CSB Analysis Tool

Dissertation ETH No.
15603
A Software Tool for the Analysis, Design

and Monitoring of Cable-Stayed Bridges
A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH
for the
degree
of
Doctor of Sciences ETH
presented by
Pietro Pedrozzi
Dipl. Bau-Ing.
ETH
born 21 November 1975

citizen of
accepted
on
Dr.
sc.
sc.
TI
the recommendation of
Prof. Dr. Edoardo

Prof. Dr.
Pregassona
Anderheggen,
techn. Mario
techn. Mike
Fontana,
Schlaich,
2004
examiner
co-examiner
co-examiner
11
Abstract
Of the
newly-built bridges, cable-stayed bridges are today very common worldwide for
spans ranging between 200 and 900 meters. Being mostly built by the cantilever method,
cable-stayed bridges have to be analyzed separately in every construction stage, taking
In addition, the forces in the cable stays have to be
into account many load cases.
determined and possibly changed during construction so as to obtain the desired deck
and mast deflections during service. Some non-linear and execution-dependent effects
such as concrete creep, shrinkage and locked-in bending in composite steel-concrete
decks
have to be considered in the case of longer spans.
Conventional structural analysis software cannot automatically handle these specific
problems, which today in many cases require the program user to manually process data.
The goal of this project is the development of a specially tailored program for the struc
tural analysis of cable-stayed bridges, which automatically handles the aforementioned
problems. The program has been designed to be used in all stages of planning and
construction: preliminary design, detail planning, construction process and retrofitting.
The program BRIDE is characterized by some newly introduced concepts allowing one
to take the aforementioned non-linear effects into account: the chronologically ordered
list of model objects accurately representing the complete construction process, which
has to be defined in the input by the user, and the stage-by-stage iteration in which all
construction stages are simulated in chronological order.
The program's functionality can be summarized as follows: the user first writes a text
input file according to the program's input syntax, which allows programming language
like constructs such as variable declarations, for-loops, if-tests, and expression evalua
-
tions for
input parameterization. The
(this
can
file
culations
user
then opens the input file from the program
opened while it is being written to check its correctness). All cal

be performed interactively through the user graphic interface. The
also be
can
then
model and the results for any desired construction stage appear
dimensional
on
the
screen
in three-
Snapshots with both graphics and text showing model

can
a highly portable Html-document. Automatic dimensioning
of cables and the calculation of required pre-camber in the deck and mast segments and
of post-tensioning forces in cables are possible. As such calculations change the model
itself, the updated model can be stored in a new input file reflecting the content of the
and results
graphics
or
text form.
be stored in
changes made to it. All the way from preliminary design to the
construction phase the bridge designer works on the same input file and, as the planning
gets more detailed, refines it either manually or automatically.
As discussed in the conclusions at the end of this thesis the program BRIDE, at least
original
one
and the
in its
present form, is
not to be viewed
main scope of this research work
non-trivial
IV
practical problem.
was
to
as
commercial software
study original approaches
product.
to
an
In
fact,
the
important and
Zusammenfassung
Unter den
neu
gebauten
Brcken sind
weiten zwischen 200 und 900 Metern.
gebaut
Schrgseilbrcken
weltweit sehr beliebt fr
Spann
Solche Brcken werden meistens in Frei-Vorbau
und mssen deshalb in allen Bauzustnden unter vielen Lastfllen untersucht
den. Zudem mssen die
Kabelvorspannkrfte
wer
ermittelt und eventuell whrend des Baus
Lage der Decke und des Mastes im

Gebrauchszustand zu erhalten. Fr lngere Spannweiten sollen einige nichtlineare- und
ausfhrungsabhngige Effekte wie zum Beispiel Kriechen, Schwinden und eingefrorene
gendert werden,
um
die erwnschte verschobene
Anfangsverschiebungen in Verbunddecken bercksichtigt werden.

Konventionelle Baustatikprogramme knnen solche spezifischen Probleme nicht be
handeln, und sie zwingen heutzutage den Benutzer die Daten manuell zu manipulieren.
Das Ziel dieses Projektes ist die Entwicklung eines Programms, das speziell fr die stati
sche Analyse von Schrgseilbrcken konzipiert ist und welches die erwhnten Probleme
automatisch behandeln kann. Das Programm wurde entworfen um in jeder Planungs
und Bauphase eingesetzt zu werden: Entwurf, Detailplanung, Bau und Sanierung.
Das Programm BRIDE ist gekennzeichnet durch einige neu eingefhrte Konzepte
welche es erlauben, die erwhnten nichtlineare Effekte zu bercksichtigen: die chrono
logisch geordnete Liste von Modellobjekte, welche vom Benutzer in der Eingabedatei
definiert wird, und die schrittweise Iteration ber die Bauzustnde, die smtliche Bau
zustnde in chronologischer Ordnung simuliert.
Die Arbeitsweise des Programms kann wie folgt zusammengefasst werden: der Be
nutzer schreibt zuerst eine Textdatei gemss der Eingabesyntax des Programms, welche
programmiersprachenhnliche Konstrukte wie Variablendeklarationen, for-Schleifen, ifPrfungen und Ausdruckevaluationen fr die Parametrisierung der Eingabe zulsst.
Danach ffnet er die Eingabedatei vom Programm aus (die Eingabedatei kann auch
geffnet werden, wenn sie noch nicht fertiggeschrieben wurde, um ihre Korrektheit zu
berprfen). Smtliche Berechnungen knnen interaktiv ber die graphische Oberflche
ausgefhrt werden. Das Modell und seine Resultate erscheinen als dreidimensionales Bild
oder als Text fr jeden beliebigen Bauzustand und Lastfall auf dem Bildschirm. Schnapp
schsse der graphischen oder der Textausgabe knnen mit dem Modell und seinen Re
sultaten als portables Html-Dokument gespeichert werden. Eine automatische Kabelbe
messung und die Ermittlung smtlicher bentigten Vorspannkrfte und berhhungen
knnen ausgefhrt werden. Sofern solche Berechnungen das Modell ndern, kann das
aktualisierte Modell als eine neue Eingabedatei abgespeichert werden, welche der ur
sprnglichen Eingabedatei entspricht plus die vom Programm ermittelten nderungen.
Whrend des ganzen Prozesses
vom
Entwurf bis
zum
Bau arbeitet der Benutzer immer
mit der
gleichen Eingabedatei und,
wenn
die
Planung
detaillierter
wird,
verfeinert
er
sie
entweder manuell oder automatisch.

Wie in den
Schlussfolgerungen dieser Dissertation erwhnt, soll das Programm BRI

DE, wenigstens in seiner jetzigen Form, nicht als kommerzielles Produkt angesehen wer
den.
Das Hauptziel dieser Forschungsarbeit war nmlich, originelle Anstze fr ein
wichtiges und nicht triviales praktisches Problem zu erforschen.
vi
Contents
1.
2.
1.1.
The
1.2.
Use of standard structural
1.3.
Technical progress of computers
1.4.
The program BRIDE and its
special
1.5.
User friendliness: BRIDE
commercial software
1.6.
Structure of this dissertation
analysis
vs.
software
features
3
4
4
description through
displacements
Conditional loads
2.3.
Chronological objects
2.4.
The
its local stiffness matrix and
set of initial
list:
composite
full
an
description
automatic
of the erection process
analysis
...
of each construction
15
beam elements
kj0
16
19
displacements
Cast beams to simulate locked-in
displacements
in concrete cast
on
an
elastic formwork without shear connection

4.2.
Composite
an
5.
beams to simulate locked-in
19
displacements
in concrete cast
on
elastic formwork with shear connection
20
21
Creep
creep behaviour of concrete
5.1.
Hereditary
5.2.
Approximation
in the
computation of
21
creep in structures with
varying
strain
6.
Concrete
6.1.
11
12
The excentric local stiffness matrices kca and
Locked-in
10
stage-by-stage iteration:
stage in chronological order
Steel-concrete
4.1.
Element
2.2.
3.1.
4.
cable-stayed bridge design problem
A tour of the program BRIDE

2.1.
3.
Introduction
25
27
shrinkage
Initial
displacements
ash
needed to take concrete
shrinkage
into account
28
vn
Contents
7.
Conditional loads
29
7.1.
The influence matrix
7.2.
The need to
algorithm
perform the stage-by-stage
31
iteration
repeatedly
to find the
conditional loads
7.3.
Convergence of the influence matrix algorithm

