − ν 2 ) b2

where h is the thickness of the plate.

4.6.1 Comparison of Theory and BRIGADE/Plus

A comparison of this theory and results provided by BRIGADE/Plus is given below

for three plates with ratios of a/b of 1, 2 and 3. Hence, for these ratios, the critical
value of compressive stress is equal for all of the plates, since the value of k is 4 (in
accordance with Figure 4.8). This value is

Nx,cr π 2 E h2
σcr = = (4.40)
h 3(1 − ν 2 ) b2

All of these comparisons are made for plates with a width of 1 m and a thickness
of 10 mm. The modulus of elasticity is assumed E = 206.9 GPa and ν = 0.3.
The models have been made of different number of elements and different element
types, giving an indication of the influence of different modelling parameters. The
load is applied as concentrated forces at 9, 17 and 65 points along the plate width
depending on the mesh size. The load is applied as consistent nodal loads for two

31
CHAPTER 4. THIN-WALLED STRUCTURES

of the models: the one with load applied at 9 nodes and a mesh of 256 elements/m2
and the one with load applied at 65 nodes and a mesh of 4096 elements/m2 . For the
other models the load is not applied to all nodes along the edge, which leads to a less
accurate result. Their magnitudes are 1 N, 0.5 N and 0.125 N, respectively, giving
a total applied load of 8 N along each side of the plate. According to Timoshenko’s
theory the value of the critical buckling stresses should be

σcr = 74.8 MPa

The results from BRIGADE/Plus are presented in Tables 4.1 and 4.2. In Table 4.1,
the analyses have been performed using the S4R element, while Table 4.2 shows the
differences for some of the analyses using other element types.

Table 4.1: Critical compressive stresses for buckling of thin plates, using S4R ele-
ments.
Stresses (MPa)
2
Elements/m Applied load aspect ratio (a/b)
1 2 3
256 1N on 9 nodes 69.9 68.3 67.7
1024 1N on 9 nodes 69.4 67.8 67.3
1N on 9 nodes 69.2 67.6 67.1
4096 0.5N on 17 nodes 71.6 70.9 70.6
0.125N on 65 nodes 73.6 73.5 73.4

Table 4.2: Critical compressive stresses for buckling of thin plates, using different
element types and with 4096 elements per m2 .
Element Stresses (MPa)
type Applied load aspect ratio (a/b)
1 2 3
1N on 9 nodes 69.2 67.6 67.1
S4 0.5N on 17 nodes 71.6 70.9 70.6
0.125N on 65 nodes 73.6 73.5 73.4
S8R5 1N on 9 nodes 69.2 67.6 67.1
S8R 1N on 9 nodes 69.0 67.5 67.0

32
 4.6. BUCKLING OF SIMPLY SUPPORTED RECTANGULAR PLATE

In Figures 4.9-4.11 the first buckling mode for ratio a/b equal to 1, 2 and 3, respec-
tively, are shown.

Figure 4.9: First buckling mode for thin rectangular plate uniformly compressed in
one direction. Aspect ratio a/b = 1.

Figure 4.10: First buckling mode for thin rectangular plate uniformly compressed
in one direction. Aspect ratio a/b = 2.

Figure 4.11: First buckling mode for thin rectangular plate uniformly compressed
in one direction. Aspect ratio a/b = 3.

33
CHAPTER 4. THIN-WALLED STRUCTURES

There is a small deviation in the results from BRIGADE/Plus and the analytical
ones. This is because in the models the load is applied as concentrated forces along
the edges of the plate while there is a uniform compressive force acting in the theo-
retical solution. It is also shown that the numerical solution from BRIGADE/Plus
tends to approach the theoretical when the load is applied to more nodes. The
best result is reached for the model with a mesh of 4096 elements/m2 and the load
applied as consistent nodal load at 65 nodes. The discrepancy from the analytical
solution in this case is only 1.6%.
The models give the best results for the ratio a/b = 1, due to the increased margin
of error when the model becomes larger. With more elements a more accurate result
is found. But more elements require more computational costs, hence the number
of elements should be chosen in such a way that the accuracy is satisfactory with as
low as possible computational cost.
Which type of element to use is not an obvious choice. There is no direct correlation
between an element using more nodes and a more accurate result, at least not when
a reasonably dense mesh is used. Instead the emphasis should lie in choosing an
element type that provides the most accurate result in the least computational
time.

34
Chapter 5

Creating Models in
BRIGADE/Plus

The procedure of creating finite element models in BRIGADE/Plus are explained in

this chapter. Aspects like choice of elements and analysis procedure are discussed.
The buckling phenomenon is studied closely and an explanation of made assumptions
is given.

5.1 Modelling Procedures in BRIGADE/Plus

BRIGADE/Plus offers a wide range of possibilities for finite element analysis of

various problems. Here are the ones used to analyse bridge Y288 over Ljungan
presented. The models have been created in a procedure as follows:

1. The parts that form the model were drawn in the GUI.

2. The different parts were given material properties, such as modulus of elastic-
ity, density, thickness and Poisson’s ratio.

3. The parts were assembled to form the structure that was analysed.

4. Dividing the analysis into steps, there are two different types of steps: general
analysis steps and linear perturbation steps. General analysis steps define
sequential events: the state of the model at the end of one general step provides
the initial state for the start of the next general step. Linear perturbation
analysis steps provide the linear response of the model about the state reached
at the end of the last general analysis step. For each step in the analysis it
is possible to choose whether ABAQUS will account for nonlinear effects from
large displacements and deformations or not. If the displacements in a model
due to loading are relatively small during a step, the effects may be small
enough to be ignored. However, in cases where the loads on a model result in
large displacements, nonlinear geometric effects can become important.

35
CHAPTER 5. CREATING MODELS IN BRIGADE/PLUS

General analysis steps and linear perturbation steps are further discussed in
section 5.3 and 5.4, respectively.

5. The assembled parts were connected to each other. These connections were
created using the tie constraint and surface-to-surface contact interactions.
The tie constraint ties two separate surfaces together so that there is no relative
motion between them. It constrains each of the nodes on the slave surface to
have the same value of displacement as the point on the master surface to
which it is closest. The surface-to-surface contact interaction establishes an
interaction between two surfaces, thus preventing them from sliding through
each other.

6. Loads and boundaries were specified. Loads can be applied as concentrated

forces, pressure loads and gravity loads. The boundaries can be defined as sym-
metry/antisymmetry/encastre boundaries or displacement/rotational bound-
aries, and can be applied to nodes or surfaces.

7. The model was meshed. Applying a mesh to the model involves choices of
element type and size. There is a variety of element types to choose from and
their size may vary for different parts of the model. The choice of element is
further discussed in section 5.5.

8. The model is submitted for analysis.

9. A critical evaluation of the results needs to be done by the user. The program
provides an enormous amount of results and it is vital that these are interpreted
correctly.

5.2 Buckling in General

Buckling means loss of the stability of an equilibrium configuration, without frac-

ture or separation of the material or at least prior to it [9]. Generally there are two
types of buckling: Bifurcation buckling and Snap-trough buckling. Bifurcation buck-
ling is the kind of buckling familiar from elementary column theory. For an axial
compressive load of magnitude Pcr , called the critical load, the straight prebuckling
configuration ceases to be a stable state of equilibrium and an alternative buck-
led configuration is also possible. Buckling may also appear without bifurcation,
as a limit point, where there is no alternative and infinitesimally close equilibrium
configuration.
A primary path is the original load-displacement line or curve and its extension.
The secondary path is the alternative path that originates when the critical load
is reached. The two paths intersect at the bifurcation point. Past the bifurcation
point, the primary path is unstable. Although it is possible mathematically that the
structure follows the primary path, a real structure will follow the secondary path
instead. If the secondary path has a positive derivative (rises), the structure has
post-buckling strength. A limit point is a maximum on a load-displacement curve.

36
 5.2. BUCKLING IN GENERAL

It is not a bifurcation point because there is no immediate adjacent equilibrium

configuration. When a limit point load is reached under increasing load, snap-
through buckling occurs, as the structure assumes a new configuration by suddenly
moving. A collapse load is the maximum load a structure can sustain without gross
deformation. It may be greater or less than the computed bifurcation buckling load
as shown in Figure 5.1.
Figure 5.1 shows possible behaviours when loads are applied to a structure. Here
P is the load magnitude and D is the displacement. The material is assumed to
remain linearly elastic and loads are gradually applied.
In Figure 5.1(a) the prebuckling path is linear. At bifurcation, two very closely
spaced equilibrium positions are possible. Thereafter, for P > Pcr , a real (imperfect)
structure follows the secondary path. Since the postbuckling path rises, the structure
can be said to have postbuckling strength. In this case, Pcr characterises a local
buckling action that has little to do with overall strength. The structure will collapse
at the limit point, which is considerably greater than the critical load.
Figure 5.1(b) describes a different type of behaviour. The structure has a nonlinear
prebuckling path and the postbuckling path falls, hence the structure has no post-
buckling strength. Closely spaced primary and falling secondary paths implies that
the structure is imperfection sensitive. Imperfection sensitivity means that small
changes in load directions, geometry and/or boundary conditions strongly affect the
collapse load.
If knowledge of the structural behaviour of the model is little or none, one must
anticipate that a computed bifurcation buckling load may be far above or far be-
low the actual collapse load, imperfections may be influential and that prebuckling
nonlinearities may be important [8].

(a) (b)

Figure 5.1: Possible load versus displacement behaviour of thin-walled structures

(from [8]).

37
CHAPTER 5. CREATING MODELS IN BRIGADE/PLUS

5.3 General Static Analysis

A general analysis step is one in which the effects of any nonlinearities present in the
model can be included. General analysis steps define sequential events: the state of
the model at the end of one general step provides the initial state for the start of
the next general step.
Nonlinear stress analysis problems can contain up to three sources of nonlinearity:
material nonlinearity, geometric nonlinearity and boundary nonlinearity.
It is possible to define a problem as a small-displacement analysis by ignoring the
geometric nonlinearity—the kinematic relationships are linearised. The alternative
to a small-displacement analysis is to include large-displacement effects by including
the NLGEOM parameter on the *STEP option.
In some geometric nonlinear analyses, buckling or collapse may occur. When the
loading can be considered proportional (the loading over the complete structure can
be scaled with a single parameter), a special approach—called the “modified Riks
method”—can be used. In other unstable analyses the instabilities are local (e.g.,
surface wrinkling, material instability or local buckling), in which case global load
control methods such as the Riks method are not appropriate.

5.4 Linear Eigenvalue Buckling Prediction

The response in a linear analysis step is the linear perturbation response about
the base state. The base state is the current state of the model at the end of the
last general analysis step prior to the linear perturbation step. If the first step
of an analysis is a perturbation step, the base state is determined from the initial
conditions.
Linear perturbation analyses can be performed from time to time during a fully
nonlinear analysis by including the linear perturbation steps between the general
response steps. The linear perturbation response has no effect as the general analysis
is continued. If geometric nonlinearity is included in the general analysis upon which
a linear perturbation study is based, stress stiffening or softening effects and load
stiffening effects are included in the linear perturbation analysis.
In an eigenvalue buckling problem the loads for which the model stiffness matrix
becomes singular are searched, so that the problem

KMN ν M = 0

has nontrivial solutions. KMN is the tangent stiffness matrix when the loads are
applied, and the ν M are nontrivial displacement solutions. Eigenvalue buckling is
generally used to estimate the critical buckling loads of stiff structures, e.g. structures
carrying their loads primarily by axial or membrane action. However, even when the
response of a structure is nonlinear prior to collapse, a general eigenvalue buckling

38
 5.4. LINEAR EIGENVALUE BUCKLING PREDICTION

analysis can provide useful estimates of collapse mode shapes. Some structures have
many buckling modes with closely spaced eigenvalues, which can cause numerical
problems. In these cases it often helps to apply enough preload, just below the
buckling load, before performing the eigenvalue extraction. In many cases a series
of closely spaced eigenvalues indicates that the structure is imperfection sensitive.
An eigenvalue buckling analysis will not give accurate predictions of the buckling
load for imperfection sensitive structures; the static Riks procedure should be used
instead. Sometimes negative eigenvalues are reported in an eigenvalue buckling
analysis. In most cases such negative eigenvalues indicate that the structure would
buckle if the load were applied in the opposite direction.
The eigenvalues given in a linear eigenvalue prediction are linear to the applied load
in the step. Hence, the structure will buckle at a load equal to the applied load,
in the linear eigenvalue prediction step, multiplied with the obtained eigenvalue.
For example, if an applied pressure load of 3 kN/m2 gives an eigenvalue of 20, the
structure will buckle at a pressure load of 60 kN/m2 .

5.4.1 The BUCKLE Command

An eigenvalue buckling step is defined by the *BUCKLE command. The buckling

command may be set to predict eigenvalues using two different methods of calcula-
tions, Lanczos or subspace. Generally the Lanczos eigensolver is said to be faster
when a large number of eigenmodes is required from a structure with many degrees
of freedom. The subspace eigensolver can be faster for a smaller system. Both meth-
ods may be used for different steps in the same analysis. The Lanczos solver can
be set to predict eigenvalues in a numeric range, i.e. the minimum and maximum
limits may be prescribed. The subspace iteration does not offer the possibility to
define a minimum limit. The Lanczos eigensolver cannot be used in cases where the
stiffness matrix is indefinite, i.e. if a model is preloaded above the bifurcation load,
if it contains contact pairs and contact elements, if it contains hybrid elements of
connector element, etc. [10].

5.4.2 The IMPERFECTION Command

Imperfections are usually introduced by perturbations in the geometry. ABAQUS

offers three ways to define an imperfection: as a linear superposition of buckling
eigenmodes, from the displacements of a *STATIC analysis, or by specifying the
node number and imperfection values directly. The usual approach involves two
analysis runs with the same model definition:

1. An eigenvalue buckling analysis is performed on the “perfect” structure to

establish probable collapse modes and to verify that the mesh discretises those
modes accurately.

2. In the second analysis run imperfections in the geometry are introduced by

39
CHAPTER 5. CREATING MODELS IN BRIGADE/PLUS

addition of the buckling modes to the “perfect” geometry using the *IMPER-
FECTION option. The imperfection has the form


M
∆xi = ωi φi
i=1

where φi is the ith eigenmode and ωi is the associated scale factor. The mag-
nitudes of the perturbations are typically a few percent of a relative structural
dimension such as a beam cross-section or shell thickness.

3. A geometrically nonlinear load-displacement analysis of the structure contain-

ing the imperfections is performed using the Riks method. In this way the Riks
method can be used to perform postbuckling analyses of “stiff” structures that
show linear behaviour prior to buckling, if perfect.

5.5 Elements

Given a structural geometry, with two dimensions considerably greater than the
third and a relatively complex design, the use of shell elements is a suitable choice.
ABAQUS shell elements assume that plane sections perpendicular to the plane of
the shell remain plane [10].
There are three different classes of shell elements, distinguished by the applicability
of the element to thin and thick shell problems. General-purpose shell elements are
valid for use with both thick and thin shell problems and they also consider finite
membrane strains. The general-purpose elements are: S4, S4R, S3/S3R, SAX1,
SAX2 and SAX2T. The special-purpose shell elements fall into two categories: thin-
only shell elements, including STRI3, STRI65, S4R5, S8R5, S9R5 and SAXA, and
thick-only shell elements, including S8R and S8RT. For thick shells, transverse shear
flexibility is important, while it is negligible for thin shells.
ABAQUS uses numerical integration to calculate the stresses and strains indepen-
dently at each integration point through the thickness of the shell, thus allowing
nonlinear material behaviour. By default, ABAQUS uses five section points through
the thickness of a homogeneous shell.
Throughout the modelling in this thesis the S4R element has been used most fre-
quent. The S4R element is a 4-node, quadrilateral, stress/displacement shell element
with reduced integration.

5.6 Improving Convergence

In nonlinear problems the challenge is always to obtain a convergent solution in

the least possible computational time. ABAQUS offers two approaches for control-
ling incrementation. Direct user control, whereby the user specifies the increment

40
 5.7. ASSUMPTIONS

scheme. Automatic control, whereby the user defines the step and, in some proce-
dures, specifies certain tolerances or error measures. ABAQUS then automatically
select the increment size against the response in the step.
Some static problems can be naturally unstable, for a variety of reasons. Instability
may also be caused by localised buckling behaviour or by material instability. The
static, general analysis procedures in ABAQUS can stabilise this type of problem if
the STABILIZE parameter is included on the procedure option.

5.7 Assumptions

Some simplifying assumptions are consistently made while producing finite element
models. A general attitude towards these are to make as few as possible and to make
them in a manner that they do not affect the correctness of the analysis. Some of
the assumptions made here are described below:

• Abutments
The abutments are assumed to be completely rigid, hence they do not deform
from the loads applied upon them. They have therefore been omitted from the
model and boundary conditions have been applied directly to the box girder
bearings.

• Loads from cast concrete

When casting the midsection of the bridge the, concrete is assumed to be
equally distributed over the whole section, providing an even pressure on the
flanges.

• Camber
The fabricated box girder has a camber (initial elevation) at mid-span. This
is omitted and the model is made level in the horizontal plane.

41
Chapter 6

Profiled Sheeting in Shear

Two models have been created for analysis. One model for a close study of the
effects of the attachment of the profiled sheeting to the top flanges of the box girder
and of the thickness of these profiled sheets. This one will be described in this
chapter, where the study of the influence of how the profiled sheeting is attached is
presented in section 6.1 and the study of the effects of variations in thickness of the
profiled sheets is presented in section 6.2. The other model aims to study the global
behaviour of the bridge, emphasising the effects of initial geometrical imperfections
and it will be described in Chapter 7.

6.1 Profiled Sheet Attachment

This model symbolises a part of the bridge, containing five profiled sheets and parts
of the underlying top flanges of the box girder. Two analyses were performed, one
where the profiled sheeting was attached along the long sides only and one where
it was attached along all sides. Dimensions and material properties of the profiled
sheets and the underlying top flanges are given in Table 6.1. The long and short
plate and the profiled sheet are shown in Figure 6.1. In Figures 6.2 and 6.3, the
models are shown when the profiled sheets are attached along the long sides only
and along all sides, respectively.

Table 6.1: Dimensions and material properties.

Outer Thickness Modulus of Poisson’s
Part dimensions elasticity ratio
(mm2 ) (mm) (GPa) (–)
Long plate 4500 × 300 30
Short plate 300 × 2700 30 210 0.3
Profiled sheet 950 × 2800 0.86

43
CHAPTER 6. PROFILED SHEETING IN SHEAR

Short plate 0,045

Long plate

Long plate
Short plate 2,8
3

0,95
1 2

(a) (b)

Figure 6.1: (a) Long and short plate. (b) Profiled sheet with dimensions.

(b)

(a)
3

1 2

Figure 6.2: The model for analysis of profiled sheet attachment along the long sides
only. Plate (a) is lock in all translational degrees of freedom. Plate (b)
is free to move in its axial direction. Concentrated forces are applied to
plate (b), in two or four nodes.

(d)

(b)

1 2
(c) (a)

Figure 6.3: The model for analysis of profiled sheet attachment along all sides. Plate
(a) is lock in all translational degrees of freedom. Plate (b) is free to
move in its axial direction. Plates (c) and (d) are lock in translation out
of its plane. Concentrated forces are applied to plate (b), in two or four
nodes.

44
 6.2. PROFILED SHEET THICKNESS

In the first analysis the profiled sheets are attached to the top flanges along the long
sides with a tie-interaction. This ties one node in each profile-bottom of the profiled
sheets to the surface of the top flange. In addition to this, a surface-to-surface
contact interaction is included between these parts. This to prevent an unrealistic
movement of the profiled sheets where it slides through the top flange. One of the
top flanges is given a fixed support, locking all of its degrees of freedom. The other
is free to move in its axial direction and a force is applied to it in this direction. The
application of this force gives rise to similar action upon the structure as an incipient
torsion of the box girder. Three different meshes have been analysed providing an
indication of the reliability of the results from the analyses1 .
In the second analysis there are, besides the interactions already defined, additional
tie-interactions along the short sides. Here nodes with c/c 400 mm on the outer
profiled sheets are tied to the additional top flanges to create an all around attach-
ment. There is also an surface-to-surface contact interaction between the short side
top flanges and the part of the profiled sheets that are in contact. Otherwise the
model is created equally to the first one, regarding boundaries, loads and mesh.

6.2 Profiled Sheet Thickness

In this analysis, the same model as in the first analysis was used, i.e. the profiled
sheets are attached along the long sides only (as shown in Figure 6.2), and the
thickness of the sheets was altered. The analysis is performed for three different
thicknesses: 0.774 mm, 0.86 mm and 0.946 mm. The thickness 0.86 mm is the
thickness of the profiled sheet that is used in the bridge and 0.774 mm and 0.946 mm
represent a decrease and an increase of 10%, respectively.

6.3 Results

The evaluation of the results focuses mainly on the stresses that arise in the trans-
verse direction in the profiled sheeting.
Sheets subjected to transverse shear forces will show stress-concentrations at the
edges, as shown in Figure 6.4. If the sheets are attached along all sides, these
stresses will decrease compared to if the sheets are attached only along two sides,
as shown in Figure 6.5.
The results of the study concerning the influence of the profiled sheet attachment
are displayed in Tables 6.2-6.4 and those of the study concerning the influence of the
profiled sheet thickness are shown in Tables 6.6-6.7. Some remarks on the results
and analysis procedure are presented below:
1
If one model with, e.g., 1000 elements and one with 2000 elements are studied, and they give
similar results one can say that the mesh is sufficient accurate. The errors from bad meshing are
small enough.

45
CHAPTER 6. PROFILED SHEETING IN SHEAR

Viewport: 1 ODB: C:/BRIGADEPlusJob/FiveTP-24.odb

S, S22
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+1.532e+08
+1.289e+08
+1.046e+08
+8.024e+07
+5.591e+07
+3.159e+07
+7.261e+06
-1.706e+07
-4.139e+07
-6.571e+07
-9.004e+07
-1.144e+08
-1.387e+08

1 2

Figure 6.4: Transverse stresses for profiled sheet attached at two sides.
S, S22
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+9.889e+06
+8.257e+06
+6.624e+06
+4.991e+06
+3.359e+06
+1.726e+06
+9.321e+04
-1.539e+06
-3.172e+06
-4.805e+06
-6.438e+06
-8.070e+06
-9.703e+06

1 2

Figure 6.5: Transverse stresses for profiled sheet attached along all sides.

• The magnitude of applied forces are not particularly interesting. Instead it is

the differences in stresses, tensions and their distribution between the models
that are of interest.
• The stresses that arise at the outer profile bottoms of the sheets for the case
where the sheets are attached only along the long sides are substantially re-
duced for the case where the sheets are attached along all four sides. This is
shown in Figures 6.4 and 6.5.
• When the profiled sheets are attached along all outer boundaries the stress
distribution is much more favourable in terms of the ability of the structure
to resist applied forces. Hence, this type of attachment creates a more stable
structure.
• For the model with an all around attachment, an average reduction of 94% in
stresses and tensions is attained. Hence, a considerably stronger structure is
provided with an overall, all around attachment of the profiled sheeting than
when it is attached only along two sides.
• If a 10% thinner profiled sheeting is attached, an average increase of about
20% of the stresses and tensions is attained. If a 10% thicker profiled sheeting
is attached, an average decrease of about 15% of the stresses and tensions is
attained.

46
 6.3. RESULTS

Table 6.2: Results from analysis of influence of trapezoidal sheet attachment with a
model composed of 2078/2138 elements.
Applied load
20000 N at two nodes 10000 N at four nodes
Longside Around Longside Around
σMises,max (MPa) 313.7 15.0 312.9 23.2
σ11,max (MPa) 187.9 9.3 187.6 15.4
σ11,min (MPa) −361.5 −17.2 −360.4 −26.7
σ22,max (MPa) 180.0 7.5 177.9 10.5
σ22,min (MPa) −162.4 −7.1 −161.4 −11.4
σ12,max (MPa) 85.7 4.2 85.7 6.3
σ12,min (MPa) −33.5 −6.2 −33.9 −3.3
11,max (·10−6 ) 910.3 45.4 909.7 66.5
11,min (·10−6 ) −1491 −71.8 −1488 −110.9
22,max (·10−6 ) 771.3 27.6 762.3 39.9
22,min (·10−6 ) −652.6 −30.3 −644.9 −44.2
12,max (·10−6 ) 1061 52.5 1061 78.5
12,min (·10−6 ) −414.4 −76.8 −419.5 −40.6
umax (mm) 14.9 0.7 14.9 1.1

Table 6.3: Results from analysis of influence of trapezoidal sheet attachment with a
model composed of 4260/4368 elements.
Applied load
20000 N at two nodes 10000 N at four nodes
Longside Around Longside Around
σMises,max (MPa) 474.0 27.3 473.4 28.5
σ11,max (MPa) 244.6 16.0 244.8 17.8
σ11,min (MPa) −517.7 −29.5 −516.3 −31.0
σ22,max (MPa) 251.0 13.7 248.4 13.8
σ22,min (MPa) −215.4 −10.2 −213.0 −10.8
σ12,max (MPa) 158.2 9.2 158.3 9.6
σ12,min (MPa) −56.2 −6.4 −55.6 −5.9
11,max (·10−6 ) 961.4 62.4 962.6 71.2
11,min (·10−6 ) −2287 −131.4 −2284 −137.6
22,max (·10−6 ) 1069 46.8 1057 47.1
22,min (·10−6 ) −891.2 −42.3 −881.6 −45.3
12,max (·10−6 ) 1958 114.1 1960 118.8
12,min (·10−6 ) −696.1 −79.2 −688.8 −72.9
umax (mm) 15.9 0.9 15.9 1.0

47
CHAPTER 6. PROFILED SHEETING IN SHEAR

Table 6.4: Results from analysis of influence of trapezoidal sheet attachment with a
model composed of 10320/10752 elements.
Applied load
20000 N at two nodes 10000 N at four nodes
Longside Around Longside Around
σMises,max (MPa) 634.0 29.4 635.1 26.9
σ11,max (MPa) 270.4 13.4 270.0 11.7
σ11,min (MPa) −661.7 −30.2 −660.3 −27.9
σ22,max (MPa) 308.9 10.7 311.8 9.5
σ22,min (MPa) −256.2 −8.2 −253.6 −7.8
σ12,max (MPa) 212.9 10.0 213.2 10.2
σ12,min (MPa) −95.5 −11.0 −95.3 −6.0
11,max (·10−6 ) 1152 57.9 1150 50.4
11,min (·10−6 ) −3003 −138.4 −3003 −126.7
22,max (·10−6 ) 1316 36.6 1329 37.0
22,min (·10−6 ) −1061 −36.5 −1050 −33.3
12,max (·10−6 ) 2636 123.7 2640 126.0
12,min (·10−6 ) −1183 −136.6 −1180 −74.8
umax (mm) 16.2 0.7 16.2 0.7

Table 6.5: Results from analysis of influence of profiled sheet thickness with a model
having a two-sided attachment of the profiled sheeting composed of 2078
elements.
Sheet thickness (mm)
0.86 0.774 0.946
σMises,max (MPa) 312.9 377.0 263.8
σ11,max (MPa) 187.6 218.2 168.6
σ11,min (MPa) −360.4 −434.1 −304.0
σ22,max (MPa) 177.9 208.9 154.3
σ22,min (MPa) −161.4 −191.9 −138.3
σ12,max (MPa) 85.7 100.1 74.4
σ12,min (MPa) −33.9 −41.7 −28.2
11,max (·10−6 ) 909.7 1031 810.9
11,min (·10−6 ) −1488 −1795 −1253
22,max (·10−6 ) 762.3 891.9 665.5
22,min (·10−6 ) −644.9 −743.0 −568.4
12,max (·10−6 ) 1061 1239 920.7
12,min (·10−6 ) −419.5 −516.1 −349.0
umax (mm) 14.9 19.7 11.6

48
 6.3. RESULTS

Table 6.6: Results from analysis of influence of profiled sheet thickness with a model
having a two-sided attachment of the profiled sheeting composed of 4260
elements.
Sheet thickness (mm)
0.86 0.774 0.946
σMises,max (MPa) 473.4 567.9 400.8
σ11,max (MPa) 244.8 293.1 208.0
σ11,min (MPa) −516.3 −618.8 −437.5
σ22,max (MPa) 248.4 292.4 214.8
σ22,min (MPa) −213.0 −247.9 −186.1
σ12,max (MPa) 158.3 186.6 135.9
σ12,min (MPa) −55.6 −67.5 −46.9
11,max (·10−6 ) 962.6 1141 828.4
11,min (·10−6 ) −2284 −2736 −1935
22,max (·10−6 ) 1057 1237 919.3
22,min (·10−6 ) −881.6 −1017 −776.4
12,max (·10−6 ) 1960 2313 1683
12,min (·10−6 ) −688.8 −835.8 −580.7
umax (mm) 15.9 21.1 12.3

Table 6.7: Results from analysis of influence of profiled sheet thickness with a model
having a two-sided attachment of the profiled sheeting composed of 10320
elements.
Sheet thickness (mm)
0.86 0.774 0.946
σMises,max (MPa) 635.1 760.1 538.6
σ11,max (MPa) 270.0 326.3 227.5
σ11,min (MPa) −660.3 −78.8 −561.9
σ22,max (MPa) 311.8 366.8 270.0
σ22,min (MPa) −253.6 −293.4 −222.2
σ12,max (MPa) 213.2 251.4 183.2
σ12,min (MPa) −95.3 −114.2 −80.8
11,max (·10−6 ) 1150 1392 973.5
11,min (·10−6 ) −3003 −3585 −2553
22,max (·10−6 ) 1329 1559 1153
22,min (·10−6 ) −1050 −1209 −924.3
12,max (·10−6 ) 2640 3113 2268
12,min (·10−6 ) −1180 −1414 −1001
umax (mm) 16.2 21.5 12.5

49
Chapter 7

Lateral Torsional Stability

This model is used to study the lateral torsional stability of the bridge. The influence
of the profiled sheeting and its attachment, and also the influence of geometric
imperfections in the box girder, mainly in shape of an initial rotation, are considered.
Three different models are studied in this analysis:

• a model composed of the steel box girder only,

• a model composed of the steel box girder and all the profiled sheets, and

• a model composed of the steel box girder and all the profiled sheets, except
the ones closest to the abutments. This last model is made to study what
importance it has to make an all around attachment of the profiled sheets to
the top flanges of the box girder. The bridge that collapsed (Y1504) had the
profiled sheets attached as in this model.