7.3.1.
Creep and shrinkage
Locked-in displacements
7.3.2.
8.
Conclusions
Cable
regarding
9.2.
an
input
33
34
the conditional loads
35
37
37
file for the program BRIDE
Analysis strategy
Monitoring the bridge
by keeping
the input file
up-to-date
....
42
49
A. How to define
numerical model for the program BRIDE
53
Syntax
for the definition of model
A. 1.1.
55
A. 1.3.
Stage object
Restart object
Remove object
A.1.4.
Nodes
56
A. 1.5.
Elements
57
A. 1.6.
Sections
58
A.l.7.
Materials
59
A.1.8.
Joints
60
A. 1.9.
Supports
61
A. 1.2.
A.1.10. Load
A.1.11. The
A.2.
39
40
erection
10. Conclusions
A.l.
33
33
dimensioning
dimensioning algorithm
Example of
9.1.
vni
Automatic cable
8.1.
9.
with non-linear models
Second order effects and cable sag
7.3.3.
7.4.
32
cases
objects
55
55
56
and load combinations
load-history
load
case
62
63
A.l.12.Node loads
63
A.1.13. Element loads
64
A. 1.14. Conditions
65
A.1.15. Cable
67
dimensioning object
A.l.16. Notes
67
Text-preprocessor
67
A.2.1.
Variables
67
A.2.2.
Expressions
68
A.2.3.
A.2.4.
[expression] integer
For-loop
A.2.5.
If.
A.2.6.
Include statement
.Else.
evaluation
construct
69
69
70
71
Contents
B.
A.2.7.
Comments
71
A.2.8.
Finish
71
Graphie
user
interface of the program BRIDE

some standard GUI components
73
B.l.
Definition of
74
B.2.
Menu bar
75
B.3.
Model state control
B.4.
Elements
common
panel
75
to all three
analysis panels
of the 3D output definition
76
B.6.
Specific analysis panel components

Static analysis panel
B.5.2. Buckling analysis panel
B.5.3. Modal analysis panel
Text panel
B.7.
Html tool-bar
80
B.8.
Syntax highlighting
80
B.5.
panels
B.5.1.
C. Information
D.
Input
files discussed in
D.I. Main
D.2.
technology
tools used to
78
78
79
79
80
develop
the program BRIDE
chapter 9
83
85
input file
Input file included (twice)
85
in the main
input file
88
IX
Contents
1.
Introduction
1.1.
The
In
cable-stayed bridge design problem

the deck is
cable-stayed bridges
which
are
fixed to the top
the cables
on one
supported
at
less
from the deck
along a mast protruding

self-anchored, i.e. the normal
of a mast is compensated by the
or
are
side
regular distances by cables

plane. In most cases
force introduced in the deck by
more or
normal force introduced
on
the
other side.
1.1.: A
Figure
The main
they can be built with very large

spans (today with a central span of up to 900 meters) by free cantilevering (see figure 1.2),
provide a large stiffness, need little material and can look quite elegant. Despite these
advantages, cable-stayed bridges only became more widespread in the nineteen-fifties
because it was impossible to analyze them with a reasonable effort and a satisfactory
accuracy using the manual methods of pre-computer days.
The reason why the design and erection of cable-stayed bridges is so computationally
intensive is that they are statically indeterminate to a degree which is approximatively
as large as the number of cables and, at the same time, because their displacements
and section forces are adjustable through the post-tensioning of each single cable. In
addition, long span cable-stayed bridges are subject to non-linear phenomena such as
advantages
of
typical cable-stayed bridge.
are
that
cable sag and concrete creep.

The
goal
of
adjusting
the
post-tensioning force
cambered mast and deck segments is to have
planned geometry.
bers
can
Even
in each cable and of
mounting
pre-
finished structure which reflects the
though cable-stayed bridges are globally very stiff, single mem

produce large displacements, especially during construc
be quite flexible and
tion.
1. Introduction
the cables
are
post-tensioned
from inside the mast
temporary supports
Figure
1.2.:
Bridge
the
Once statical
erection
cover
by
free
cantilevering (analogous
analysis
on
for several construction stages and for the final service state
might
still be needed
with measurements in situ in order to correct
erection in combination
during
alignment
and to
errors
keep
the rather
behaviour of concrete and supports under close control.
Use of standard structural
1.2.
bridge
page).
has been carried out, computations
unpredictable
to the Uddevalla
analysis
software
Cable-stayed bridges can, in principle, be successfully analyzed with standard structural

analysis software. This however is not straightforward. Standard software, in general,
cannot automatically determine the post-tensioning forces needed in the cable-stays or
the pre-camber values for the mast and deck segments. In many cases these still have
to be found by the bridge designer by means of time-consuming and error-prone manual
computation. Here are a few possible strategies:
Insertion of fictitious
mast is not
hinges in the mast so

subjected to bending moments.
to find cable forces for which the
as
Use of very high artificial stiffnesses for the normal

ments in the deck and the pylons.
to avoid axial
force,
displace
Use of supports instead of cables to obtain the post-tensioning forces from the
support reactions.
every cable and
mulation of
as
many
allowing
bridge's displaced shape caused by

unitary pre-camber value for every deck
Determination of the
an
equation system (to be solved
compatibility
one
conditions
to find the
actually
as
in
an
and mast segment. For
external
unitary shortenings
needed cable
unitary shortening of
shortenings
program) representing
and
and
pre-camber values,
pre-camber
values
1.3.
(this
is the method
adopted
These computations, when

are
1.3.
our
program BRIDE but
performed
either
time-consuming and inhibit
very
implemented
in
totally
way).
automatic
grams,
in
manually or using
the investigation of
additional external pro

several
project
variants.
The
performance of today's computer technology is impressive: not only the hardware

has become faster, also the available programming techniques have progressed consider
ably, allowing one to describe complex systems in a much more compact and expressive
solution
algorithms can be developed which, relying on the computa

tional power available on everybody's desk, model the physical phenomena in a more
natural and accurate way than algorithms developed when computational power was
Nowadays
way.
limited.
This opens up
lowing
new
research
paths
which
the researcher to be innovative
structural
1.4.
were
in
even
technically not conceivable before, al

seemingly well-explored field such as
an
analysis.
The program BRIDE and its
The difficulties encountered in the
analysis
of
special features
with standard software
and the
opportunity offered by the technical progress in computer hard- and software ap

peared
good reasons to start the development of a computer program specially
to be two
tailored to the requirements of

The fact that
are now
growing
provide
assess
for further research in

programs
The program has been called

tradition of
designer
institute to
our
in the
following
model all kinds of

account in
of cable-
give
or
"BRIDE",
as
an
seed
on
technically newly
development for
stayed bridges
this topic: this
accessible
new
fields,
commercial
BRIdge DEsigner (it is a

names), and its goal is to help the
acronym for
programs romantic
tasks:
cable-stayed bridges, taking
all main non-linear effects into
consistent way,
allow the
easily
find all needed
analysis
of each construction stage,
post-tensioning forces for the cables and all pre-camber values for
the mast and deck segments in
analysis
available shows that the time is ripe for fundamental research
knowledge base
commercially available
products.
can
cable-stayed bridge designers.
number of programs for the
correct deviations in the
fully
automatic way and
geometry measured
in situ
by changing
the cable forces.
1. Introduction
Five main ideas determined the
of the program BRIDE and
design
it from
distinguish
others programs:
As the
displacements
stayed bridges
are
finite element
plus
set of "initial"
can
Post-tensioning
by the program
tudes
are
found
The erection of
described
fully
displacements.
be added to
corresponding segment
forces
can
can
or
or
In order to take
since cable-
structure modelled
element does not
physics
depend
on
the
of the modelled segment of
just
as
the
removed in real life.
values which have to be found
as
initial element
automatically
displacements whose magni
conditions to be fulfilled.
cable-stayed bridge
chronological
in
an
the
is
an
way to describe it with accuracy is with
system
acceptable
the element's local stiffness matrix
by
Such
be added
be formulated
through
a
which is
removed from the numerical model
pre-camber
or
small,
the behaviour of every segment of
be
can
and, fully representing
structure
the structure,
assumed to be
stiff,
very
with
global
are
evolutionary
a
list of the
process and the
simplest
changes occurring in the
order.
displacements locked in concrete and creep into account it is

displacements of the structure in the stages prior to the
necessary to know the
stage being analyzed.
The program BRIDE should be

the
1.5.
bridge
construction
sufficiently user-friendly to
industry and for teaching purposes.
User friendliness:
With programs
to concentrate
interface.
developed
For this
the
reason
the
applicable
both in
commercial software
vs.
research context in
implementation of
project, however, the goal
more on
In this
friendliness.
in
BRIDE
be
computational
mechanics it is usual
computational
core
was
set to achieve
than of the
user
acceptable userdevelopment of a
an
much effort has been invested in the
graphic user interface with a three-dimensional representation of the numerical models.