The box girder is made as one homogeneous part of shell elements, with the external
dimensions 57 × 3.9 × 2.0 m3 , as shown in Figure 7.1.

57

2,0

3,9

Figure 7.1: Model of steel box girder with external dimensions.

51
CHAPTER 7. LATERAL TORSIONAL STABILITY

0,045 0,045

3,4 2,8

0,95 0,95

(a) (b)

Figure 7.2: Model and dimensions of the long profiled sheet (a) and the short profiled
sheet (b).

The profiled sheets are made as a homogeneous part of shell elements, with the ex-
ternal dimensions 0.95 × 3.4 × 0.045 m3 or 0.95 × 2.8 × 0.045 m3 , as shown in Fig-
ure 7.2. There are 11 long profiled sheets, attached from each abutment and towards
the centre of the bridge, and 38 short profiled sheets attached at the centre part of
the bridge.
Dimensions and material properties for the parts are given in Table 7.1. The profiled
sheets are connected to the box girder in each profile bottom with a tie-interaction,
locking all translational degrees of freedom of the node in the profile bottom to the
top flange of the box girder. They are also tied mutually in the same manner as for
the small model. The connections for the profiled sheets are shown in Figure 7.3.
The boundary conditions are defined equally for all three models. The bearings at
one end of the bridge are locked in all translational degrees of freedom while the
bearings at the other end are locked only in the axial and transverse direction. Apart
from the formwork and reinforcement, which are modelled as loads, the self-weight,
wind load and the load from the cast concrete are applied.

(2)

(3) (1)

Figure 7.3: Profiled sheet with connections: (1) are the tie interactions between the
profiled sheet and the box girder top flange. (2) are the slave nodes
in the tie interaction between the profiled sheets, and (3) represent the
master surface in the same interaction. The slave nodes in (2) are tied
to the master surface of the next profiled sheet.

52
 7.1. IMPLEMENTATION OF GEOMETRIC IMPERFECTIONS

Table 7.1: Dimensions and material properties.

Outer Thickness Modulus of Poisson’s
Part dimensions elasticity ratio
(m) (mm) (GPa) (–)
Box girder 57 × 3.9 × 2 10 – 38
Central profiled sheets 0.95 × 2.7 × 0.045 0.86 210 0.3
Outer profiled sheets 0.95 × 3.3 × 0.045 0.86

7.1 Implementation of Geometric Imperfections

The implementation of geometric imperfections is performed by first performing a

linear eigenvalue prediction, which provides the buckling modes for the structure.
These buckling modes are then added to the model as imperfections in the geom-
etry, giving the “perfect” structure displacements that correspond to the applied
eigenmodes.
In the linear eigenvalue prediction the first overall buckling mode is the most vital
to find, since this buckling mode corresponds well to the form that the box girder
has due to production imperfections, hence the way the bridge will deform in case of
collapse. However, this particular buckling mode is not trivial to find in the model
containing both the box girder and the profiled sheets because a large amount of
local buckling modes, mainly in the profiled sheets, will appear prior to any overall
buckling mode. In order to avoid a linear eigenvalue prediction that only provides
numerous local buckling modes, a linear eigenvalue prediction can be performed on
a model composed of the steel box girder only. In this way, an overall buckling mode
is obtained much earlier in the analysis. The buckling modes of interest can then be
implemented to the model composed of both the steel box girder and the profiled
sheets, given that the node numbers are the same in both the linear eigenvalue
prediction analysis and the other model.
The buckling modes of interest, i.e. mainly the first overall buckling mode but also
random local buckling modes, are introduced to the structure with a suitable scaling
factor providing the geometric imperfections to the structure. The scaling factor is
varied in different analyses, providing an indication of the influence of the magnitude
of the geometric imperfections.

7.1.1 Linear Eigenvalue Prediction

The linear eigenvalue prediction is performed using the Lanczos eigensolver. The
first 100 buckling modes larger than zero are sought. Three concentrated forces at
each top flange, as shown in Figure 7.4, are applied to create the perturbation.
The first overall buckling mode given by this analysis is shown in Figure 7.5. This
mode, and some random local buckling modes like those in Figure 7.6, are introduced
to the models with different scaling factors for the postbuckling analysis.

53
CHAPTER 7. LATERAL TORSIONAL STABILITY

Figure 7.4: Loads in the linear eigenvalue prediction analysis.

U, Magnitude
+1.196e+00
+1.096e+00
+9.963e-01
+8.967e-01
+7.971e-01
+6.974e-01
+5.978e-01
+4.982e-01
+3.985e-01
+2.989e-01
+1.993e-01
+9.963e-02
+0.000e+00

Figure 7.5: First overall buckling mode from the linear eigenvalue prediction.

U, Magnitude U, Magnitude
+1.022e+00 +1.024e+00
+9.372e-01 +9.388e-01
+8.520e-01 +8.534e-01
+7.668e-01 +7.681e-01
+6.816e-01 +6.827e-01
+5.964e-01 +5.974e-01
+5.112e-01 +5.121e-01
+4.260e-01 +4.267e-01
+3.408e-01 +3.414e-01
+2.556e-01 +2.560e-01
+1.704e-01 +1.707e-01
+8.520e-02 +8.534e-02
+0.000e+00 +0.000e+00

(a) (b)

Figure 7.6: Typical local buckling modes.

54
 7.2. POSTBUCKLING ANALYSIS

7.2 Postbuckling Analysis

The postbuckling analysis aims to analyse the bridge response, regarding stresses
and deformations, to the applied loads with different geometrical imperfections and
different attachments of the profiled sheeting. Analyses are carried out on the three
different models as described below:

• The first overall buckling mode is introduced with a scaling factor w = 0.05.
Additionally, a few random local buckling modes are introduced with a scaling
factor of w = 0.001. The loads are applied as:

– A pressure load of 5 kN/m2 at the top flanges representing the wooden

formwork.
– A pressure load of 1.5 kN/m2 , acting transversally on the structure, rep-
resenting the wind load.
– An additional pressure load of 45 kN/m2 at the central part of the top
flanges representing the cast concrete.

The model is meshed with an approximate global element size of 0.48 meter
and the S4R element is chosen. The loads from the wooden formwork and wind
load are applied in a non-linear static general step, while the cast concrete is
applied in a Riks step.

• Additional analyses are performed with the following changes:

– The overall buckling mode is introduced with a scaling factor of 0.01, 0.1
and 0.2.
– The model is meshed with an approximate global element size of 0.24 me-
ter.

55
CHAPTER 7. LATERAL TORSIONAL STABILITY

7.3 Results

First, the results from the analyses are presented individually for the models and
then a comparison between these models in given.

7.3.1 Box Girder with Open Cross-Section

This model aims to demonstrate the sensitivity of a box girder with open cross-
section, principally regarding its lateral torsional stability. The stresses and dis-
placements from
the analyses are presented and compared. The displacements and stresses for the
whole model are presented. Some paths, as shown in Figure 7.7, are used for a study
of displacements and stresses at different stages of the analyses.

Analysis 1

This analysis is performed with a scaling factor of 0.05 for the overall buckling mode.
Three local buckling modes are introduced with a scaling factor of 0.001 and the
model is meshed with global seeds of 0.48.
In Figure 7.8 the von Mises stresses for the whole model are shown, in Figures 7.9-
7.11 the displacements, in Figure 7.12 the stresses in the 11-direction, in Figure 7.13
the stresses in the 22-direction and
in Figure 7.14 the stresses in the 12-direction. The figures show the stresses and
displacements when full load from the cast concrete is applied. The stresses are
displayed in Pa and the displacements in metres.

Path: Top flange 1

Path: Top flange 2

Figure 7.7: Paths for which displacements and stresses are shown.

56
 7.3. RESULTS

S, Mises
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+3.596e+08
+3.297e+08
+2.998e+08
+2.699e+08
+2.399e+08
+2.100e+08
+1.801e+08
+1.502e+08
+1.203e+08
+9.035e+07
+6.044e+07
+3.052e+07
+6.047e+05

1 2

Figure 7.8: von Mises stresses.

U, Magnitude
+3.794e-01
+3.478e-01
+3.162e-01
+2.846e-01
+2.529e-01
+2.213e-01
+1.897e-01
+1.581e-01
+1.265e-01
+9.485e-02
+6.323e-02
+3.162e-02
+0.000e+00

1 2

√
Figure 7.9: Magnitude of displacements, U = U12 + U22 + U32 .

57
CHAPTER 7. LATERAL TORSIONAL STABILITY

U, U2
+1.087e-01
+9.955e-02
+9.041e-02
+8.128e-02
+7.214e-02
+6.300e-02
+5.386e-02
+4.472e-02
+3.558e-02
+2.644e-02
+1.730e-02
+8.164e-03
-9.751e-04

1 2

Figure 7.10: Lateral displacements, U2.

U, U3
+1.599e-03
-2.897e-02
-5.955e-02
-9.012e-02
-1.207e-01
-1.513e-01
-1.818e-01
-2.124e-01
-2.430e-01
-2.736e-01
-3.041e-01
-3.347e-01
-3.653e-01

1 2

Figure 7.11: Vertical displacements, U3.

58
 7.3. RESULTS

S, S11
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+2.282e+08
+1.784e+08
+1.286e+08
+7.878e+07
+2.898e+07
-2.082e+07
-7.062e+07
-1.204e+08
-1.702e+08
-2.200e+08
-2.698e+08
-3.196e+08
-3.694e+08

1 2

Figure 7.12: Stresses in the 11-direction.

S, S22
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+1.139e+08
+8.003e+07
+4.619e+07
+1.235e+07
-2.149e+07
-5.533e+07
-8.916e+07
-1.230e+08
-1.568e+08
-1.907e+08
-2.245e+08
-2.584e+08
-2.922e+08

1 2

Figure 7.13: Stresses in the 22-direction.

59
CHAPTER 7. LATERAL TORSIONAL STABILITY

S, S12
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+1.324e+08
+1.149e+08
+9.741e+07
+7.990e+07
+6.239e+07
+4.487e+07
+2.736e+07
+9.846e+06
-7.667e+06
-2.518e+07
-4.269e+07
-6.021e+07
-7.772e+07

1 2

Figure 7.14: Stresses in the 12-direction.

The first path, top flange 1, is created in the bridge longitudinal direction at one of
the top flanges. In Figure 7.15 the magnitudes of the displacements along path top
flange 1 are shown. Graphs are drawn for different stages of the casting, starting
when no concrete is cast and until the casting is finished. In Figure 7.16 the lateral
displacements along path top flange 1 are shown.

Cast concrete: 0%
Cast concrete: 20%
Cast concrete: 40%
Cast concrete: 60%
Cast concrete: 80%
Cast concrete: 100%

Figure 7.15: Magnitude of node displacement along path top flange 1.

60
 7.3. RESULTS

Cast concrete: 0%
Cast concrete: 20%
Cast concrete: 40%
Cast concrete: 60%
Cast concrete: 80%
Cast concrete: 100%

Figure 7.16: Lateral node displacement along path top flange 1.

The other top flange is also used to create a path, top flange 2, for which the same
variables are studied. In Figure 7.17 the magnitude of the displacements along path
top flange 2 are shown. In Figure 7.18 the lateral displacements along path top
flange 2 are shown.

Cast concrete: 0%
Cast concrete: 20%
Cast concrete: 40%
Cast concrete: 60%
Cast concrete: 80%
Cast concrete: 100%

Figure 7.17: Magnitude of node displacement along path top flange 2.

61
CHAPTER 7. LATERAL TORSIONAL STABILITY

Cast concrete: 0%
Cast concrete: 20%
Cast concrete: 40%
Cast concrete: 60%
Cast concrete: 80%
Cast concrete: 100%

Figure 7.18: Lateral node displacement along path top flange 2.

There is an obvious difference between the magnitude of node displacement for top
flange 1 and top flange 2, while their lateral displacements are almost identical. This
is due to the rotation about the longitudinal axis, which makes top flange 1 descend
twice as much as top flange 2.
To calculate the ultimate load for the construction, the analysis is proceeded further.
Instead of stopping at a stage where the full load from the cast concrete is applied,
the analysis is carried on until the ultimate load is reached. The Riks method
initially adds 0.1 of the total casting load in each increment. When the applied load
approaches the ultimate load, the applied load in each increment is reduced so that
there still is a converging solution. As the applied load approaches the ultimate load,
the displacements continue to grow and the curve shown in Figure 7.20 is obtained.

62
 7.3. RESULTS

U, U2
+6.472e-01
+5.912e-01
+5.351e-01
+4.790e-01
+4.229e-01
+3.668e-01
+3.108e-01
+2.547e-01
+1.986e-01
+1.425e-01
+8.644e-02
+3.036e-02
-2.572e-02

1 2

Figure 7.19: Lateral displacement at ultimate load.

In Figure 7.19 the lateral displacements when the ultimate load is applied are shown.
In Figure 7.20 the maximum lateral displacement along path top flange 1 is shown
on the x-axis, while the y-axis shows a load proportionality factor, λ, of the cast
concrete.
2.5

Pcr
2

1.5
λ

0.5

0
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
δ

Figure 7.20: Load-displacement curve.

63
CHAPTER 7. LATERAL TORSIONAL STABILITY

Plotting these displacements against a quota of the displacement divided by the

applied load, Southwell plot is given. This curve is shown in Figure 7.21. From this
curve, the critical load proportionality factor is given by
1
λcr =
θ
where θ is the gradient of the curve. Using the marked points in Figure 7.21 to
calculate the gradient, θ = 0.432 which gives λcr = 2.31. This value can also be
compared with the maximum value of the critical load proportionality factor given
directly by BRIGADE/Plus, which is λcr = 2.16 as shown in Figure 7.20.

0.35

0.3

0.25
δ/P

0.2
θ

0.15

0.1
0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55
δ

Figure 7.21: Southwell plot.