During the design of the program BRIDE a rigorous transparency has been actively
attempted in the way the program generates the numeric models from the user speci
fications and in the way computations are performed. This transparency is a kind of
user-friendliness which many commercial programs do not offer and which, apart from
bridge designers, can be of pedagogical advantage for both teachers and students.
1.6.
Chapter
explains
gives
an
overview of the main features of the program BRIDE.
the composite beam element.
tributions to the initial
displacements
Chapters 4,
S-miUai
Chapter
5 and 6 deal in detail with the
con
discussed in the section 2.1 due to locked-in
1.6.
displacements (chapter 4),
creep
(chapter 5)
and
shrinkage (chapter 6). Chapter
2.2). Chap
explains the way cable stays are automatically dimensioned. In chapter 9 a practical
example is given. Chapter 10 presents the conclusions. The appendices contain the user
manual of the program BRIDE: in appendix A the input syntax is defined, in appendix B
the graphic user interface is explained, appendix C presents the information technology
with the computation of the intensities of conditional loads
(see
7 deals
subsection
ter 8
tools used to
implement
for the model shown in
the program BRIDE and in
chapter
are
appendix
listed and commented
D the
complete input
files
on.
1. Introduction
2.
The
was
A tour of the program BRIDE

approach adopted in the development of the algorithms for the program BRIDE
to keep as close as possible to the physics of cable-stayed bridge construction, so as
to avoid inconsistencies.
This lead to
"object-oriented" solution in which the model

consists of a list of data objects representing the different components of the real bridge
ordered with the same chronological order with which their physical counterparts are
added to the real bridge. All needed computations are performed by the program BRIDE
This chapter gives an overview of
on this chronologically consistent numerical model.
an
them.
2.1.
Element
description through
matrix and
set
of initial
its local stiffness
displacements
Cable-stayed bridges are modelled in the program BRIDE as space frames where 6
displacement parameters (3 displacements and 3 rotations) are introduced in each node.
The geometry of the model is described through the Cartesian coordinates of its nodes
and through the element's incidence nodes. In addition, every element requires section
and material properties. Once the coordinates of the incidence nodes, the material and
the section of the element are defined, its elastic response can be formulated by means
of
If
local 12x12 elastic stiffness matrix.

we
were
to consider
(see figure 2.1),
take
an
existing bridge already modelled
concrete
cutter,
cut out
element used for the
Figure
a
single
2.1.:
Step
1. Choose
finite element
one
in
segment of the
structure modelled with
thought experiment
element in the numerical model of
(see figure 2.2)
simulation program
an
existing bridge,
and determine its geometry in the unloaded state,
2. A tour of the program BRIDE
we
would notice that its geometry does not
element.
correspond
to the
geometry of the finite
fact, displacements locked-in during the hardening of the concrete, creep,

shrinkage and the pre-camber would have slightly changed the shape of the cut out and
thus unloaded segment (see figure 2.3).
Figure
In
2.2.:
Step
2. Cut out the structural
The vector amrfmi of size
node) containing
geometry
can
12(3
segment corresponding
translation and 3 rotation components for each incidence
the difference between the cut out
be formulated
as
segment's
<*h
"-c
as a
(2.1)
ash
where a/,, is the contribution of the locked-in concrete

considered
and the finite element's
follows:
"initial
and as/j is that of
to the chosen element.
displacements,
acr that of creep
shrinkage (the pre-camber

load", see section 2.2).
is not taken into account here because it is
"conditional
structure
segment
cut out
finite element used to model the element cut out
Figure
2.3.:
Step
3. The
geometry of the
cut out structural
the finite element used to model it differ

One way to model the cut out segment
subjecting
the finite element to
set of
by
segment and the geometry of
the initial
displacements
amrfmi.
correctly is to reproduce its displacements by

initial self-equilibrating forces f (see figure 2.4)
2.1.
Element
producing
description through
the initial
its local stiffness matrix and
displacements
element local stiffness matrix k
by
2.4.:
Step
4.
assumes
The finite element

the
same
be
f,
found
displacements
by multiplying
the
as
to the initial
subjected
as
(2.2)
ainitial
self-equilibrating
force f
the cut out structural segment.
for any kind of element
load, have to be assembled

model, its local stiffness
load vector every time the element is part of the
global
being
matrix k
to
geometry
The components of the vector

in the
are
set of initial
S-mitiai'-
Figure
The forces f
amttm|.
assembled in the
global
stiffness matrix.
Being initial,
i.e.
corresponding
deformed but stress-free state, the contribution from f to the section forces has to
ignored during post-processing,
i.e.
the section forces induced
have to be subtracted from the section forces found from the nodal
by f in the element
displacements of the
global solution. The subtraction does not affect the equilibrium of the structure because
f is self-equilibrating: this is a property of local stiffness matrices which, being singular,
always deliver a self-equilibrating forces group if multiplied by any nodal displacement
vector.
cast concrete A
ca'
It
ca
Iz
ca'
G
ca
ca'
centra id of the cast concrete
eca
Dz
centroid of the combined section
efo
centroid of the formwork
formwork
Figure
Initial
Afo, ltfo, lzfo, Efo, Gf,
2.5.: Section idealization used for cast and
displacements
composite beam elements.
have to be taken into account in
composite beam element specially developed for the
slightly different way for

program BRIDE, which allows
a
the
one
to take the bond between concrete and its
underlying steel profile into account (see

explained in chapter 3, takes bonding into
figure 2.5). The composite beam element, as

account by merging into a single combined section the section of the cast concrete and
the section of the steel profile, which is modelled by a so-called formwork beam (see
4.2).
section
composite beam elements f
For
kco
a-
where kco is the stiffness matrix

the
displacements
kca and
kca
(a
computed
a)
as
follows:
k/o (a
a)
(2.3)
a and q{ are
during the casting stage,
with the combined section,
and the nodal forces of the formwork beam
the stiffness matrices of the cast concrete part and of the formwork
k/0
beam, respectively,
are
displacements
is found
with
no
bond between them,
a^, a^, eJc
and
&
of the formwork beam and of the cast concrete beam
have in addition initial
due to creep and
are
(it
is
the initial
justified
to
for the formwork since
displacements
shrinkage
might be the case in pre-cast concrete). All these forces, stiffness matrices and
displacements are expressed relative to the centroid of the combined section. The element
is then excentrically connected to its nodes.
this
The fact of not
taking bonding into account for kca and kj0 is a obvious simplification
acceptable because the self-equilibrating load f is applied to the undisplaced
composite element (whose stiffness takes bonding into account) hence differences between
the initial displacements of the cast concrete a^ + a^ and those of the formwork a^ + a^
only lead to differences in the section forces but not to bond-producing discontinuities in
which is
the total strain

must
elastic and the initial strain summed
together).
The elastic strain
compensate any initial strain because in the finite element model the element is
forced into its
2.2.
(the
undisplaced shape
for the formulation of
global equilibrium
conditions.
Conditional loads
The so-called conditional loads represent the
pre-camber
by
of mast
post-tensioning of the cable-stays and the

and deck segments. Their intensities are automatically computed
the program BRIDE.

Both conditional loads and
regular
loads
are
model
objects
to be inserted in the
chronological objects
only
regular loads
(see section 2.3).
the load intensity is explicitly defined while in conditional loads it is defined through a
condition to be fulfilled, and that conditional loads always belong to the standard load
case load-history (see section 2.4). Their required intensities are found automatically by
the program BRIDE by fulfilling appropriate conditions specified by the program user
(e.g. "post-tension a cable with a force such that its anchoring point on the deck exhibits
no vertical displacement in the final construction stage, taking into account the dead
load and all other conditional loads", see figure 2.6).
list
10
The
differences is that in
2.3.
list:
full
description
of the erection process
*-4Figure
2.6.: Deflected
shapes of the same model without and with the automatically

computed pre-camber of the deck and mast segments and the post-tensioning
of the cables.
2.3.
list:
full
description of
the
erection process
After
attempts it became clear that the
model of
been introduced to allow the program

he feels that
object.
model
enough
objects
The stage
objects
user
list of data
to define
define the states of the
objects representing
objects
bridge
To
stage whenever
where
analysis
by
is
possible.
the user, it constructs
it.
the list of data
stage, the (object-oriented)

from the
program
sends to each
through
objects starting
beginning,
1
signal to assemble itself to the actual model and stops when it encoun
data object corresponding to the desired construction stage (see figure 2.7).
object
ters the
construction
list defined
To build the numerical model of any construction

BRIDE goes
data
a new
have been added to the model since the last stage
Once the program BRIDE has read the

a
most convenient way to
specify
the
evolving
cable-stayed bridge during construction is a list of all the model's objects
(nodes, elements, supports, forces, ...) ordered in the same chronological order in which
their real counterparts are added to the bridge being constructed. The object "stage" has
some
the
analyze
the model for
given load
case
the program BRIDE
proceeds
in
an
analo
assembling of the finite element model: it goes through the chrono

logical objects
beginning, takes into account the load objects belonging to
the chosen load case and stops when the stage object corresponding to the stage being
analyzed is encountered (see figure 2.8).
gous way
as
for the
list from the
The fact of
oriented
viewing units of data as active "objects" rather than passive units distinguishes the object
programming paradigm supported by programming languages like Java and C++ from the
traditional
imperative paradigm of programming languages such
the units of data
are
as
passive and manipulated through procedures
Fortran,
Pascal and
C,
in which
not included in their definition.
Start from the beginning of

the chronological
objects list
Go to the next
object
yes
Is this an object representing

the stage to De analyzed?
The assembling of the finite element model

for the stage to be analyzed is finished
The load case still has to be chosen
and assembled
Is this a model object to be assembled

the finite element model?
on
yes
Assemble it
Figure
2.4.
The
on
the actual model
2.7.: Flow
diagram
of the
stage-by-stage
assembling
iteration:
of each construction stage in
of
an
construction
stage.
automatic
chronological
analysis
order
To take creep and locked-in concrete
displacements in a given construction stage into

account it is necessary to know the initial displacements acr and a/j for every element
(see section 2.1), which, in turn, depend on the displaced shapes of the bridge in all
previous
construction
stages. Since it is important
of any construction stage
to allow the
user
to view the results
immediately, without having to wait for the analysis of all

displacements of every construction stage due to all relevant
loads are stored in the corresponding stage data object.
Such nodal displacements are found during what is called stage-by-stage iteration, in
which every construction stage is automatically assembled and analyzed in chronological
order. Figure 2.9 shows the flow diagram of such an iteration.
To denote the load objects which should be taken into account during the stage-bystage iteration the standard load case called load-history has been introduced. The
conditional loads are automatically assigned to this standard load case because they are
to be computed during the stage-by-stage iteration. The user, however, is still free to
assign them to other load cases or load combinations.
former stages, the nodal
12
2.4.
The
stage-by-stage
iteration:
an
automatic
analysis
of each construction stage in
chronological