Additional analyses

Five additional analyses are performed on this model as described below:

• Analysis 2. This analysis is performed with a scaling factor of 0.01 for the
overall buckling mode. Three local buckling modes are introduced with a
scaling factor of 0.001 and the model is meshed with a global seed of 0.48.

• Analysis 3. This analysis is performed with a scaling factor of 0.1 for the
overall buckling mode. Three local buckling modes are introduced with a
scaling factor of 0.001 and the model is meshed with a global seed of 0.48.

64
 7.3. RESULTS

• Analysis 4. This analysis is performed with a scaling factor of 0.2 for the
overall buckling mode. Three local buckling modes are introduced with a
scaling factor of 0.001 and the model is meshed with a global seed of 0.48.
• Analysis 5. This analysis is performed without any geometrical imperfections.
The model is meshed with a global seed of 0.48.
• Analysis 6. This analysis is performed with a scaling factor of 0.05 for the
overall buckling mode. Three local buckling modes are introduced with a
scaling factor of 0.001 and the model is meshed with a global seed of 0.24.

The results from these analyses are compared with the results from the first analysis.
In Figure 7.22 the magnitude of the displacements along path top flange 1 are shown.
Graphs are drawn for when the total load from the casting is applied for different
initial imperfections. In Figure 7.23 the lateral displacements along path top flange
1 are shown.

Overall imperfection 0
Overall imperfection 0.01
Overall imperfection 0.05
Overall imperfection 0.10
Overall imperfection 0.20

Figure 7.22: Magnitude of node displacement along path top flange 1 for different
initial imperfections.

Overall imperfection 0
Overall imperfection 0.01
Overall imperfection 0.05
Overall imperfection 0.10
Overall imperfection 0.20

Figure 7.23: Lateral node displacement along path top flange 1 for different initial
imperfections

65
CHAPTER 7. LATERAL TORSIONAL STABILITY

The maximal lateral displacements in Figure 7.23 becomes relatively larger when the
initial imperfection is increased. Increasing the initial imperfection from 0.01 to 0.05
provides a growth in lateral displacement of 36%. A further increase from 0.05 to 0.1
yields an additional growth in lateral displacement of 24% and an increase from 0.1
to 0.2 finally produces a growth in lateral displacement of 52%. This clearly shows
that an increased initial imperfection drastically increases the lateral displacement.
In Figure 7.24 the maximum lateral displacement along path top flange 1 is shown
on the x-axis, while the y-axis shows a load proportionality factor, λ, of the cast
concrete. Curves are drawn for different initial imperfections. The Southwell plot
for the curves of Figure 7.24 are shown in Figure 7.25.

2.5

1.5
λ

0.5
Initial imperfection 0
Initial imperfection 0,01
Initial imperfection 0,05
Initial imperfection 0,1
Initial imperfection 0,2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
δ

Figure 7.24: Load-displacement curves for different initial imperfections.

66
 7.3. RESULTS

0.55
Initial imperfection 0
Points to calculate θ0
0.5 Initial imperfection 0,01
Points to calculate θ0,01
Initial imperfection 0,05
0.45 Points to calculate θ0,05
Initial imperfection 0,1
Points to calculate θ0,1
0.4 Initial imperfection 0,2
Points to calculate θ0,2

0.35
δ/P

0.3

0.25

0.2

0.15

0.1

0.05
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
δ

Figure 7.25: Southwell plot for various initial imperfections.

From these curves, the critical load proportionality factor is given by

1
λcr,i =
θi
where θi is the gradient of the curve i. Using the marked points in Figure 7.25
to calculate the gradient, θi , and the corresponding critical load proportionality
factor, λi , gives the results presented in Table 7.2. In Figure 7.26, the axial rotation
for the bridge is shown for different initial imperfections. These curves can be
approximated with quadratic curves, producing the curves shown in Figure 7.27.
From these curves, it is clear that an increased initial imperfection also produces a
relatively larger rotation of the structure.

Table 7.2: Curve gradients and their corresponding critical load proportionality fac-
tors for various initial imperfections.
Initial θi λi
imperfection, i (rad) (–)
0 0.409 2.446
0.01 0.412 2.427
0.05 0.432 2.314
0.1 0.464 2.153
0.2 0.503 1.988

67
CHAPTER 7. LATERAL TORSIONAL STABILITY

0.07
Initial imperfection 0
Initial imperfection 0,01
Initial imperfection 0,05
0.06 Initial imperfection 0,1
Initial imperfection 0,2

0.05

0.04
θ (rad)

0.03

0.02

0.01

0
0 10 20 30 40 50 60
x (m)

Figure 7.26: Axial rotation for various initial imperfections.

0.07
Initial imperfection 0
Initial imperfection 0,01
Initial imperfection 0,05
0.06 Initial imperfection 0,1
Initial imperfection 0,2

0.05

0.04
θ (rad)

0.03

0.02

0.01

0
0 10 20 30 40 50 60
x (m)

Figure 7.27: Axial rotation, approximated with quadratic curves, for various initial
imperfections.

68
 7.3. RESULTS

To verify that the error from bad meshing is small enough to be neglected, an
analysis with an improved mesh is performed. The box girder is meshed with global
seeds of 0.24. In Figure 7.28, the magnitude of the displacements and the von Mises
stresses from analyses with different meshes are shown. Since the discrepancy is
very small, the coarser mesh can be assumed to provide trustworthy results.

Global seeds 0.48

Global seeds 0.48
Global seeds 0.24
Global seeds 0.24

(a) (b)

Figure 7.28: (a) Magnitude of displacements along top flange 1. (b) von Mises
stresses along top flange1.

69
CHAPTER 7. LATERAL TORSIONAL STABILITY

7.3.2 Box Girder with Closed Cross-Section

The same box girder as in the model with the open cross-section is used in this model.
Profiled sheets are attached at the top of the box girder and given an all around
attachment, producing a closed cross-section. The stresses and displacements from
the analyses are presented and compared.

Analysis 1

This analysis is performed with a scaling factor of 0.05 for the overall buckling mode.
Three local buckling modes are introduced with a scaling factor of 0.001 and the
model is meshed with global seeds of 0.48.
In Figure 7.29 the von Mises stresses for the whole model are shown, in Figures 7.30-
7.32 the displacements, in Figure 7.33 the stresses in the 11-direction, in Figure 7.34
the stresses in the 22-direction and in Figure 7.35 the stresses in the 12-direction.

S, Mises
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+4.753e+08
+4.357e+08
+3.961e+08
+3.565e+08
+3.169e+08
+2.773e+08
+2.377e+08
+1.981e+08
+1.585e+08
+1.189e+08
+7.927e+07
+3.966e+07
+6.184e+04

1 2

Figure 7.29: von Mises stresses.

70
 7.3. RESULTS

U, Magnitude
+2.987e-01
+2.738e-01
+2.489e-01
+2.240e-01
+1.991e-01
+1.742e-01
+1.493e-01
+1.244e-01
+9.956e-02
+7.467e-02
+4.978e-02
+2.489e-02
+0.000e+00

1 2

√
Figure 7.30: Magnitude of displacements, U = U12 + U22 + U32 .

U, U2
+1.767e-02
+1.605e-02
+1.443e-02
+1.281e-02
+1.119e-02
+9.564e-03
+7.942e-03
+6.320e-03
+4.698e-03
+3.075e-03
+1.453e-03
-1.687e-04
-1.791e-03

1 2

Figure 7.31: Lateral displacements, U2.

71
CHAPTER 7. LATERAL TORSIONAL STABILITY

U, U3
+1.363e-03
-2.362e-02
-4.860e-02
-7.359e-02
-9.857e-02
-1.236e-01
-1.485e-01
-1.735e-01
-1.985e-01
-2.235e-01
-2.485e-01
-2.734e-01
-2.984e-01

1 2

Figure 7.32: Vertical displacements, U3.

S, S11
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+3.328e+08
+2.624e+08
+1.920e+08
+1.216e+08
+5.121e+07
-1.919e+07
-8.959e+07
-1.600e+08
-2.304e+08
-3.008e+08
-3.712e+08
-4.416e+08
-5.120e+08

1 2

Figure 7.33: Stresses in the 11-direction.

72
 7.3. RESULTS

S, S22
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+2.571e+08
+2.096e+08
+1.621e+08
+1.146e+08
+6.716e+07
+1.968e+07
-2.780e+07
-7.529e+07
-1.228e+08
-1.703e+08
-2.177e+08
-2.652e+08
-3.127e+08

1 2

Figure 7.34: Stresses in the 22-direction.

S, S12
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+1.300e+08
+1.089e+08
+8.775e+07
+6.665e+07
+4.555e+07
+2.444e+07
+3.340e+06
-1.776e+07
-3.887e+07
-5.997e+07
-8.107e+07
-1.022e+08
-1.233e+08

1 2

Figure 7.35: Stresses in the 12-direction.

The profiled sheeting is subjected to large stresses in the connections to the box
girder, as described in Chapter 6. If these stresses become to large, bolt hole yielding
will occur. The stresses at the outer part of the profiled sheeting are shown below.

73
CHAPTER 7. LATERAL TORSIONAL STABILITY

In Figure 7.36 the von Mises stresses for the four outer profiled sheets are shown.
In Figure 7.37 the stresses in the axial direction for the four outer profiled sheets
are shown. In Figure 7.38 the stresses in the transverse direction for the four outer
profiled sheets are shown.

S, Mises
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+3.545e+08
+3.251e+08
+2.956e+08
+2.662e+08
+2.367e+08
+2.072e+08
+1.778e+08
+1.483e+08
+1.189e+08
+8.941e+07
+5.995e+07
+3.049e+07
+1.036e+06

2
1

Figure 7.36: von Mises stresses.

S, S11
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+2.531e+08
+1.999e+08
+1.466e+08
+9.337e+07
+4.012e+07
-1.313e+07
-6.639e+07
-1.196e+08
-1.729e+08
-2.261e+08
-2.794e+08
-3.327e+08
-3.859e+08

2
1

Figure 7.37: Stresses in the 11-direction.

S, S22
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+1.351e+08
+1.146e+08
+9.421e+07
+7.378e+07
+5.335e+07
+3.292e+07
+1.249e+07
-7.937e+06
-2.837e+07
-4.880e+07
-6.922e+07
-8.965e+07
-1.101e+08

2
1

Figure 7.38: Stresses in the 22-direction.

74
 7.3. RESULTS

7.3.3 Box Girder with Partly Closed Cross-Section

The same box girder as in the model with the open cross-section is used in this
model. Profiled sheets are attached at the top flanges of the box girder, providing
a partly closed cross-section. The stresses and displacements from the analyses are
presented and compared.

Analysis 1

This analysis is performed with a scaling factor of 0.05 for the overall buckling mode.
Three local buckling modes are introduced with a scaling factor of 0.001 and the
model is meshed with global seeds of 0.48.
In Figure 7.39 the von Mises stresses for the whole model are shown, in Figures 7.40-
7.42 the displacements, in Figure 7.43 the stresses in the 11-direction, in Figure 7.44
the stresses in the 22-direction and in Figure 7.45 the stresses in the 12-direction.

S, Mises
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+4.829e+08
+4.426e+08
+4.024e+08
+3.622e+08
+3.220e+08
+2.818e+08
+2.416e+08
+2.013e+08
+1.611e+08
+1.209e+08
+8.069e+07
+4.047e+07
+2.518e+05

1 2

Figure 7.39: von Mises stresses.

75
CHAPTER 7. LATERAL TORSIONAL STABILITY

U, Magnitude
+2.991e-01
+2.742e-01
+2.493e-01
+2.244e-01
+1.994e-01
+1.745e-01
+1.496e-01
+1.246e-01
+9.971e-02
+7.478e-02
+4.986e-02
+2.493e-02
+0.000e+00

1 2

√
Figure 7.40: Magnitude of displacements, U = U12 + U22 + U32 .

U, U2
+1.814e-02
+1.652e-02
+1.489e-02
+1.326e-02
+1.163e-02
+1.001e-02
+8.378e-03
+6.751e-03
+5.123e-03
+3.496e-03
+1.868e-03
+2.408e-04
-1.387e-03

1 2

Figure 7.41: Lateral displacements, U2.

76
 7.3. RESULTS

U, U3
+1.391e-03
-2.363e-02
-4.865e-02
-7.367e-02
-9.870e-02
-1.237e-01
-1.487e-01
-1.738e-01
-1.988e-01
-2.238e-01
-2.488e-01
-2.738e-01
-2.989e-01

1 2

Figure 7.42: Vertical displacements, U3.

S, S11
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+3.115e+08
+2.413e+08
+1.711e+08
+1.008e+08
+3.062e+07
-3.960e+07
-1.098e+08
-1.800e+08
-2.503e+08
-3.205e+08
-3.907e+08
-4.609e+08
-5.312e+08

1 2

Figure 7.43: Stresses in the 11-direction.

77
CHAPTER 7. LATERAL TORSIONAL STABILITY

S, S22
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+2.591e+08
+2.109e+08
+1.626e+08
+1.144e+08
+6.618e+07
+1.796e+07
-3.027e+07
-7.849e+07
-1.267e+08
-1.749e+08
-2.232e+08
-2.714e+08
-3.196e+08

1 2

Figure 7.44: Stresses in the 22-direction.

S, S12
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+1.317e+08
+1.102e+08
+8.879e+07
+6.733e+07
+4.587e+07
+2.441e+07
+2.946e+06
-1.851e+07
-3.997e+07
-6.143e+07
-8.290e+07
-1.044e+08
-1.258e+08

1 2

Figure 7.45: Stresses in the 12-direction.

78
 7.3. RESULTS

The stresses at the outer part of the profiled sheeting are shown below. In Figure 7.46
the von Mises stresses for the three outer profiled sheets are shown. In Figure 7.47
the stresses in the axial direction for the three outer profiled sheets are shown. In
Figure 7.48 the stresses in the transverse direction for the three outer profiled sheets
are shown.

S, Mises
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+3.693e+08
+3.387e+08
+3.081e+08
+2.776e+08
+2.470e+08
+2.164e+08
+1.858e+08
+1.552e+08
+1.246e+08
+9.404e+07
+6.345e+07
+3.286e+07
+2.278e+06

2
1

Figure 7.46: von Mises stresses.

S, S11
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+3.115e+08
+2.525e+08
+1.935e+08
+1.344e+08
+7.540e+07
+1.637e+07
-4.266e+07
-1.017e+08
-1.607e+08
-2.197e+08
-2.788e+08
-3.378e+08
-3.968e+08

2
1

Figure 7.47: Stresses in the 11-direction.

S, S22
SNEG, (fraction = -1.0)
(Ave. Crit.: 75%)
+2.407e+08
+2.021e+08
+1.636e+08
+1.251e+08
+8.660e+07
+4.808e+07
+9.568e+06
-2.895e+07
-6.746e+07
-1.060e+08
-1.445e+08
-1.830e+08
-2.215e+08

2
1

Figure 7.48: Stresses in the 22-direction.