the chronologicafobjects list
Go to the next
object
Is this a stage object representing

the stage being analyzed?
Is this
case or
Analyze the finite element model
yes
-
according to the options
specifiecTinteractively by the
user
load object belonging to the load

load combination chosen?
yes
Take it into account
Figure
in
the
analysis
2.8.: Flow
into
diagram of the selection of

account during the analysis.
the load
objects
which have to be taken

the chronological
objects list
Are there
more
objects
in
One stage-by-stage iteration finished,

the user can look at any construction stage with any
load case and study the result or iterate again
the list?
yes
][
Go to the next
Is this
object
"construction
stage" object?
yes
Assemble the finite element model for the

construction stage represented by the
object just foundand for the load case
"load_history"
Find for every element all initial
displacements
Analyze the actual model to find the

conditional loads having a condition to be
fulfilled in the actual stage
Analyze the model again taking the
newly found conditional loads into account
Store the nodal
displacements
stage object just found
Figure
in
the
2.9.: Flow
diagram
of
one
stage-by-stage
iteration.
13
order
14
3.
Steel-concrete
beam
composite
elements
beam elements have been
Composite
model steel-concrete composite decks with

locked-in
displacement
decks than for
generally
is
plain
shear connection assumed to be
This is
into account.
generally
(see chapter 4),
cast decks
for the program BRIDE to
specially developed
more
rigid taking
important for composite
because the stiffness of the steel
lower than the stiffness of the temporary formworks used for the
profiles
plain cast
segments.
The main difference between
the steel
work, i.g.
profile,
is
cast deck and
never
composite deck
bonded to it. A combined model for composite decks has been

the assumption that for the force range
on
is that here the form-
removed from the cast section but remains
usually
allowed for
adopted,
rigidly
which is based
the behaviour of
bridges
composite deck segments remains elastic.
When, during stage-by-stage iteration, a stage is selected in which a new composite

beam is defined, its formwork beam (the beam object used to simulate the steel profile,
see chapter 4) is removed from the model and the properties of the combined section
resulting
object)
and
from the
merging of the
cast concrete section
and the formwork beam section
assigned
to the
are
automatically
(AE)co
AfoEfo
AcaEca
efo
eca
(lyE)co
(lyjo
(IZE)co
ca
in the
by
composite beam
the program BRIDE
are
calculated
(3.1)
Dz
(3-2)
-Dz- efo
_
Af0efo)Ef0
(IytCa
hjoEfo
Iz,caEca
follows:
AcaEca
ItJoGfo
as
HE-
AfoEfo -Dz
{AE)co
(ItG)co
section,
found
composite beam element.
The section properties for the combined section
where the
(defined
(3.4)
It,caGca
+
(3.3)
Acaeca)Eca
(3.5)
(3.6)
denotes the combined section, fo denotes the formwork beam

denotes the cast concrete section, A is the section's area, e is the excentricity
subscript
co
of the centroid of either cast concrete and formwork section from the centroid of the
combined section, E the
Young-modulus,
G the
shear-modulus, It
the the torsional
15
3. Steel-concrete
rigidity, Iy
figure 2.5).
and Iz
composite beam elements

the moments of inertia with respect to the local y and
are
The normal stress distribution
by superposing
four components
the normal stresses
during
along the local

(see figure 3.1):
the
axes
(see
axis in the combined beam is found
casting stage, where the composite beam
section is
unstressed because the concrete is still wet and the formwork beam has its elastic
normal stress
the elastic normal stresses of the combined beam from the actual
global
solution
and from the actual local load
the stresses obtained

would have if it
and to the local
by subtracting the elastic normal stresses the combined beam

was subjected to the hardening displacements a^ (see chapter 4)
load acting on the formwork beam during the casting stage
the stresses due to creep and
shrinkage.
normal stress
the
normal stress found

with the stiffness of
the combined section,
the nodal displacements
from the actual global
solution and the
actual local load
during
casting stage
Figure
3.1.:
Superposition
normal stress from

normal stress found
with the stiffness of
creep and shrinkage
the combined section,
the hardening
displacements and
the local load acting
on the formwork beam
during the casting stage
of the different normal stress components in
concrete-steel
composite beam.
3.1.
The excentric local stiffness matrices
As mentioned in section 2.1 to compute the

initial
displacements
into account for
the local stiffness matrix kca for
self-equilibrating
hypothetical
beam
and the stiffness matrix of the formwork element
16
i.e.
excentrically (see
eq.
force f needed to take the
composite beam elements it is necessary to have
between formwork and cast concrete and
centroid,
km and kf0
2.3).
having
kj0,
expressed
the cast concrete
both
as
if there
was
as
section
no
bond
relative to the combined section's
The excentric local stiffness matrices kca and
3.1.
kca
be derived
can
as
follows:
assuming
connection between the formwork beam
rigid
and the cast concrete, the

can
be
expressed
as a
displacements aca in the centroid of

function of the displacement parameters
the cast concrete section

a
of the combined beam:
(3.7)
a
or
the nodal forces from the centroid of the cast concrete section qca
to q, the
same
forces set but
can
be transformed
to the combined beam:
applied
r1
^ca
(3.8)
nca
With 3
displacement parameters in global X-, Y-, and Z-directions and 3

respect to the corresponding coordinate axis introduced (in that order)
the 12x12 transformation matrix Cca is defined
T is the
where
in the
following
kj,
as
rotations with
at each
node,
follows:
6x6 transformation matrix:

1
dz
-dy
dz
dx
dy
dx
dx, dy and dz are the components

global coordinates system.
Since the local stiffness matrix
kca)Ceratc
of the
of
excentricity
beam
having
vector
[0,0, eca] expressed
the cast concrete section is
known and since:

q
by inserting
resulting
section
V""^/
"
ca
**-ca,centric
^^ca
following
"
relation
applies:
(3.10)
stiffness matrix is the excentric stiffness matrix for the cast concrete
kca:
K-ca
kf0
'
eq. 3.7 in eq. 3.9 and eq. 3.9 in eq. 3.8 the
The
**-ca,centric
can
be derived in
^ca
'
K-ca,centric
completely analogous
'
^ca
V"
way.
17
3. Steel-concrete
18
composite beam elements
4.
Locked-in
Any
displacements
deformation of the formwork remains "stored" in the cast concrete
In contrast to the usual case, where the formwork is stiff
this has to be taken into account in the
deflections,
those with
cially
construction
an
the
lead to
systematic accumulation of
displacements
two different
causing
types of elements have been imple
shear connection with its formwork and
concrete
Cast beams to simulate locked-in
4.1.
error
a
having no
having a rigid shear connection with a steel profile which
role of non-removable formwork (see chapter 3). In the program BRIDE the
is modelled by a regular beam element.
formwork
can
cast beam for concrete
composite beam for
plays
enough to avoid excessive

cable-stayed bridges (espe
curvature to be locked-in in the deck.
To simulate locked-in
mented:
of
it hardens.
composite steel-concrete deck) because the repetitive, incremental
by free-cantilevering
important
case
as
in
displacements
formwork without shear
concrete cast on an elastic
connection
In
cast beam
element,
the beam element
instead of
representing
defining
the element's incidence nodes the
the formwork and
value Dz
user
defines
(see figure 2.5) representing
the distance between the cast beam and the formwork beam centroids. The cast beam
element retrieves its incidence nodes and orientation from the formwork element and the
initial
hardening displacements the formwork beam had during the casting stage the
stage coming before the one in which the new cast beam is added to the chronological ob
jects list accordingly transformed, contribute to each cast beam's initial displacements
-
"initial
This is how cast beams
The
a
user
defines
are
treated
user
modeling the formwork (the formwork beam)

object simulating the weight of the wet concrete.
beam element
distributed element force
The
numerically:
defines the cast beam
object
for the construction stage
right
and
after the
casting stage.
performing the stage-by-stage iteration the program BRIDE checks if in

the stage following the currently selected one new cast beams are introduced in
the model. If this is the case it transforms the displacement parameters of the
corresponding formwork beams to the displacements of the (not yet assembled)
While
19
4. Locked-in
displacements
rigid connection (analogous to eq. 3.7)

and stores them within the cast beam objects. These transformed displacement
parameters are the locked-in displacements a/j. They represent a contribution to
centroids of the cast beams
(see
the cast beam's S-mitiai
4.2.
Composite
assuming
eq.
2.1).
beams to simulate locked-in
displacements
in concrete cast
on
an
elastic
formwork with shear connection

The locked-in
displacements
cast beams up to
the model after the

section's
having
stiffness),
point:
casting stage (since
The
hardening displacements
way
k/0
analogous
as
for
already included in the combined

initial self-equilibrating forces f after
(see
eq.
2.3):
during
the
casting
analogously as in eq. 3.7 for the centroid of the
transformed hardening displacements are denoted by a.
to be transformed
The nodal forces q/0 applied by the formwork beam to its incidence nodes
the casting stage as a result of its elastic deformation are found as follows:
where
is the formwork's
to eq.
and denoted
3.8, q/0
(centric)
kfo
ah
during
(4.1)
stiffness matrix.
is transformed for the centroid of the combined section
by qf0.
the initial
in the
self-equilibrating force f/j)CO needed to take the locked-in displacements

composite beam, which is the first term of eq. 2.3, is found as follows:
fh>co
20
formwork,
same
which is removed from
a/,, which the formwork beam had
q/o
be found in the
its stiffness is
have to be subtracted from the
combined section. The
can
been transformed for the centroid of the combined section
stage have
for composite beams
the nodal forces of the
kco
a-
qc/0
(4.2)
Creep
5.
Concrete creep often has to be taken into account in the
According
stayed bridges.
0.4/cm1,
<
concrete may be considered
the program BRIDE

while
[CEB/FIP,
to
only
ageing (i.e. changing
is the maximum
4>oo
half the maximum

as
shown in
figure
(p^
to
time)
the
ignored2.
(pit) which
has been
the creep factor
by
<p>(t)
is
approximated by
("5))
the
(5.1)
is the number of
is reached. Both
object (see
(p^
subsection
days
in which
and 2 have to be
specified
A.1.7).
The
graph
for
<p>(t)
is
and 2 should be obtained from the
day
corresponding design code in

is loaded can also be taken into account. Typically
value around 40 days. The time t (see figure 5.1) is
the concrete first
under stress.
comes
[Wittmann, IV.9]
and
[Shaw
haviour for viscoelastic materials such

that
aTCtan
in the material
Hereditary
According
In
and the creep component ecr of strain
eei
asymptotic value of (p and 2
value around 2 and t2
measured from the
5.1.
described
0oo tanh(t
which the age of concrete when it

has
stresses, i.e.
5.1.
The values for
(poo
over
In the program BRIDE
asymptotic value (p^
input parameters
cable-
function:
4>(t)
where
usually
component
uniaxial stress-strain state.
following hyperbolic
wide-span
ageing linear viscoelastic material.
an
of material characteristics
is the ratio between the elastic

an
as
of
the linear viscoelastic behaviour has been taken into account,
The creep behaviour of materials is

in
46]
p.
analysis
in the range of service
as
and
Whiteman,
concrete
Boltzmann superposition of stress
can
be
can
p.
185] hereditary
be taken into account
applied,
creep be
by assuming
where these stress increments
by Hooke's law to corresponding strain increments. If a uniaxial stress-strain

state is considered, the stress tensor a%J and the strain tensor e%J are reduced to a scalar
and Hooke's law can be expressed as follows:
are
related
ait)
1
fem
is the
mean
value of the
D{t) e(t0)
(5.2)
of concrete
cylinder compressive strength
fc
at
an
age of 28
days,
is
the normal tension.
2This
is
constraint due to the
adoption
of Boltzmann's
superposition principle explained
in
sec
tion 5.1.
21
Creep
5.
^(t)=^.tanh(t-arctyh(a5);
.12
time
where
time
decreasing
a
constant
Let
of
us
a new
bled. If
dependence
used to approximate the creep factor.
has been introduced in
function of t. The loss of tension
e(t)
e(t0)
is known
as
denote
by an increasing
element, starting with i
we
between
<p>(t)
5.1.: Function
Figure
assume
one
which is
smooth
monotonically
decreasing of D(t) with
stress relaxation.
index i all the construction stages since the

=
that the nodal
0 for the
e(t)
assembling
in which the element has been
stage
displacements
stage and the next, the strain
the step function
D(t)
time due to the
over
and thus the strain
in the element
can
remain constant
be
assem
approximated by
e(t):
e{t)
e{U)
in[ti,ti+i[fori
(5.3)
0,1,2,...,
Each strain increment
Ae(i+i)
induces
stress increment
according
Aa(U, tj)
where
Aa(ti,t0)
and
D(U
being
assembled.
the time
e(ti+i)
e(U)
(5.4)
to Hooke's law:
Ae(0) represent
when the element passes from not
tj) Ae(tj)
for0<j
the stress and the strain increments at time t0
being
assembled
(a
of
D(t).
state)
to
according
to
stress and strain free
Notice that each of these stress increments also relaxes
dependence
(5.5)
<i
The total stress at time U is
now
given by superposition:
(5.6)
3=0
In
material
subject
components of strain
to the constant strain
eei(t)
and
e(to)
22
j=0
ecr(t)
=
e(t0)
the
ei(t)
cr(t)
of the elastic and the creep
sum
is constant and must
equal e(t0):
constant
(5.7)
5.1.
since
(p(t)
is the ratio between
e(to)
the elastic strain
Young's
eei(t)
can
eei(t)
ecr(t), ecr(t)
and
eei(t)
Hereditary
(1
be formulated
<f>(t))
as a
formulated
as
follows for t
>
the ratio between
(5.9)
place,
the elastic stress
(to)
can
be
(5.10)
a(t)
and
(to)
E
=
e(tp)
YTW)
is the relaxation factor
D(t)
be written
as
ip(t) (see figure 5.2):

<5'12)
TTW)
of eq. 5.12 for uniaxial stress-strain states
help
can
e(t0)
5.6
and the
0:
a{t)
Eq.
a(t)
follows:
a(t0)
With
(5.8)
constant
function of the elastic stress
hence for to, when the relaxation has not yet taken
as
be eliminated from eq. 5.7:
modulus E:
Sel{t)
and
can
D(t)
can
be determined
as
follows:
E-ip(t)
(5.13)
Y,E-4>(tt-tJ)Ae(tJ)
(5.14)
follows:
i
a(tt)
3=0
Ignoring
eq. 5.14
the contribution to creep from the element's local
can
be
generalized
difference between the nodal
1 and
displacements Aa(tJ)
of
the
relationship of
proportional to the
Ae(t3)
two subsequent stages numbered
load,
for the nodal forces q, since
is
j:
i
q(tt)
^2k-iP(tt-t3)-Aa(t3)
(5.15)
3=0
where k is the element's local stiffness matrix and the difference
Aa(t,)
is found
as
follows:
Aa(ij)
where
a(t)
(a(t3)
a.sh(t3)
auser(t3))
is the vector of the nodal
(a(i,_i)
displacements
ash(t3-i)
from the
auser(t3-i))
global solution, ash(t)
(5.16)
is the
shrinkage and auser(t) is the vector of the initial

displacement defined by the user (e.g. to model a pre-camber or a cable post-tensioning).
a(t_i), ash(t-i) and aMser(_1) needed to find Aa(0) are zero vectors representing the
vector of the initial
displacements
due to
unloaded state of the element before it is assembled in the structure.
23
5.
Creep
AI
o(0)(l-v(t))
0(00)
time
5.2.: Relaxation in
Figure
Every geometry and load change in

and j induces a change Aa(t,) which is
increment of the nodal forces
Aa(tj)
is
computed
as
Aq(tj)
bar
the structure
found
be
expressed
we
as
corresponds
to
an
initial
can
is
superposition of all
global
equation system. The
displacements
(5.17)
Aa(t3)
magnitude
of
Aq(t,)
due to
(1
tP(tt
tj))
displacement
acr
(1
iP(U
2.2 and
Y,âcr(U,Aq(t3))
3=0
24
Aa(t3)
(see eq.
Aacr(ti, Aq(t,)):
displacement
acr(tt)
by solving
the
show that the loss of
Aacr(tt, Aq(t3))
the total initial
occurring between the stages j
follows:
Aqcr(t3)
which
t0.
follows:
From eq. 5.17 and eq. 5.15

can
at the time
related to the increment in the nodal
Aq(t3)
relaxation
lengthened by A/
2.3)
needed to take creep into account
Y, Aafe)
3=0
(5.18)
tj)) Aa(t3)
(1
"
4>{U
tj))
(5.19)
Approximation
5.2.
5.2.
in the
computation of
computation of
in the
Approximation
structures with a
creep in structures with
varying
varying
strain
creep in
strain
While creep is considered
exactly with eq. 5.19 for systems having a constant strain,

having
varying strain an error is introduced because the contributions to
acr from the nodal displacements due to creep between subsequent stages are ignored.
However this approximation has proved to be good and to improve rapidly if the time is
subdivided in more steps through the definition of more possibly fictitious construction
for system
stages.
The increase of the
extreme
at
once
case
for
displacements
due to creep in
varying strains) having
at the time to
can
iteration
can
statically
displacements
found by
implicitly
(p(ti)
vector
be formulated
as
determinate model
material and in which all load is
be summarized with the factor
any component of the
The factor
single
4>b(U)
a7ta\
(the
applied
where
is
a.
the program BRIDE
during
the
stage-by-stage
follows:
(p(ti)
YSl'a
(5.20)
-ipih -t0)
(5.21)
'
~~
va=0
where:
Si.O
Sl,l
i-2
y^ Si-2,a
(1
1p(U
U-i)
(5.22)
a=0
"3,3
"^
This
complicated formulation,
by-stage iteration,
"
^(t3
which
*,-l)
exactly
(i
V(*.
(5.23)
*j-i)
expresses what
can
be visualized for four construction
happens during the stagestages by means of the following
table:
3=0
S*,J
l-V(*i-*o)
l-V(*2-*0)
(i-V(*i--t0))(l-V'(i2--tl))
((1-^2 _to))
l-V(*3-*0)
(i-V(*i--t0))(l-ip(t3--h))
error
r(tt)
which
occurs
the creep function and the actual
between the function
(p(t)
r(ti)
can
be
4>b(U)
expressed
-
<KU)
4>(U)
(l_ ^(*i
*0))(l-^(*2-*l))-(l--V(tito))
The relative
-i
U-V>te-t2))
(p(t) adopted
as
to
approximate
follows:
(5.24)
25
5.
Creep
days
100
200
400
300
500
600
700
800
..00*08^"
-0 1
<>
-0 2
<>
o *
[]
x
o
-0 3
-0 4
<>
o o o
[]
"'
032
slages
D16 slages
A8
slages
X4 slages
02 slages
i i
>;
1 slage
-0 5
-0 6
);
C)
-0 7
Figure
5.3.: Relative
and for
error
between
material
(pB(t)
having
and
(p^
(p(tt)
2 and
with different subdivisions of time