79
CHAPTER 7. LATERAL TORSIONAL STABILITY

7.3.4 Comparison of Results

From the model with a quadrilaterally connected profiled sheeting and the model
with a laterally connected profiled sheeting, two main conclusions are drawn. The
first is that if the profiled sheeting is only laterally attached, large axial forces
appear in parts of the profiled sheeting closest to the edge, additionally straining
the connections. This is demonstrated in Figure 7.49.
The other conclusion is that in both these models the stresses in the profile-bottoms
of the profiled sheeting become alarmingly large, already at a relatively moderate
loading. These stresses are larger for the model with a two sided attachment of the
profiled sheeting than for the one with a quadrilaterally attached profiled sheeting.
In Table 7.3 the stresses in the profiled sheeting are compared between the model
with the closed cross-section and the partly closed cross-section.

Table 7.3: Stresses in the profiled sheeting with an initial imperfection of 0.05 when
full load from cast concrete is applied.
Stresses (MPa) Closed cross-section Partly closed cross-section
All Sheets Outer Sheets All Sheets Outer Sheets
von Mises 475 355 483 369
11-direction (+) 333 253 312 312
11-direction (−) 512 386 531 397
22-direction (+) 257 135 259 241
22-direction (−) 313 110 320 222

These values should be interpreted with some caution. The connections in the
model have been made in such a manner that the part of the profile-bottom that is
in contact with the top flange has been tied to this top flange. In the actual bridge,
the profiled sheets are nailed to the top flange with one nail at each profile-bottom,
providing a much more concentrated connection. Consequently, the stresses that

S, S22 S, S22
SNEG, (fraction = -1.0) SNEG, (fraction = -1.0)
(Ave. Crit.: 75%) (Ave. Crit.: 75%)
+1.351e+08 +2.407e+08
+1.146e+08 +2.021e+08
+9.421e+07 +1.636e+08
+7.378e+07 +1.251e+08
+5.335e+07 +8.660e+07
+3.292e+07 +4.808e+07
+1.249e+07 +9.568e+06
-7.937e+06 -2.895e+07
-2.837e+07 -6.746e+07
-4.880e+07 -1.060e+08
-6.922e+07 -1.445e+08
-8.965e+07 -1.830e+08
-1.101e+08 -2.215e+08

3
3

2
1 2
1

(a) (b)

Figure 7.49: Stresses in the 22-direction for the outer profiled sheets for the model
with in (a) a quadrilaterally connected profiled sheeting and in (b) a
laterally connected profiled sheeting.

80
 7.3. RESULTS

arise in the real bridge are probably somewhat larger than the ones given by the
finite element models, thus the finite element model provides a lower bound for the
stresses.
A comparison of the displacements and rotations between the three different mod-
els (open cross-section, closed cross-section and partly closed cross-section) shows
that for the models with the profiled sheeting, the displacements and rotations are
very small, while they are quite large for the open box girder. In Figure 7.50 the
lateral displacements along top flange 1 for these three different models, with an
initial imperfection of 0.05, are shown. The model with the open cross-section has
a maximum lateral displacement of 0.109 m, the one with the closed cross-section
0.0177 m and the one with the partly closed cross-section 0.0181 m.
In Figure 7.51, the rotations for all three models are shown. The model with the
open cross-section experiences rotations that are about ten times the rotations for
the models with the profiled sheeting.
When the stresses in the profile-bottoms become to large, bolt hole yielding will
occur which swiftly reduces the shearing rigidity and the continued action will be as
if the box girder cross-section was open. It is therefore of great importance that the
box girder with open cross-section is rigid enough to avoid a collapse. In the case
of an open cross-section it is not the stresses that are the main problem, but rather
rotations and lateral displacements.
The box girder with an open cross-section has been studied for various initial imper-
fections and comparison of the results given by the finite element models and analyt-
ical analyses considering second degree differential equations have been made [15].

Closed cross-section
Partly closed cross-section
Open cross-section

Figure 7.50: Lateral displacements along top flange 1 for the three different models
with an initial imperfection of 0.05

81
CHAPTER 7. LATERAL TORSIONAL STABILITY

0.035
Open corsssectio n
Closed crosssectio n
Partly closed crosssectio n
0.03

0.025

0.02
θ (rad)

0.015

0.01

0.005

0
0 10 20 30 40 50 60
x (m)

Figure 7.51: Rotations for the three different models with an initial imperfection of
0.05

The analytical analysis, with an initial imperfection of 0.05, gives an ultimate load
of 2.13 times full casting load. The corresponding value from the finite element anal-
ysis is 2.15. The analyses with different initial imperfections show that relatively
larger rotations and lateral displacements are obtained with an increased initial im-
perfection. Consequently, the ultimate load is also decreased with increased initial
imperfections.

82
Chapter 8

Conclusions

The main objective of this thesis was to study the lateral torsional stability of bridge
Y288 over river Ljungan, using finite element analysis performed in BRIGADE/Plus.
A secondary objective was to examine the possible usefulness of BRIGADE/Plus
for these types of designs.

8.1 Finite Element Calculations

Three different types of finite element analyses were performed: a simply supported
rectangular plate uniformly compressed in one direction, profiled sheeting subjected
to shear forces and finally the lateral torsional stability of the bridge.
The aim of the first analysis was to use a simple model for which analytical re-
sults were available and compare them with finite element modelling results. The
analytical solution of this problem is due to Timoshenko, and the finite element
analysis provided results well in accordance with the analytical results. According
to Timoshenko’s analytical solution, the critical buckling stress should be 74.8 MPa.
The finite element analyses provided results of the critical buckling stress between
67.1 MPa and 73.6 MPa, depending on the accuracy of the model. The difference
can be derived to that the load is applied as a uniformly compressive force in the
analytical solution, while it was applied as a number of concentrated forces in the
finite element analysis. The more concentrated forces that were applied, the more
accurate were the results. The critical buckling stress of 73.6 MPa was obtained
when 65 concentrated forces were applied at each side.
The analysis of the profiled sheeting in shear aimed to study the importance of differ-
ent attachment techniques of the profiled sheeting. With an all around attachment,
the structure is much more capable to resist shear forces. If the profiled sheeting is
attached merely in one direction (here the axial direction of the bridge) large axial
forces arise at the edges of the profiled sheets. This might lead to a pull through ac-
tion at the fasteners, leading to a loss of the stabilising effects of the profiled sheeting
which ultimately might lead to a collapse of the bridge. These force concentrations
are considerably reduced with an all around attachment of the profiled sheeting.

83
CHAPTER 8. CONCLUSIONS

These analyses were performed on models containing five profiled sheets and two or
four plates, depending on type of attachment. When the profiled sheeting is given
an all around attachment, a 94% reduction of stresses is achieved compared to when
it is attached only at two sides. Large stress concentrations also appear along the
free edges of the profiled sheeting for the model that has an attachment along two
sides. These concentrations are avoided with an all around attachment.
Furthermore, the importance of the thickness of the profiled sheeting was studied.
This study showed are clear correlation between decreased thickness and increased
stresses in the profiled sheets. When a 10% thinner profiled sheeting was used,
an about 20% increase of the stresses was obtained. If the profiled sheeting was
made 10% thicker, the stresses were decreased about 15%. Although it is a well
known fact that a thinner profiled sheeting provides a weaker structure, it is of
great importance to consider the thickness in this case where the structure already
is weak. An additional attenuation, from the use of too thin profiled sheeting, might
give rise to a pull through action at the fasteners.
The third, and final, main aim was to study the overall behaviour of the bridge.
Primarily the lateral torsional stability and how it is affected by initial imperfections
and various attachment choices of the profiled sheeting was studied. When a profiled
sheeting is used to create a closed cross-section, the structure becomes much more
rigid compared to an open cross-section. The rotations and lateral displacements
are significantly reduced when a closed cross-section is used. In order to successfully
get the profiled sheeting to create a closed cross-section that is sustainable, two
aspects are especially important: First, that the profiled sheeting is given an all
around attachment. If it is attached along two sides only, large axial forces arise
in parts of the profiled sheeting closest to the edge. These forces might lead to a
pull through action at the fasteners, leading to a loss of the stabilising effect of the
profiled sheeting. If that happens, the cross-section of the bridge behaves like an
open, thin-walled cross-section. Second main consideration is to keep the stresses
below the yield point. Even though an all around attachment is made, large stresses
arise in the profile-bottoms at the connections to the top flanges of the box girder.
Small initial imperfections help reduce the stresses while thicker profiled sheeting
together with sturdy connections between the profiled sheeting and box girder makes
the profiled sheeting more capable to resist the arisen stresses. To obtain accurate
quantitative values of the arisen stresses in the profile bottoms a more detailed
modelling of the connections between the profiled sheeting and the top flanges is
needed. In this study, two different variants were analysed. One where a part of
the surface of the profile bottom is tied to the top flange and one where just one
node is tied. The former assumes a smeared-out connection that is more widespread
than in reality, leading to an underestimation of the stresses. The latter produces
a connection that is more concentrated than in reality, providing too large stresses.
However, it can be concluded that it is vital to carefully consider both the thickness
of the profiled sheeting and the way in which it is attached to the box girder.
Another aspect is that if the stabilising effect from the profiled sheeting is lost, the
box girder with the open cross-section ought to be able to withstand the continued
action. From the analyses of the open cross-section it was shown that the magnitude

84
 8.2. SUGGESTED DIRECTION FOR FURTHER RESEARCH

of the initial imperfections has great influence of the bridge behaviour. An increasing
initial imperfection leads to larger rotations, lateral displacements and decreasing
ultimate load. What furthermore will happen, which is not shown in these models,
is that if the rotations become large the wet concrete will run towards one side of
the bridge, acting as an eccentric load that further increases the rotations.

Modelling Problems

Performing finite element analyses of this kind involves much time spend on trial
and error before a model that provides results that are trustworthy has been created.
Since this is very time-consuming it is helpful to find out about problems that others
have had in order to avoid them. Therefore, one of the modelling problems that cost
me a lot of time will be described below.
In the study of the lateral torsional stability geometric imperfections were to be
implemented in the analysis. A linear eigenvalue buckling prediction was performed
in order to find the overall buckling mode corresponding to the geometric initial
imperfection of the box girder. This overall buckling mode, together with some
random local buckling modes, was then implemented as an initial imperfection in
a postbuckling analysis. Finding this overall buckling mode for the model with the
profiled sheeting included turned out to be quite complicated. The linear eigenvalue
buckling prediction only resulted in numerous local buckling modes, mainly in the
profiled sheeting. To circumvent this problem, a linear eigenvalue buckling predic-
tion was performed on the box girder without the profiled sheeting and the overall
buckling mode from that analysis was implemented in the postbuckling analysis of
the model with the profiled sheeting included. This course of action led to another
problem, since the linear eigenvalue buckling prediction and the postbuckling analy-
sis were performed on models with different geometric definitions. Furthermore, the
connections between the profiled sheeting and the box girder in the postbuckling
analysis have to be defined prior to implementing the initial imperfection from the
linear eigenvalue buckling prediction. This two circumstances created vast difficul-
ties creating a converging solution. The final solution of this problem was found
when the connections were made with tie-interactions without adjusting the initial
position of the slave nodes.

8.2 Suggested Direction for Further Research

It would be interesting to perform a study of the influence of the warping restraint

casting at the abutments. This was originally intended to be studied in this thesis,
but time constraints forced an omission of this part. Such a study would include
modelling of the concrete and consideration of the action between the steel box girder
and the concrete. That study could determine how much the torsional stability
would be increased by an initial casting at the abutments.
A study that includes a careful modelling of the connections between the profiled

85
CHAPTER 8. CONCLUSIONS

sheeting and the top flanges of the box girder would give more accurate results
of the magnitude of the stresses at the connections. However, there are about
600 connections between the profiled sheeting and the top flanges, and about 420
between the profiled sheets. Thus, careful modelling these connections would be
extremely time-consuming.

86
Bibliography

[1] http://www.rydab.se (May 2003).

[2] Letter from Skandinavisk bro- & industriteknik AB (March 2003).

[3] http://www.scanscot.com (April 2003).

[4] Private communication with Johan Kölfors at Scanscot Technology (June 2003).

[5] Boverket. Boverkets handbok om snö- och vindlast, utgåva 2. Tryckeri Balder,
Karlskrona, 1997.

[6] Boverket. Boverkets handbok om stålkonstruktioner, BSK 99. Tryckeri

Balder, Karlskrona, 1999.

[7] Chen, W. F., and Lui, E. M. Structural stability: theory and implementa-
tion. Elsevier Science Publishing, New York, 1987.

[8] Cook, R. D., Malkus, D. S., and Plesha, M. E. Concepts and appli-
cations of finite element analysis, third ed. John Wiley & Sons, New York,
1989.

[9] Cook, R. D., Malkus, D. S., Plesha, M. E., and Witt, R. J. Concepts
and applications of finite element analysis, fourth ed. John Wiley & Sons, New
York, 2002.

[10] Hibbitt, Karlsson, and Sorensen. ABAQUS/Standard User’s Manual,

6.3 ed. Hibbitt, Karlsson and Sorensen Inc, Pawtucket, 2002.

[11] Huebner, K. H., and Thornton, E. A. The Finite Element Method for
Engineers. John Wiley & Sons, New York, 1982.

[12] Murray, N. W. Introduction to the theory of thin-walled structures. Oxford

University Press, Oxford, 1984.

[13] Oehlers, D. J., and Bradford, M. A. Composite Steel and Concrete

Structural Members: Fundamental Behaviour. Elsevies Science Ltd, Oxford,
1995.

[14] Ryall, M. J., Parke, G. A. R., and Harding, J. E. Manual of Bridge

Engineering. Thomas Telford Publishing, London, 2000.

87
BIBLIOGRAPHY

[15] Thomasson, P-O. MathCAD program: “Kontroll av sidostabilitet i byg-

gskedet” (January 2003).

[16] Timoshenko, S. Theory of Elastic Stabililty. McGraw-Hill, New York, 1936.

[17] Tyréns AB. Design drawings for bridge Y288 over river Ljungan at Erikslund.

[18] Vägverket. Publikation 1994:1. Allmän teknisk beskrivning för broar,

BRO 94. Division Väg & Trafik, Borlänge, 1994.

[19] Zienkiewicz, O. C., and Cheung, Y. K. The Finite Element Method in

Structural and Continuum Mechanics. McGraw-Hill, London, 1967.

88
Appendix A

ABAQUS Template

A.1 Template

This template provides general information of various aspects of an ABAQUS/Stand-

ard analysis. For more detailed understanding of included, or general understanding
of excluded options and parameters, see the ABAQUS manual.