a
100
days.
numerically. Figure 5.3 shows a chart of r(tt) for

different subdivisions of time: with an increasing subdivision, (p(t) converges towards
the exact solution and for subdivision into 32 steps r(tt) has a magnitude comparable
to the approximation in the estimation of 0oo and 2 in the design codes.
and its
26
magnitude
can
be controlled
6.
Concrete
Concrete
shrinkage is
cable-stayed bridges.
the model in a simple
shrinkage
difficult to
predict
but has to be taken into account for wide span
The program BRIDE allows

and effective way
exploiting
one
the
element has been added to the structure thanks to the

For
our
to introduce
complex phenomenon can be simplified

approximated with the same function as for creep:
sh(t)
(see figure 6.1)
shoo
tanh(t
values in
advantage of knowing when

chronological object list.
purposes this
initial strain
empirical
to
each
time-dependent
arctanh(0.5)
object (see A.1.7) by two values: the maximum

asymptotic value shoo and the number of days 2 in which it reaches half the maximum
asymptotic value. The time , however, has to be measured from the day the concrete
hardens (even if it is not stressed, which is the case when prefabricated panels are used).
defined in the material
The values shoo and 2 should be obtained from the relevant

the value for eshoo is around 0.0005 and the value for 2 is
design code. Typically

between 30 days and 2 years.
time
Figure
6.1.: Function
esh(t)
used to approximate the initial strain due to
shrinkage.
27
Concrete
6.
6.1.
shrinkage
Initial
displacements
a./, needed to take concrete
into account
shrinkage
In contrast to locked-in
stage
iteration to be
cast. In the
of the
case
casting
displacements and creep, shrinkage does not need the stage-bycomputed because it depends only on the time 0 the concrete was
of cast and composite beams the time to
stage, otherwise it
in which the material
(a
model
corresponds
object
as
to the
any
day
other)
corresponds
to the
day
number
number of the construction stage
has been defined.
However,
shrinkage into account during the stage-by-stage iteration

displaced shape which, in turn, affects locked-in displacements,
it is
necessary to take
because it
affects the
creep and
conditional loads.
The initial
are
it
displacements
obtained from the nodal
was
ash is
fully clamped
a
shrinkage into account (see eq. 2.1 and 2.3)

displacements the shrinkage-affected element would have if
ash
needed to take
in the first node and free to
vector in which the
only
non-zero
d.sh,tx,2
move
component
.sh(t
to)
corresponding to the translation in the longitudinal

node, where 0 is the time at which creep begins,
element length.
28
in the second.
is the
For this
reason
component
direction of the second incidence

is the
analysis
time and / is the
7.
Conditional loads
In most structures the
displacements are so small that they do not cause an excessive

planned geometry. Quite contrary to this rule, displacements play
deviation from the
cable-stayed bridges and are usually corrected by post-tensioning the

cables and by mounting mast and deck segments with a pre-camber. Such adjustments
in geometry can be simulated with initial displacements whose magnitude is found au
tomatically by the program BRIDE and which are denoted as conditional loads.
Conditional loads are load objects (see subsections A.1.12 and A.1.13) to which a
condition to be fulfilled has been appended (see subsection A. 1.14). These are a few
examples of the statements which can be formulated in the program input with such
a
major role
in
conditions:
The initial
condition is
the
whose
appended
anchoring
the load
should have
point of the cable
specified
in the load
value such that the
equals
zero
object to which the

vertical displacement of
in the "last" construction
"loacLhistory" and taking all other conditional

intensity are also found in the "last" construction stage.
case
The initial
condition is
shape
of the cable
displacement
displacement of the cable specified

appended should have a value, such
in the load
stage, under
loads into account
object
to which the
that the rotation of the deflected
of the formwork beam to which the loaded cable is anchored is
equal
at the
beginning and ending during the formwork's casting stage1, under

"loacLhistory" and taking all other conditional loads into account
intensity are also found in the formwork's casting stage.
formwork beam's
the load
whose
case
The initial force

the condition is
acting
on
appended
the deck segment

and
modelling
assumptions made at input should have

ment of the deck at the measurement
construction
load
case
intensity
"loacLhistory"
are
can
object
to which
value such that the vertical
point equals the measured value

to the measurement
day),
displace
in
given
under the
all other conditional loads into account whose
also found in that construction stage.
The imposition of absolute

which
taking
in the load
the deviation of the real creep from the

a
stage (the stage corresponding

and
specified
displacements
and section forces
are
not the
only
conditions
be formulated. The program BRIDE also allows relative and "same factor as"
conditions
(see
1This condition
explained
in
section A. 1.14).
can
be used to avoid the accumulation of locked-m
bending
in
composite decks,
as
[Schlaich]
29
7.
Conditional loads
typical case in which a "same factor as" condition should be used is the one in
which two symmetric cables, both anchored to the ground on the same anchor block
but one on the right and the other on the left of the carriageway, and both reaching the
top of the same mast, have to be post-tensioned to avoid a longitudinal displacement
of the mast's top. This should be formulated using an absolute condition for one cable
(i.e. "post-tension the first cable in such a way that the top of the mast does not move
longitudinally") and a same factor condition for the second cable (i.e. "post-tension the
second cable
so as
to have the
For the definition of
same
conditional factor
conditional load the
as
for the first
cable").
is needed:
following
A load
object whose correct intensity is not yet known and thus still arbitrary:
the arbitrary intensity load object. During the stage-by-stage iteration the correct
intensity is automatically found by the program BRIDE in the form of a conditional
factor with which the load represented by the arbitrary intensity load object is
multiplied. Actually after such multiplication the arbitrary intensity load object
does not deserve the adjective "arbitrary
anymore, because its intensity has been
computed; however we keep calling them an arbitrary intensity load object because
this computed intensity is used to perform the influence matrix algorithm (see
section 7.1) in the following stage-by-stage iterations in the same way it was used
in the first stage-by-stage iteration, when it really was arbitrary.
"
A condition to be fulfilled whose definition is
arbitrary intensity
object (see
load
three bits of information
the condition
subsections A. 1.12 and A. 1.13) and contains
subsection
A.1.14):
specifying which displacement or section

value at a given point of the structure
specifying
condition value
rameter has to
to the definition of the
condition parameter
assume
(see
appended
condition
the absolute
force has to
relative value the condition pa
or
assume
in which the condition has to be fulfilled.
stage
algebraic point of view it would not be strictly necessary to couple the

conditions to an arbitrary intensity load.
This coupling however is reasonable
because it makes sure that in the input there are always as many conditions as
arbitrary intensity loads to be updated in a given stage. This forces the user to
think about the influence of the arbitrary intensity load on the condition parameter
From
an
(reducing
the risk that
non-solvable equation system is
the program to check that all
during,
but not
load
fixed
the load
the
30
same
after,
arbitrary intensity
are
and allows
activated before
or
the condition stage.
consisting of all loads
case
loads
formulated)
active
load-history excluding
condition stage.
during
the condition stage
those whose
intensity
is
belonging
being computed
to
in
7.1.
7.1.
The
section forces:
one can
displacements
and
knowing the displacements and the section forces caused by given loads
displacements and the section forces caused by any linear combination
just by combining the respective displacements and section forces with
find the
of those loads
the
algorithm
of first order static models allows the superposition of
linearity
algorithm
same
linear factors used to combine the loads.
Conversely, knowing the displacements and the section forces caused by n loads of ar
bitrary intensity and having n conditions to fulfil in the form of prescribed displacements
and section forces, one may find the linear factors needed for a conditions-fulfilling load
combination. These linear factors are called conditional factors and they are found au
tomatically by the program during the stage-by-stage iteration with the influence matrix
algorithm developed in this project.
This is how the influence matrix algorithm works if all condition values are absolute:
1.
stage-by-stage iteration, given n arbitrary intensity loads associated to

having as a common condition stage, the stage actually selected by
the stage-by-stage iteration, the nxn influence matrix N has to be built, where Nl3
is the value assumed by the condition parameter defined in the condition i under
the arbitrary intensity load j.
During
n
2.
the
conditions
The nxl vector L has to be
parameter of condition
3.
built,
where Ct is the
following equation system has to

containing all the conditional factors:
NF + L
parameter which is
not
absolute,
another point of the structure
bending moment here

corresponding factors
Nl3
C ^F
but
has to induce
portion of
(e.g. "post-tension
is half the
bending
section force
or
(i.e.
moment
"the
displacement multiplied by
moment
there")
under the
(7.1)
a
value of the condition
section force
the cable with
for the matrices of eq. 7.1
ciated to the conditional load
condition value of
N"1(C-L)
there").
are
found
is the difference between the value assumed
bending
required
be built and solved to find the vector F
Relative conditions state that the conditional load
the
the condition
i.
4. The
the
by
under the fixed loads.
The nxl vector C has to be

condition
where Lt is the value assumed
built,
by
bending
an
or
displacement
intensity,
If the condition is relative

as
follows:
the condition parameter

moment
section force
bending
or
arbitrary intensity
(i.e.
"the
displacement multiplied by
moment
there")
asso
here") and the other

(i.e. "half of the
the proportion factor

load j.
Lt is the difference between the value assumed by the condition parameter

ciated to the conditional load
at
such that
bending
moment
asso
here") and the other

(i.e. "half of the
the proportion factor
under the fixed load.
31
Conditional loads
7.
C\
is set to 0.
The "same factor as" condition is used to avoid

condition parameter
tional factor
as
by stating
conditional loads
more
having
that the conditional load i must have the
j (Ft
another conditional load
F3).
Such conditions
are
the
same
same
condi
expressed
as
follows in the influence matrix:

the i-th
of
on
all other factors
7.2.
row
N; Nn and Nl3
on
The need to
repeatedly
As
the
explained
row
set to 1 and -1
i inclusive of
perform
to
are
the
L% and C%
respectively,
are
set to 0.
stage-by-stage
iteration
find the conditional loads
in the introduction to this
chapter
conditional load is constituted
by
arbitrary intensity load coupled with a condition. The conditional factor needed to
multiply this load is found with the influence matrix algorithm when, during stage-bystage iteration, the condition stage specified in the condition is selected (see section 7.1).
Frequently the condition stage does not correspond to the activation stage of the
conditional load (i.e. the stage in which the conditional load is added to the chronological
objects list and begins to act on the structure) but it is a posterior one, e.g. often a cable
has to be tensioned during construction to avoid a vertical displacement of its anchoring
point on the deck in the final stage. If the condition does not correspond to the activation
stage there are stages in between (activation stage included) in which, during the stageby-stage iteration, the conditional load is already acting on the structure but with its
still unmodified arbitrary intensity.
a
an
If the conditions of all conditional loads have to be fulfilled in the

if
new
conditional loads
conditional
because
no
inserted in the
same
stage
or
list after the stage
chronological objects
formerly defined, the
factors found with the first stage-by-stage iteration are the correct ones
conditional load of unmodified arbitrary intensity was considered as part of
objects representing
are
the condition stages of all conditional loads
the fixed load.
By contrast,
if conditions have to be fulfilled for
ditional loads whose condition has to be fulfilled in

factors of the conditional loads
an
a
earlier stage in which other
later stage
in the earlier
are
already active,
con
the
incorrect after the first
computed
stage
stage-by-stage iteration because they have been computed with a fixed load containing
also the still arbitrary intensity of the conditional loads determined at a later stage.
When during the stage-by-stage iteration conditional factors are found, they are im
mediately used to correct the arbitrary intensity of the arbitrary intensity loads they are
associated to; hence the influence matrix algorithm can converge if the whole stage-bystage iteration is performed repeatedly (leaving out the case in which locked-in displace
ments are taken into account explained in subsection 7.3.2) because the errors manifest
themselves as wrong values of the vector L and are compensated in the next stage-bystage iteration (see equation 7.1).
32
are
Convergence
7.3.
of such
practical example
which the cables
shape
of the influence matrix
post-tensioned
are
in the formwork element
construction
twice:
during
stage (see [Schlaich]
for
the
the first time to induce
casting stage and
details).
more
In this
case
part of the fixed load, induce
loads with
the
condition to be fulfilled
stage-by-stage
7.3.
an
iteration two
Convergence
error
during
or more
in
given deflected
second time for the final

the
the conditional loads with the condition to be fulfilled in the final

as
with non-linear models
given by bridges with composite decks
situation is
algorithm
arbitrary intensity of
stage, being considered
in the conditional factors of the conditional
the
casting stage and
it is necessary to
repeat
times.
of the influence matrix
algorithm
with
non-linear models
As shown in the
practical example
of
9 the influence matrix
chapter
not need to converge for non-linear models to be
intensities found with
linear model
the conditional parameters
However, the
the
even
are
exact
if non-linear effects
does
since the conditional load
applicable,
enough
algorithm
and lead to
acceptable
values of
taken into account.
are
convergence behaviour of the influence matrix
algorithm
combined with
iteration has been studied and the
repeatedly performed stage-by-stage

explain the results of this investigation.
following
sub-sections
Creep
7.3.1.
and
shrinkage
The influence matrix combined with the
quickly also
if concrete creep and shrinkage are being taken into account, because these phenomena
do not change the material stiffness. The errors they induce in the values assumed by
the condition parameters appear in the vector L and are also automatically corrected by
the influence matrix algorithm when performed during the next stage-by-stage iteration.
7.3.2.
During
Locked-in
the
iteration it is
compute the conditional
stage of every conditional load is not

one.
In this
intensity
is
case
(once
assigned
stresses found in all
the
possible to
factors, and
the
always
stage-by-stage iteration
physically
newly found
take locked-in
it may
same as
happen
displacements
into
that the condition
its activation
stage but
reaches the condition
stage)
later
a new
to the conditional load which would have affected the locked-in
stages between the activation stage and the condition stage. The
consequence of this is that at the end of such

a
iteration converges
displacements
stage-by-stage
account and to
stage-by-stage
inconsistent
stage-by-stage iteration the model is in

displacements do not reflect the
state, because the locked-in
conditional load intensities.
This
physical inconsistency leads to the necessity of always performing the stageby-stage iteration twice (which is automatically done by the program BRIDE) when
computing the conditional factors taking locked-in displacements into account:
33
7.
Conditional loads
1.
During
the first
stage-by-stage
iteration locked-in
the intensities of the conditional loads

2.
During
the second
stage-by-stage
displacements
displacements
are
just locked-in displacements
iteration
updated
considered and
computed.
are
ered while the intensities of the conditional loads

the locked-in
are
to the
are
new
kept unchanged.
are
consid
In this way
intensities of the conditional
loads.
There is another curious effect which
impedes
the convergence of the influence matrix
repeatedly performing the stage-by-stage iteration.

displacements amplify the effect on the deflected shape caused by the
changes of the conditional load intensities between subsequent stage-by-stage iterations
in such a way that the self-correcting behaviour of the influence matrix algorithm ex
plained in section 7.2 cannot take place. In other words, when taking locked-in dis
placements into account the structure "over-reacts" to the changes in the conditional
load intensities which occur when the stage-by-stage iteration is performed repeatedly
and in each stage-by-stage iteration the conditional factors are computed again: instead
of converging to the condition values the values of the condition parameters oscillate
around the condition values with growing distance. This fact can be observed by the
user with the graphic representation of the deflected shape which is updated every time
a stage-by-stage iteration is completed.
The error feed-back factor a ranging from 0 to 1 has been introduced to avoid such
oscillation and has to be set interactively by the user when the stage-by-stage iteration is
repeatedly performed on the basis of the observed variation of the deflected shape. Let's
call R the residuum affecting the condition parameters after a stage-by-stage iteration,
the residuum R multiplied with the feed-back factor a is added as follows to the right
side of eq. 7.1 in the next stage-by-stage iteration:
algorithm
when
Locked-in
NF + L
If
change,
is set to 0 eq. 7.2
hence the
equals
the
(error-affected)
(7.2)
C + R-a
self-correcting
eq.
7.1.
solution is stabilized.
1 makes the influence matrix
Any
If
is set to 1 F cannot
value of
less
between 0 and
While
algorithm converge
quickly.
repeatedly
performing the stage-by-stage iteration computing both conditional loads and locked-in
displacements the user should keep a as close as possible to 0 but big enough to avoid
more or
the oscillation of the values of the condition parameters about the condition values.
7.3.3.
Second order effects and cable sag
Second order effects and cable

eq. 7.1
as
2The cable
sag2
are
stiffness
changing phenomena
follows:
sag is taken into account in the program BRIDE
tangent Young's modulus Etan derived in
[Gimsing,
p.
using the parabola-based formula for the
141]:
~/2l2E
12<r3
34
which transform
7.4.
(N
where AN is the
Conclusions
AN)F
+ L
regarding
(7.3)
in the influence matrix due to
change
in the stiffness of the
change
structure.
This stiffness
change can be taken into account by computing the influence matrix

using the updated stiffness of the structure.
If locked-in displacements are not taken into account (hence abrupt changes of the
conditional factors between subsequent stage-by-stage iterations do not cause the overreaction of the structure explained in section 7.3.2) the result F can be improved by
performing a final stage-by-stage iteration in which the error R is subtracted from the
condition values (in eq. 7.2 this would correspond to a
1). With this last iteration
the error in the newly computed conditional factors is reduced dramatically. This is
what happens numerically:
=
In the last
stage-by-stage iteration, once the error has stabilized itself, instead of

explicitly finding N + AN we can assume that ANF equals the residuum R, being
ANF the only term in eq. 7.3 differing from eq. 7.1
1.
The correct F
2.
can
be found
by re-formulating equation
NF + L
which is what is done in the last
stage-by-stage
7.3
follows:
as
(7.4)
R
iteration if the
user
has chosen this
option.
7.4.
Conclusions
regarding
Not every conditional load leads to
solvable equation system: if the
user
tries to induce
arbitrary intensity load j having no influence on the condition

parameter i (i.e. Nl3
0), it is quite possible that the matrix N is going to be singular.
Another unsolvable situation is given if the same condition parameter is used more than
once, as this leads to two identical rows in N (which should be avoided using a "same
factor as" condition).
More generally it can be said that the major problem in the
formulation of conditional loads is that the user has to have a good understanding of
a
displacement
with
an
=
the behaviour of the structure.