• *BOUNDARY
Prescribes boundary conditions at nodes. Boundary conditions can be given as
model input data to define zero-valued boundary conditions or as history input
data to add, modify, or remove zero-valued or nonzero boundary conditions.
Options for this command,

Direct: Specify the node set or node number to which the boundary con-
ditions apply, degree of freedom or the first and last of a range
of degrees of freedom to be specified, and the magnitude of the
boundary condition. Set the TYPE parameter to DISPLACE-
MENT to give a displacement history, default in ABAQUS. Set
TYPE=VELOCITY to give a velocity history. Set
TYPE=ACCELERATION to give a acceleration history.
Type: The type of boundary condition can be specified instead of degrees
of freedom. The following boundary condition types are available:

XSYMM Symmetry about a plane X = constant (dof 1, 5, 6 = 0)

YSYMM Symmetry about a plane Y = constant (dof 2, 4, 6 = 0)
ZSYMM Symmetry about a plane Z = constant (dof 3, 4, 5 = 0)
ENCASTRE Fully built-in (dof 1, 2, 3, 4, 5, 6 = 0)
PINNED Pinned (dof 1, 2, 3 = 0)
XASYMM Antisymmetry about a plane X = constant (dof 2, 3, 4 = 0)
YASYMM Antisymmetry about a plane Y = constant (dof 1, 3, 5 = 0)
ZASYMM Antisymmetry about a plane Z = constant (dof 1, 2, 6 = 0)

89
APPENDIX A. ABAQUS TEMPLATE

Op: The optional parameter OP can be set to one of two values, MOD
or NEW.
Set OP=MOD to modify or OP=NEW to add zero-valued or
nonzero boundary conditions. If OP=NEW is used, all previ-
ously defined boundary conditions will be removed.

Examples: *BOUNDARY, TYPE=DISPLACEMENT, OP=MOD

TOP, 2, 2, 0.5E-3
BOTTOM, 1, 3, 0
RSIDE 1, 1, 0
Here the node set TOP is set to move 0.5 mm in dof 2, and the
node set BOTTOM is prescribed a zero-value in dof:s 1-3.

*BOUNDARY, TYPE=DISPLACEMENT, OP=NEW

TOP, 1, 1, 3E-3
BOTTOM, ENCASTRE
By changing to OP=NEW, the node set TOP is here set to move
3 mm in dof 1, and the node set BOTTOM is prescribed a zero-
value in all dof:s. Note that the node set RSIDE is left out, thus
removing previously defined boundary conditions for this node set.

• *BUCKLE
Controls eigenvalue buckling estimation.
Options for this command,

Eigensolver: This parameter can be set one of two values, SUBSPACE or

LANCZOS. Below follows examples of the Lanczos and the Sub-
space eigensolver.
Examples: *BUCKLE, EIGENSOLVER=LANCZOS
10, 0.2, 0.5,,
The first digit sets the number of eigenvalues to be estimated.
The second digit sets the minimum value of interest.
The third digit sets the maximum value of interest.
The fourth digit sets the block size. Here, this entry is omitted
and a default value is created.
The fifth digit sets the maximum number of block Lanczos steps
within each Lanczos run. Here, this entry is omitted and a default
value is created.

*BUCKLE, EIGENSOLVER=SUBSPACE
10, 0.2, 18, 30
The first digit sets the number of eigenvalues to be estimated.
The second digit sets the maximum value of interest.
The third digit sets the number of vectors in the iteration. In gen-
eral, the convergence in solving the eigenproblem is more rapid if

90
 A.1. TEMPLATE

more vectors are carried in the iteration. The default number of

vectors used is the minimum of 2n or n + 8
The fourth digit sets the maximum number of iterations. The de-
fault is 30.

• *CLOAD
Specifies concentrated forces and moments.
Options for this command,
Amplitude: Set the parameter AMPLITUDE equal to the name of the am-
plitude curve that defines the magnitude of the load during the
step. If omitted, the reference magnitude is applied immediately
at the beginning of the step or linearly over the step.
Follower: Include this parameter if the direction of the load is assumed to
rotate with the rotation at this node. This parameter should be
used only for large-displacement analysis and can be used only
at nodes with active rotational degree of freedom, such as beams
and shells.
Op: The optional parameter OP can be set to one of two values, MOD
or NEW.
Set OP=MOD to modify or OP=NEW to add concentrated forces.
If OP=NEW is used, all previously defined concentrated forces
will be removed.

Example: *CLOAD, OP=MOD

PickedSet1330, 2, -2000
PickedSet1330, 3, -2000
Here the node set PickedSet1330 is set to apply a concentrated
force of magnitude -2000 at d.o.f. 2 and 3.

• *HEADING
This option is used to define a title to the analysis.

Example: *HEADING
Torsional stability of box girder without profiled sheets.

• *IMPERFECTION
Introduces a geometric imperfection into a model for postbuckling analysis.
Options for this command,
FILE: Set the parameter FILE equal to the name of the results file
from a previous analysis containing either the mode shapes from
a *BUCKLE or *FREQUENCY analysis or the nodal displace-
ments from a *STATIC analysis

91
APPENDIX A. ABAQUS TEMPLATE

STEP: Set the parameter STEP equal to the step number from which
model or displacement data are to be read.
Examples: *IMPERFECTION, FILE=BIG-Buckle, STEP=1
1, 0.001
2, 0.001
3, 0.001
9, 0.1
The first digit sets the mode number to be introduced.
The second digit sets the scaling factor for this mode.
Here mode number 1, 2, and 3 are introduced with the scaling fac-
tor 0.001 and mode number 4 with the scaling factor 0.1.

This section deals with material properties and gives an idea of how it could
look like in a input file.

• *MATERIAL, *DENSITY, and *ELASTIC

Defines material properties.
Required parameter,

Name: Set this parameter equal to the name of the parameter to be used
for the material when referenced in the element property option.
The name must be unique and cannot begin with a number.
Example: *MATERIAL, NAME=Steel
**
*DENSITY
7850
**
*ELASTIC
**
2.1e+11, 0.3

• *STATIC
Specifies that the step should be analysed as a static load step.
Options for this command,

Direct: Include this parameter if direct control over the time incrementa-
tion is desirable throughout the step.
Riks: Include this parameter if it is desirable to use the Riks method
when performing the static load step. The load case should be
proportional.
Stabilize: This parameter provides automatic stabilisation to the structure.
Set this parameter equal to the dissipated energy fraction desir-
able. If this parameter is omitted, ABAQUS defaults to a value
of 2 × 10−4 . If the FACTOR parameter is used, any value of the

92
 A.1. TEMPLATE

dissipated energy fraction will be overridden.

This parameter cannot be used if the RIKS parameter is included.
Example: *STATIC
0.1, 1., 1e-05, 0.1
The first digit sets initial time increment.
The second digit sets the time period of the step.
The third digit sets the minimum time increment allowed.
The fourth digit sets the maximum time increment allowed.

• *STEP
Specifies a step definition. It must be followed by a procedure definition.
Options for this command,

Nl-geom: If the NLGEOM (Nonlinear Geometry) parameter is specified on

the *STEP option it will be active in all subsequent steps of the
analysis.
Amplitude: This parameter can be set one of two values, STEP or RAMP.
By setting AMPLITUDE=STEP, Figure A.1 a, the loads are ap-
plied instantaneously at the start of the step. By setting AM-
PLITUDE=RAMP, Figure A.1 b, the load magnitude will vary
linearly over the step.
Inc: Set this parameter equal to the maximum number of increments
in the step. The default value is 100.
Example: *STEP, NLGEOM, AMPLITUDE=RAMP, INC=200
The step defined in this example states that the loading will vary
linearly over the step, that large-displacement will be included in
the step and that the step will consist of 200 time increments.

(a) (b)

Figure A.1: Stepload and Rampload

93
APPENDIX A. ABAQUS TEMPLATE

• *TIE
Constrains each node on the slave surface to have the same value of displace-
ment as the point on the master surface to which it is closest.
Required parameter,

Name: Set this parameter equal to the name that will be used to identify
this constraint.

Options for this parameter,

Position tolerance: Set this parameter equal to the distance within which nodes
on the slave surface must lie from the master surface in order to
be tied. The default value for this tolerance distance is 5% of
the typical element size in the master surface. If a node-based
master surface is used, the default tolerance distance is based on
the average distance between nodes in the master surface.
Adjust: Set ADJUST=YES (default) to move all tied nodes on the slave
surface onto the master surface in the initial configuration, with-
out any strain.
No rotation: Include this parameter if rotation degrees of freedom should not
be tied. If this parameter is omitted, any existing rotation degrees
of freedom will be tied if applicable, in addition to the translation
degrees of freedom.
Example: *TIE, NAME=name
slave surface name, master surface name
*TIE, NAME=tpy9-tpy10, ADJUST=yes, NO ROTATION
PickedSet1350-CNS, PickedSurf1349
Here the node set PickedSet1350-CNS will be tied in degrees of
freedom 1, 2 and 3 to PickedSurf1349. All nodes in the slave node
set within the default position tolerance will be moved to the mas-
ter surface.

94
Appendix B

ABAQUS Input Files

This template provides a summary of the input files used of the finite element
analyses. The assembly has been left out from each input file to make this appendix
reasonable short. An example of how the assembly might be created is given after the
input files. For more detailed understanding of included, or general understanding
of excluded parts of the input files, contact the author of this thesis.

B.1 Profiled Sheets in Shear

B.1.1 Long side Attachment Input File

*Heading
PROFILED SHEETS IN SHEAR, ATTACHED AT LONGSIDES
Units: Length - m, Force - N, Stress - N/m2
** Job name: FiveTP-26 Model name: Model-1
**
**-----------------DEFINING THE PARTS-----------------
**
*Part, name=Plate
*End Part
*Part, name=TrapSheet
*End Part
**
**----------------ASSEMBLY OF THE PARTS----------------
**
The assembly is left out.
*End Assembly
**
**---------------------MATERIAL DATA---------------------
**
** MATERIALS

95
APPENDIX B. ABAQUS INPUT FILES

**
*Material, name=PlattaSection
*Density
7850.,
*Elastic
2.1e+11, 0.3
*Material, name=TRP-Plåt
*Density
7850.,
*Elastic
2.1e+11, 0.3
**
**-----------------INTERACTION DEFINITIONS-----------------
**
** INTERACTION PROPERTIES
**
*Surface Interaction, name=IntProp-1
1.,
*Surface Behavior, pressure-overclosure=HARD
**
**-------------------BOUNDARY CONDITIONS-------------------
**
** BOUNDARY CONDITIONS
**
** Name: Fixlager Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet491, ENCASTRE
** Name: rull2 Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet492, XASYMM
**
** INTERACTIONS
**
** Interaction: Int-1
*Contact Pair, interaction=IntProp-1
PickedSurf505, PickedSurf504
** Interaction: Int-2
*Contact Pair, interaction=IntProp-1
PickedSurf507, PickedSurf506
** -------------------------------------------------------
**
********************************************************
**ANALYSIS STEP - APPLYING LOADS
********************************************************
** STEP: General
**
*Step, name=General

96
 B.1. PROFILED SHEETS IN SHEAR

*Static
0.1, 1., 1e-05, 0.1
**
** LOADS
**
** Name: DisplacementLoad Type: Concentrated force
*Cload
PickedSet515, 1, 10000.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*El Print, freq=999999
*Node Print, freq=999999
*End Step

B.1.2 All Around Attachment Input File

*Heading
PROFILED SHEETS IN SHEAR, ATTACHED ALL AROUND
Units: Length - m, Force - N, Stress - N/m2
** Job name: FiveTParound-13 Model name: Model-1
**
**-----------------DEFINING THE PARTS-----------------
**
*Part, name=Plate
*End Part
*Part, name=PlateShort
*End Part
*Part, name=TrapSheet
*End Part
**
**----------------ASSEMBLY OF THE PARTS----------------
**
The assembly is left out.
*End Assembly
**

97
APPENDIX B. ABAQUS INPUT FILES

**---------------------MATERIAL DATA---------------------
**
** MATERIALS
**
*Material, name=PlattaSection
*Density
7850.,
*Elastic
2.1e+11, 0.3
*Material, name=TRP-Plåt
*Density
7850.,
*Elastic
2.1e+11, 0.3
**
**-----------------INTERACTION DEFINITIONS-----------------
**
** INTERACTION PROPERTIES
**
*Surface Interaction, name=IntProp-1
1.,
*Surface Behavior, pressure-overclosure=HARD
**
**-------------------BOUNDARY CONDITIONS-------------------
**
** BOUNDARY CONDITIONS
**
** Name: Fixlager Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet491, ENCASTRE
** Name: rull2 Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet549, XASYMM
**
** INTERACTIONS
**
** Interaction: Int-1
*Contact Pair, interaction=IntProp-1
PickedSurf505, PickedSurf504
** Interaction: Int-2
*Contact Pair, interaction=IntProp-1
PickedSurf507, PickedSurf506
** -------------------------------------------------------
**
********************************************************
**ANALYSIS STEP - APPLYING LOADS

98
 B.2. TORSIONAL STABILITY

********************************************************
** STEP: General
**
*Step, name=General
*Static
0.1, 1., 1e-05, 0.1
**
** LOADS
**
** Name: DisplacementLoad Type: Concentrated force
*Cload
PickedSet584, 1, 10000.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*El Print, freq=999999
*Node Print, freq=999999
*End Step

B.2 Torsional Stability

B.2.1 Box Girder with Open Cross-Section

Prebuckling Input File

*Heading
TORSIONAL STABILITY OF BOX GIRDER WITH OPEN CROSS-SECTION
Units: Length - m, Force - N, Stress - N/m2
** Job name: BIG-Buckle Model name: Model-1
**
**-----------------DEFINING THE PARTS-----------------
**
The assembly is left out.
*Part, name=Y288
*End Part

99
APPENDIX B. ABAQUS INPUT FILES

**
**----------------ASSEMBLY OF THE PARTS----------------
**
The assembly is left out.
*End Assembly
**
**---------------------MATERIAL DATA---------------------
**
*Material, name=Y288-Steel
*Density
7850.,
*Elastic
2.1e+11, 0.3
**
**-------------------BOUNDARY CONDITIONS-------------------
**
** Name: Support,Fix Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet888, PINNED
** Name: Support,Moveable Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet889, XASYMM
** -------------------------------------------------------
**
********************************************************
** STEP - EIGENVALUE BUCKLING PREDICTION
********************************************************
** STEP: Buckling
**
*Step, name=Buckling, perturbation
*Buckle, eigensolver=lanczos
100, 0., ,
**
** BOUNDARY CONDITIONS
**
** Name: Support,Fix Type: Symmetry/Antisymmetry/Encastre
*Boundary, op=NEW, load case=1
PickedSet888, PINNED
*Boundary, op=NEW, load case=2
PickedSet888, PINNED
** Name: Support,Moveable Type: Symmetry/Antisymmetry/Encastre
*Boundary, op=NEW, load case=1
PickedSet889, XASYMM
*Boundary, op=NEW, load case=2
PickedSet889, XASYMM
**
** LOADS

100
 B.2. TORSIONAL STABILITY

**
** Name: BucklingLoad Type: Concentrated force
*Cload
PickedSet887, 3, -25000.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-2
**
*Output, field, variable=PRESELECT
*El Print, freq=999999
*Node Print, freq=999999
*Node File, freq=999999
U
*End Step

Postbuckling Input File

*HEADING
TORSIONAL STABILITY OF BOX GIRDER WITH OPEN CROSS-SECTION
** Job name: BIG-pb1 Model name: Model-1
**
**-----------------DEFINING THE PARTS-----------------
**
*Part, name=Y288
*End Part
**
**----------------ASSEMBLY OF THE PARTS----------------
**
The assembly is left out.
*End Assembly
**
**---------------------MATERIAL DATA---------------------
**
*Material, name=Y288-Steel
*Density
7850.,
*Elastic
2.1e+11, 0.3
**
**---------------INTRODUCING IMPERFECTIONS---------------
**
*Imperfection, file=BIG-Buckle-2, Step=1
1, 1e-3

101
APPENDIX B. ABAQUS INPUT FILES

2, 1e-3
3, 1e-3
9, 0.05
**
**-------------------BOUNDARY CONDITIONS-------------------
**
** Name: Support,Fix Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet889, PINNED
** Name: Support,Moveable Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet890, XASYMM
** ------------------------------------------------------
**
********************************************************
** STEP 1 - APPLYING GRAVITY, FORMWORK AND WIND
********************************************************
** STEP: General
**
*Step, name=General, nlgeom
*Static
0.1, 1., 1e-05, 0.1
**
** LOADS
**
** Name: GravityLoad Type: Gravity
*Dload
PickedSet875, GRAV, 9.81, 0., 0., -1.
** Name: Formwork Type: Pressure
*Dsload
PickedSurf903, P, 5000.
** Name: Wind Type: Pressure
*Dsload
PickedSurf905, P, 1500.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT

102
 B.2. TORSIONAL STABILITY

*El Print, freq=999999

*Node Print, freq=999999
*End Step
** ------------------------------------------------------
**
********************************************************
** STEP 2 - CAST CONCRETE
********************************************************
** STEP: Casting
**
*Step, name=Casting, nlgeom
*Static, riks
0.2, 1., 1e-05, 0.2, 3,
**
** LOADS
**
** Name: Casting Type: Pressure
*Dsload
PickedSurf904, P, 45000.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*End Step

B.2.2 Box Girder with Closed Cross-Section

Prebuckling Input File

The same file is used for the prebuckling analysis as in the

analysis with the open cross-section

Postbuckling Input File

*Heading
TORSIONAL STABILITY OF BOX GIRDER WITH CLOSED CROSS-SECTION
Units: Length - m, Force - N, Stress - N/m2

103
APPENDIX B. ABAQUS INPUT FILES

** Job name: BIG-TP-pb2 Model name: Model-1

**
**-----------------DEFINING THE PARTS-----------------
**
*Part, name="TP,inner"
*End Part
*Part, name="TP,ytter"
*End Part
*Part, name=Y288
*End Part
**
**----------------ASSEMBLY OF THE PARTS----------------
**
The assembly is left out.
*End Assembly
**
**---------------------MATERIAL DATA---------------------
**
*Material, name=Sheet
*Density
7850.,
*Elastic
2.1e+11, 0.3
*Material, name=Y288-Steel
*Density
7850.,
*Elastic
2.1e+11, 0.3
**
*Imperfection, file=BIG-TP-Buckle, Step=1
1, 1e-3
2, 1e-3
3, 1e-3
9, 0.05
**
**-------------------BOUNDARY CONDITIONS-------------------
**
** Name: Support,Fix Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet1243, PINNED
** Name: Support,Moveable Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet1244, XASYMM
** -------------------------------------------------------
**
********************************************************

104
 B.2. TORSIONAL STABILITY

** STEP 1 - APPLYING GRAVITY, FORMWORK AND WIND

********************************************************
** STEP: General
**
*Step, name=General, nlgeom
*Static
0.1, 1., 1e-08, 0.1
**
** LOADS
**
** Name: Formwork Type: Pressure
*Dsload
PickedSurf1469, P, 5000.
** Name: Gravity Type: Gravity
*Dload
PickedSet1480, GRAV, 9.81, 0., 0., -1.
** Name: Wind Load Type: Pressure
*Dsload
PickedSurf1470, P, 1500.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*El Print, freq=999999
*Node Print, freq=999999
*End Step
** ------------------------------------------------------
**
********************************************************
** STEP 2 - CAST CONCRETE
********************************************************
** STEP: Casting
**
*Step, name=Casting, nlgeom
*Static, riks
0.1, 1., 1e-08, 0.1, 1.,
**
** LOADS

105
APPENDIX B. ABAQUS INPUT FILES

**
** Name: Casting Type: Pressure
*Dsload
PickedSurf1471, P, 45000.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*End Step

B.2.3 Box Girder with Partly Closed Cross-Section

Prebuckling Input File

The same file is used for the prebuckling analysis as in the

analysis with the open cross-section

Postbuckling Input File

*Heading
** Job name: BIG-TP-Missing-pb4 Model name: Model-1
**
** PARTS
**
*Part, name="TP,inner"
*End Part
*Part, name="TP,ytter"
*End Part
*Part, name=Y288
*End Part
**
** ASSEMBLY
**
The assembly is left out.
*End Assembly
**
** MATERIALS

106
 B.2. TORSIONAL STABILITY

**
*Material, name=Sheet
*Density
7850.,
*Elastic
2.1e+11, 0.3
*Material, name=Y288-Steel
*Density
7850.,
*Elastic
2.1e+11, 0.3
**
*Imperfection, file=BIG-TP-Buckle-3, Step=1
1, 1e-3
2, 1e-3
3, 1e-3
9, 0.05
**
** BOUNDARY CONDITIONS
**
** Name: Support,Fix Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet1243, PINNED
** Name: Support,Moveable Type: Symmetry/Antisymmetry/Encastre
*Boundary
PickedSet1244, XASYMM
** ----------------------------------------------------------------
**
********************************************************
** STEP 1 - APPLYING GRAVITY, FORMWORK AND WIND
********************************************************
** STEP: General
**
*Step, name=General, nlgeom
*Static
0.1, 1., 1e-08, 0.1
**
** LOADS
**
** Name: Formwork Type: Pressure
*Dsload
PickedSurf1469, P, 5000.
** Name: Gravity Type: Gravity
*Dload
PickedSet1480, GRAV, 9.81, 0., 0., -1.
** Name: Wind Load Type: Pressure
*Dsload

107
APPENDIX B. ABAQUS INPUT FILES

PickedSurf1470, P, 1500.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*El Print, freq=999999
*Node Print, freq=999999
*End Step
** ----------------------------------------------------------------
**
********************************************************
** STEP 2 - CAST CONCRETE
********************************************************
** STEP: Casting
**
*Step, name=Casting, nlgeom
*Static, riks
0.1, 1., 1e-08, 0.1, 1.,
**
** LOADS
**
** Name: Casting Type: Pressure
*Dsload
PickedSurf1471, P, 45000.
**
** OUTPUT REQUESTS
**
*Restart, write, frequency=1
**
** FIELD OUTPUT: F-Output-1
**
*Output, field, variable=PRESELECT
**
** HISTORY OUTPUT: H-Output-1
**
*Output, history, variable=PRESELECT
*End Step

108
 B.2. TORSIONAL STABILITY

B.2.4 Example of Assembly

The following is a shorten version of the assembly for the postbuckling analysis of
the model with the partly closed cross-section.
*Heading
** Job name: BIG-TP-Missing-pb4 Model name: Model-1
**
** PARTS
**
*Part, name="TP,inner"
*End Part
*Part, name="TP,ytter"
*End Part
*Part, name=Y288
*End Part
**
** ASSEMBLY
**
*Assembly, name=Assembly
**
*Instance, name=Y288-1, part=Y288
*Node
1, -14.25, 1.368, 2.015
2, -14.25, 1.418, 2.015
3, -14.25, 1.818, 2.015
...
10075, -27.9713, -1.683858, 0.
10076, -27.95896, -1.714647, 0.
*Element, type=S4R
1, 1, 2, 3138, 384
2, 2, 3, 363, 3138
...
10001, 10063, 10062, 3132, 3133
10002, 10062, 3131, 362, 3132
** Region: (Överfläns0,038:Picked)
*Elset, elset=I1, internal
1, 2, 3,..., 1117, 1118
** Section: Överfläns0,038
*Shell Section, elset=I1, material=Y288-Steel
0.038, 5
** Region: (Överflänsplåt0,025:Picked)
*Elset, elset=I2, internal
217, 218, 219,...,1357, 1358
** Section: Överflänsplåt0,025
*Shell Section, elset=I2, material=Y288-Steel
0.025, 5

109
APPENDIX B. ABAQUS INPUT FILES

** Region: (Överfläns0,025:Picked)
*Elset, elset=I3, internal
247, 248, 249,..., 1505, 1506
** Section: Överfläns0,025
*Shell Section, elset=I3, material=Y288-Steel
0.025, 5
** Region: (Livplåt0,019:Picked)
*Elset, elset=I4, internal
1507, 1508, 1509,..., 9855, 9856
** Section: Livplåt0,019
*Shell Section, elset=I4, material=Y288-Steel
0.019, 5
** Region: (Y288-BoxGirder:Picked)
*Elset, elset=I5, internal
1859, 1860, 1861,..., 9911, 9912
** Section: Y288-BoxGirder
*Shell Section, elset=I5, material=Y288-Steel
0.03, 5
** Region: (Livplåt0,016:Picked)
*Elset, elset=I6, internal
1864, 1865, 1866,..., 7020, 7021
** Section: Livplåt0,016
*Shell Section, elset=I6, material=Y288-Steel
0.016, 5
** Region: (Underfläns0,02:Picked)
*Elset, elset=I7, internal
2435, 2436, 2437,..., 8299, 8300
** Section: Underfläns0,02
*Shell Section, elset=I7, material=Y288-Steel
0.02, 5
** Region: (Underfläns0,032:Picked)
*Elset, elset=I8, internal
2599, 2600, 2601,..., 9025, 9026
** Section: Underfläns0,032
*Shell Section, elset=I8, material=Y288-Steel
0.032, 5
** Region: (Avstyvning0,025:Picked)
*Elset, elset=I9, internal
3112, 3113, 3114,..., 10001, 10002
** Section: Avstyvning0,025
*Shell Section, elset=I9, material=Y288-Steel
0.025, 5
** Region: (Underfläns0,02:Picked)
*Elset, elset=I10, internal
3402, 3403, 3448,..., 9883, 9884
** Section: Underfläns0,02

110
 B.2. TORSIONAL STABILITY

*Shell Section, elset=I10, material=Y288-Steel

0.02, 5
** Region: (Ändplåt0,02:Picked)
*Elset, elset=I11, internal
3404, 3405, 3406,..., 9452, 9453
** Section: Ändplåt0,02
*Shell Section, elset=I11, material=Y288-Steel
0.02, 5
** Region: (Tvärskott0,02:Picked)
*Elset, elset=I12, internal
3540, 3541, 3542,..., 7244, 7245
** Section: Överflänsplåt0,015
*Shell Section, elset=I14, material=Y288-Steel
0.015, 5
** Region: (Tvärskott0,01:Picked)
*Elset, elset=I15, internal
4883, 4884, 4885,..., 7658, 7659
** Section: Tvärskott0,01
*Shell Section, elset=I15, material=Y288-Steel
0.01, 5
*End Instance
**
*Instance, name="TP,ytter-2", part="TP,ytter"
27.1125, 1.70031441048035, 2.015
27.1125, 1.70031441048035, 2.015
28.1125, 1.70031441048035, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.03495415
2, -0.4775, 0., 0.0349083
3, -0.4775, 0., 3.368
...
689, 0.4925, 0., 3.034437
690, 0.4925, 0., 3.201218
*Element, type=S4R
1, 1, 2, 117, 116
2, 116, 117, 118, 115
3, 115, 118, 119, 114
...
637, 2, 1, 576, 112
638, 1, 15, 113, 576
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 638, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5

111
APPENDIX B. ABAQUS INPUT FILES

*End Instance
**
*Instance, name="TP,ytter-3", part="TP,ytter"
26.1625, 1.70031441048035, 2.015
26.1625, 1.70031441048035, 2.015
27.1625, 1.70031441048035, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.03495415
2, -0.4775, 0., 0.0349083
3, -0.4775, 0., 3.368
...
689, 0.4925, 0., 3.034437
690, 0.4925, 0., 3.201218
*Element, type=S4R
1, 1, 2, 117, 116
2, 116, 117, 118, 115
3, 115, 118, 119, 114
...
637, 2, 1, 576, 112
638, 1, 15, 113, 576
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 638, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
**
*Instance, name="TP,ytter-11", part="TP,ytter"
18.5625, 1.70031441048035, 2.015
18.5625, 1.70031441048035, 2.015
19.5625, 1.70031441048035, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.03495415
2, -0.4775, 0., 0.0349083
3, -0.4775, 0., 3.368
...
689, 0.4925, 0., 3.034437
690, 0.4925, 0., 3.201218
*Element, type=S4R
1, 1, 2, 117, 116
2, 116, 117, 118, 115
3, 115, 118, 119, 114
...
637, 2, 1, 576, 112
638, 1, 15, 113, 576

112
 B.2. TORSIONAL STABILITY

** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 638, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
**
*Instance, name="TP,inner-1", part="TP,inner"
17.6125, 1.40040611353712, 2.015
17.6125, 1.40040611353712, 2.015
18.6125, 1.40040611353712, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.025
2, -0.4775, 0., 0.025
3, -0.4775, 0., 2.775
...
569, -0.4575, 0., 2.43125
570, -0.4575, 0., 2.603125
*Element, type=S4R
1, 1, 2, 115, 114
2, 114, 115, 116, 4
3, 4, 116, 117, 136
...
521, 10, 130, 466, 467
522, 130, 11, 72, 466
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 522, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
**
*Instance, name="TP,inner-2", part="TP,inner"
16.6625, 1.40040611353712, 2.015
16.6625, 1.40040611353712, 2.015
17.6625, 1.40040611353712, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.025
2, -0.4775, 0., 0.025
3, -0.4775, 0., 2.775
...
569, -0.4575, 0., 2.43125
570, -0.4575, 0., 2.603125
*Element, type=S4R

113
APPENDIX B. ABAQUS INPUT FILES

1, 1, 2, 115, 114
2, 114, 115, 116, 4
3, 4, 116, 117, 136
...
521, 10, 130, 466, 467
522, 130, 11, 72, 466
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 522, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
**
*Instance, name="TP,inner-38", part="TP,inner"
-17.5375, 1.40040611353712, 2.015
-17.5375, 1.40040611353712, 2.015
-16.5375, 1.40040611353712, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.025
2, -0.4775, 0., 0.025
3, -0.4775, 0., 2.775
...
569, -0.4575, 0., 2.43125
570, -0.4575, 0., 2.603125
*Element, type=S4R
1, 1, 2, 115, 114
2, 114, 115, 116, 4
3, 4, 116, 117, 136
...
521, 10, 130, 466, 467
522, 130, 11, 72, 466
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 522, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
**
*Instance, name="TP,ytter-12", part="TP,ytter"
-18.4875, 1.69782532751092, 2.015
-18.4875, 1.69782532751092, 2.015
-17.4875, 1.69782532751092, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.03495415

114
 B.2. TORSIONAL STABILITY

2, -0.4775, 0., 0.0349083

3, -0.4775, 0., 3.368
...
689, 0.4925, 0., 3.034437
690, 0.4925, 0., 3.201218
*Element, type=S4R
1, 1, 2, 117, 116
2, 116, 117, 118, 115
3, 115, 118, 119, 114
...
637, 2, 1, 576, 112
638, 1, 15, 113, 576
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 638, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
**
*Instance, name="TP,ytter-13", part="TP,ytter"
-19.4375, 1.69782532751092, 2.015
-19.4375, 1.69782532751092, 2.015
-18.4375, 1.69782532751092, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.03495415
2, -0.4775, 0., 0.0349083
3, -0.4775, 0., 3.368
...
689, 0.4925, 0., 3.034437
690, 0.4925, 0., 3.201218
*Element, type=S4R
1, 1, 2, 117, 116
2, 116, 117, 118, 115
3, 115, 118, 119, 114
...
637, 2, 1, 576, 112
638, 1, 15, 113, 576
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 638, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
**

115
APPENDIX B. ABAQUS INPUT FILES

*Instance, name="TP,ytter-21", part="TP,ytter"

-27.0375, 1.69782532751092, 2.015
-27.0375, 1.69782532751092, 2.015
-26.0375, 1.69782532751092, 2.015, 90.
*Node
1, -0.495, 0.02125, 0.03495415
2, -0.4775, 0., 0.0349083
3, -0.4775, 0., 3.368
...
689, 0.4925, 0., 3.034437
690, 0.4925, 0., 3.201218
*Element, type=S4R
1, 1, 2, 117, 116
2, 116, 117, 118, 115
3, 115, 118, 119, 114
...
637, 2, 1, 576, 112
638, 1, 15, 113, 576
** Region: (y288-TP:Picked)
*Elset, elset=I1, internal, generate
1, 638, 1
** Section: y288-TP
*Shell Section, elset=I1, material=Sheet
0.00086, 5
*End Instance
*Nset, nset=PickedSet1241, internal, instance=Y288-1
68, 69, 78
*Nset, nset=PickedSet1243, internal, instance=Y288-1
305, 309
*Nset, nset=PickedSet1244, internal, instance=Y288-1
220, 226
*Nset, nset=PickedSet1336, internal, instance="TP,ytter-2"
4, 10, 1
*Nset, nset=PickedSet1338, internal, instance="TP,ytter-3"
4, 10, 1
*Nset, nset=PickedSet1354, internal, instance="TP,ytter-11"
4, 10, 1
*Nset, nset=PickedSet1356, internal, instance="TP,inner-1"
4, 10, 1
*Nset, nset=PickedSet1358, internal, instance="TP,inner-2"
4, 10, 1
*Nset, nset=PickedSet1430, internal, instance="TP,inner-38"
4, 10, 1
*Nset, nset=PickedSet1432, internal, instance="TP,ytter-12"
4, 10, 1
*Nset, nset=PickedSet1434, internal, instance="TP,ytter-13"

116
 B.2. TORSIONAL STABILITY

4, 10, 1
*Nset, nset=PickedSet1480, internal, instance="TP,ytter-21"
1, 690, 1
*Elset, elset=PickedSet1480, internal, instance=Y288-1
1, 10002, 1
*Elset, elset=PickedSet1480, internal, instance="TP,ytter-2"
1, 638, 1
*Elset, elset=PickedSet1480, internal, instance="TP,ytter-3"
1, 638, 1
*Elset, elset=PickedSet1480, internal, instance="TP,inner-1"
1, 522, 1
*Elset, elset=PickedSet1480, internal, instance="TP,ytter-13"
1, 638, 1
*Elset, elset=PickedSet1480, internal, instance="TP,inner-2"
1, 522, 1
*Elset, elset=PickedSet1480, internal, instance="TP,inner-38"
1, 522, 1
*Elset, elset=PickedSet1480, internal, instance="TP,ytter-11"
1, 638, 1
** Constraint: LS1-inner
*Tie, name=LS1-inner, adjust=no, position tolerance=10., no rotation
PickedSurf1487, PickedSurf1312
** Constraint: LS1-ytter1
*Tie, name=LS1-ytter1, adjust=no, position tolerance=10., no rotation
PickedSurf1489, PickedSurf1320
** Constraint: LS1-ytter2
*Tie, name=LS1-ytter2, adjust=no, position tolerance=10., no rotation
PickedSurf1490, PickedSurf1322
** Constraint: LS2-inner
*Tie, name=LS2-inner, adjust=no, position tolerance=10., no rotation
PickedSurf1488, PickedSurf1316
** Constraint: LS2-ytter1
*Tie, name=LS2-ytter1, adjust=no, position tolerance=10., no rotation
PickedSurf1491, PickedSurf1324
** Constraint: LS2-ytter2
*Tie, name=LS2-ytter2, adjust=no, position tolerance=10., no rotation
PickedSurf1492, PickedSurf1326
** Constraint: tpi1-tpi2
*Tie, name=tpi1-tpi2, adjust=yes, no rotation
PickedSet1356-CNS, PickedSurf1355
** Constraint: tpi10-tpi11
*Tie, name=tpi10-tpi11, adjust=yes, no rotation
PickedSet1374-CNS, PickedSurf1373
** Constraint: tpi11-tpi12
*Tie, name=tpi11-tpi12, adjust=yes, no rotation
PickedSet1376-CNS, PickedSurf1375

117
APPENDIX B. ABAQUS INPUT FILES

** Constraint: tpi12-tpi13
*Tie, name=tpi12-tpi13, adjust=yes, no rotation
PickedSet1378-CNS, PickedSurf1377
** Constraint: tpi13-tpi14
*Tie, name=tpi13-tpi14, adjust=yes, no rotation
PickedSet1380-CNS, PickedSurf1379
** Constraint: tpi14-tpi15
*Tie, name=tpi14-tpi15, adjust=yes, no rotation
PickedSet1382-CNS, PickedSurf1381
** Constraint: tpi15-tpi16
*Tie, name=tpi15-tpi16, adjust=yes, no rotation
PickedSet1384-CNS, PickedSurf1383
** Constraint: tpi16-tpi17
*Tie, name=tpi16-tpi17, adjust=yes, no rotation
PickedSet1386-CNS, PickedSurf1385
** Constraint: tpi17-tpi18
*Tie, name=tpi17-tpi18, adjust=yes, no rotation
PickedSet1388-CNS, PickedSurf1387
** Constraint: tpi18-tpi19
*Tie, name=tpi18-tpi19, adjust=yes, no rotation
PickedSet1390-CNS, PickedSurf1389
** Constraint: tpi19-tpi20
*Tie, name=tpi19-tpi20, adjust=yes, no rotation
PickedSet1392-CNS, PickedSurf1391
** Constraint: tpi2-tpi3
*Tie, name=tpi2-tpi3, adjust=yes, no rotation
PickedSet1358-CNS, PickedSurf1357
** Constraint: tpi20-tpi21
*Tie, name=tpi20-tpi21, adjust=yes, no rotation
PickedSet1394-CNS, PickedSurf1393
** Constraint: tpi21-tpi22
*Tie, name=tpi21-tpi22, adjust=yes, no rotation
PickedSet1396-CNS, PickedSurf1395
** Constraint: tpi22-tpi23
*Tie, name=tpi22-tpi23, adjust=yes, no rotation
PickedSet1398-CNS, PickedSurf1397
** Constraint: tpi23-tpi24
*Tie, name=tpi23-tpi24, adjust=yes, no rotation
PickedSet1400-CNS, PickedSurf1399
** Constraint: tpi24-tpi25
*Tie, name=tpi24-tpi25, adjust=yes, no rotation
PickedSet1402-CNS, PickedSurf1401
** Constraint: tpi25-tpi26
*Tie, name=tpi25-tpi26, adjust=yes, no rotation
PickedSet1404-CNS, PickedSurf1403
** Constraint: tpi26-tpi27

118
 B.2. TORSIONAL STABILITY

*Tie, name=tpi26-tpi27, adjust=yes, no rotation

PickedSet1406-CNS, PickedSurf1405
** Constraint: tpi27-tpi28
*Tie, name=tpi27-tpi28, adjust=yes, no rotation
PickedSet1408-CNS, PickedSurf1407
** Constraint: tpi28-tpi29
*Tie, name=tpi28-tpi29, adjust=yes, no rotation
PickedSet1410-CNS, PickedSurf1409
** Constraint: tpi29-tpi30
*Tie, name=tpi29-tpi30, adjust=yes, no rotation
PickedSet1412-CNS, PickedSurf1411
** Constraint: tpi3-tpi4
*Tie, name=tpi3-tpi4, adjust=yes, no rotation
PickedSet1360-CNS, PickedSurf1359
** Constraint: tpi30-tpi31
*Tie, name=tpi30-tpi31, adjust=yes, no rotation
PickedSet1414-CNS, PickedSurf1413
** Constraint: tpi31-tpi32
*Tie, name=tpi31-tpi32, adjust=yes, no rotation
PickedSet1416-CNS, PickedSurf1415
** Constraint: tpi32-tpi33
*Tie, name=tpi32-tpi33, adjust=yes, no rotation
PickedSet1418-CNS, PickedSurf1417
** Constraint: tpi33-tpi34
*Tie, name=tpi33-tpi34, adjust=yes, no rotation
PickedSet1420-CNS, PickedSurf1419
** Constraint: tpi34-tpi35
*Tie, name=tpi34-tpi35, adjust=yes, no rotation
PickedSet1422-CNS, PickedSurf1421
** Constraint: tpi35-tpi36
*Tie, name=tpi35-tpi36, adjust=yes, no rotation
PickedSet1424-CNS, PickedSurf1423
** Constraint: tpi36-tpi37
*Tie, name=tpi36-tpi37, adjust=yes, no rotation
PickedSet1426-CNS, PickedSurf1425
** Constraint: tpi37-tpi38
*Tie, name=tpi37-tpi38, adjust=yes, no rotation
PickedSet1428-CNS, PickedSurf1427
** Constraint: tpi38-tpy12
*Tie, name=tpi38-tpy12, adjust=yes, no rotation
PickedSet1430-CNS, PickedSurf1429
** Constraint: tpi4-tpi5
*Tie, name=tpi4-tpi5, adjust=yes, no rotation
PickedSet1362-CNS, PickedSurf1361
** Constraint: tpi5-tpi6
*Tie, name=tpi5-tpi6, adjust=yes, no rotation

119
APPENDIX B. ABAQUS INPUT FILES

PickedSet1364-CNS, PickedSurf1363
** Constraint: tpi6-tpi7
*Tie, name=tpi6-tpi7, adjust=yes, no rotation
PickedSet1366-CNS, PickedSurf1365
** Constraint: tpi7-tpi8
*Tie, name=tpi7-tpi8, adjust=yes, no rotation
PickedSet1368-CNS, PickedSurf1367
** Constraint: tpi8-tpi9
*Tie, name=tpi8-tpi9, adjust=yes, no rotation
PickedSet1370-CNS, PickedSurf1369
** Constraint: tpi9-tpi10
*Tie, name=tpi9-tpi10, adjust=yes, no rotation
PickedSet1372-CNS, PickedSurf1371
** Constraint: tpy10-tpy11
*Tie, name=tpy10-tpy11, adjust=yes, no rotation
PickedSet1352-CNS, PickedSurf1351
** Constraint: tpy11-tpi1
*Tie, name=tpy11-tpi1, adjust=yes, no rotation
PickedSet1354-CNS, PickedSurf1353
** Constraint: tpy12-tpy13
*Tie, name=tpy12-tpy13, adjust=yes, no rotation
PickedSet1432-CNS, PickedSurf1431
** Constraint: tpy13-tpy14
*Tie, name=tpy13-tpy14, adjust=yes, no rotation
PickedSet1434-CNS, PickedSurf1433
** Constraint: tpy14-tpy15
*Tie, name=tpy14-tpy15, adjust=yes, no rotation
PickedSet1436-CNS, PickedSurf1435
** Constraint: tpy15-tpy16
*Tie, name=tpy15-tpy16, adjust=yes, no rotation
PickedSet1438-CNS, PickedSurf1437
** Constraint: tpy16-tpy17
*Tie, name=tpy16-tpy17, adjust=yes, no rotation
PickedSet1440-CNS, PickedSurf1439
** Constraint: tpy17-tpy18
*Tie, name=tpy17-tpy18, adjust=yes, no rotation
PickedSet1442-CNS, PickedSurf1441
** Constraint: tpy18-tpy19
*Tie, name=tpy18-tpy19, adjust=yes, no rotation
PickedSet1444-CNS, PickedSurf1443
** Constraint: tpy19-tpy20
*Tie, name=tpy19-tpy20, adjust=yes, no rotation
PickedSet1446-CNS, PickedSurf1445
** Constraint: tpy2-tpy3
*Tie, name=tpy2-tpy3, adjust=yes, no rotation
PickedSet1336-CNS, PickedSurf1335

120
 B.2. TORSIONAL STABILITY

** Constraint: tpy20-tpy21
*Tie, name=tpy20-tpy21, adjust=yes, no rotation
PickedSet1448-CNS, PickedSurf1447
** Constraint: tpy3-tpy4
*Tie, name=tpy3-tpy4, adjust=yes, no rotation
PickedSet1338-CNS, PickedSurf1337
** Constraint: tpy4-tpy5
*Tie, name=tpy4-tpy5, adjust=yes, no rotation
PickedSet1340-CNS, PickedSurf1339
** Constraint: tpy5-tpy6
*Tie, name=tpy5-tpy6, adjust=yes, no rotation
PickedSet1342-CNS, PickedSurf1341
** Constraint: tpy6-tpy7
*Tie, name=tpy6-tpy7, adjust=yes, no rotation
PickedSet1344-CNS, PickedSurf1343
** Constraint: tpy7-tpy8
*Tie, name=tpy7-tpy8, adjust=yes, no rotation
PickedSet1346-CNS, PickedSurf1345
** Constraint: tpy8-tpy9
*Tie, name=tpy8-tpy9, adjust=yes, no rotation
PickedSet1348-CNS, PickedSurf1347
** Constraint: tpy9-tpy10
*Tie, name=tpy9-tpy10, adjust=yes, no rotation
PickedSet1350-CNS, PickedSurf1349
*End Assembly

121
Royal Institute of Technology
Department of Mechanics
SE-100 44 Stockholm, Sweden
http://www.mech.kth.se