The important aspect to be considered when
not its absolute
even
intensity which
can
if it is very different from the
be corrected in
specially
problematic aspect of the influence matrix algorithm

representing section forces (and not displacements) can
where E is the
cable's
few
an
the condition parameter.
Another
rameters
appropriate value, but
load influences the static system and
arbitrary intensity load is

linear stage-by-stage iterations
the way the arbitrary intensity
formulating
length
Young's
and
modulus and 7 is the
weight
is that condition pa
lead to
per unit volume of the cable's
near-singular
material,
/ is the
the stress in the cable.
35
7.
Conditional loads
delivering the conditional factors exact condition fulfilment. Also

in this case a good understanding of the structure's behaviour is required from the
user, who, in some cases, might be forced to choose alternative conditions prescribing
displacements.
influence matrices not
36
8.
Automatic cable
dimensioning
Automatic cable
dimensioning has been implemented in the program BRIDE. In fact,

cable dimensioning is a tedious and time-consuming task if performed manually.
To perform an automatic cable dimensioning the user has to specify the cable dimen
sioning objects (see subsection A.1.15) in the chronological objects list and simply click
on a button of the graphic user interface.
8.1.
Cable
When the
user
dimensioning algorithm
clicks
on
the cable
dimensioning
button the program finds for every cable
to be dimensioned the smallest sections which resist all normal forces
the cables in every construction stage and load combination pair
experienced by
specified in the cable
dimensioning objects.
To dimension
by
cable it is necessary to know its normal force which
of structural
turn, structural
be
be found
only
performed only if
can
analysis.
analysis
already been assigned to the cable. Thus the cable sections cannot be
dimensioned directly but only iteratively by first adopting an approximative section. If
this is very different from the adequate one it may be necessary to iterate (by simply
clicking the cable dimensioning button) two or three times. In addition sections must
be chosen from an appropriate pool of predefined cable types.
This is how the section re-dimensioning problem has been solved:
a
means
In
can
section has
1.
Every
tion
2.
section
object
is
assigned
section group with
an
identifier
(see
subsec
A.l.6).
As for all other elements the
they
to
are
going
user
has to define
to be re-dimensioned
(see
section for every
subsection A.
cable,
even
if
1.5).
3. While
performing the automatic cable dimensioning the program uses the section
specified by the user as the initial approximative section and looks for the smallest
section under the sections belonging to the same section group.
37
8. Automatic cable
38
dimensioning
9.
Example of
input file for
an
the
program BRIDE
In this
chapter
an
of
example
computation performed with the
discussed. The model used for the computation has been

different models used to test the program
all
implemented
Our
during
its
program BRIDE is
obtained
freely
development
so
by merging
that it
can
show
features.
imaginary bridge
is
symmetric and has been defined using
At the
figure 9.1) all materials and sections

definitions, a second input file (see
section D.2 and figure 9.2) in which one half of the model is defined keeping the signs
of the X and Y coordinates and all identifiers dependent on a single variable "s" (to
be defined before inclusion, as mentioned in figure 9.1) is included twice (once for each
model side). Between the first and the second file inclusion instruction a restart object
is inserted. After this double inclusion all objects needed to join together the two halves
of the main file
beginning
needed
(see
two files.
are
defined.
section D.l and
After the material and section
and to dimension the cables
are
defined.
joints joining together
the two
second
first
bridge
Figure
half
(the
variable "s" is set to 1
9.1.: The two
bridge
bridge
bridge
halves
half
(the
variable "s" is set to
2)
halves
joined together
in the main file.
39
of
Example
9.
Figure
an
9.2.: Half
input file for the
defined in the included file with the convention used to define
bridge
the identifiers: the

line and
number
the identifier.
(where
span
program BRIDE
objects on a dashed line have the letter assigned to that

growing along the line in the direction of the arrow in
Notice that the formwork beams in the middle of the mid-
the letters
"d" and "e"
"c",
the composite beams which

excentric
relatively
are
are)
have not been
defined in the outer
to the other's
merged yet with

file, hence they appear
composite beams.
This definition with two files has the

model for
a
single
side
(which
better overview of the way the two

The model has
Another
the
use
example
composite deck modelled with composite beams (see figure 9.3).
with cast beams instead of composite beams is not shown because
of composite and cast beams is
beams instead of composite beams

all
advantage of allowing an easy formulation of the

during construction, see section 9.2) and to keep
sides are joined together.
is useful
occurrences
of the
can
In fact
fully analogous.
model with cast
be obtained from the model
keyword "CompositeBeam"
with the
by simply replacing
keyword "CastBeam" (see
figure 9.4).
9.1.
After
Analysis strategy
writing of the input file(s) the user has to open it with the
(actually it is advisable to open the input file also when it is not yet
having completed
program BRIDE
completed,
so as
the
to track all
errors as soon as
possible using
the
error
messages from the
program).
As mentioned in
when
taking
convergence is
40
chapter 7,
the
stage-by-stage iteration does not always converge

However, it has been observed that such
all non-linear effects into account.
actually
not necessary since the intensities of the conditional loads found
9.1.
Analysis strategy
4 I fiitim-im
"iirllai iils
Figure
sip
9.3.: The model listed and commented in
The box
of the element
shape
the element's section and has
appendix
D with its
composite deck.
only shows the maxima and

nothing to do either with its
the minima of
stiffness
or
its
self-weight.
with
satisfactory precision in the conditional parameters also

when,
having
linearly, non-linearities are taken into
account in the model. This is plausible since otherwise it would be incomprehensible
that cable-stayed bridges could be successfully built with the manual methods listed in
a
linear model lead to
after
section 1.2. For this
All
found the conditional loads
reason a
two
step strategy has been used:
del-changing computations such as the computation of the conditional load

intensities and the cable dimensioning are first performed using a linear model.
The model modified by the computation of the conditional loads and of the cable
sections is then stored as an expanded input file (i.e. an input file without pre
mo
processor instructions and in which the intensities of the conditional loads and the
cable sections have been
updated,
The stored model is

such
as
second
opened
order, cable
taken into account with

1Tension/compression-only
engineer
history-
is
are
figures
from 9.5 to
9.10).
with the program BRIDE. All non-linear

locked-in
phenomena1
and creep
are
displacements
shrinkage,
which
the
of
intensities
iteration,
stage-by-stage
during
left unchanged (see figures from 9.11 to 9.14).
sag,
behaviour
asssumed to have
see
is
designed
relevant loss of contact of
not taken into account at this
the structure
such
support and the loss of
stage of the analysis
way that
tension
m a
an
since
the
unforseen construction-
cable, respectively,
does not
occur
41
Example
Figure
of
input file for the
an
program BRIDE
9.4.: The model listed and commented in
appendix
D where its
composite beams
by cast beams simply by replacing all occurrences of the

keyword "CompositeBeam" with the keyword "CastBeam" in the input files.
are
9.2.
Monitoring
file
During
substituted
the
erection
by keeping
the
input
up-to-date
the construction of the
the events in situ:

conditional
bridge
if
bridge
cable has been
the input file has to be
post-tensioned
with its
kept up-to-date
with
intensity found using
this conditional load definition in the input file has to be transformed
load,
regular load by removing the condition and by substituting the computed intensity
with the intensity measured while post-tensioning the cable.
If a deviation from the planned geometry is observed, e.g. due to an unforseen creep
behaviour of concrete, the engineer on site should update the input file by producing
this deviation in the model by means of a conditional load. Once the deviation has been
introduced into the model the engineer can investigate which adjustments have to be
undertaken, e.g. introducing additional conditional loads.
At the end of construction such an input file kept up-to-date also represents a detailed
report of all relevant events that occurred during the erection of the bridge.
into
42
9.2.
i...^_>
.-_
Monitoring
t,
the
bridge
%,^% ###,#s
by keeping
erection
^b***
ft
the input hie
up-to-date
r ? /
Figure
9.5.:
Step
1.
/ :x\ .l
The file listed in
,.^v
appendix
D is
the last construction stage and under the

A
displaced
t t,i \ vv t\
opened
and its
load-history
'
F~-
*****
displaced shape in
case is presented.
load
node in the middle of the mid-deck has been selected and its
displayed in the text panel (the units are meters and

radians). The model has no pre-camber, no post-tensioning and the cables
are too flexible thus the displacements are obviously too big.
displacement
values
are
FT
fFr^JTOg WMS^WT ITS &
<#
BB*i*l |Q*|P
Ssiuia oil *tii
Mf
^__y_
Figure
9.6.:
Step
into
2.
stage-by-stage iteration (i.e. not taking non-linearities

performed in which the conditional loads are found.
One linear
account)
is
43
9.
of
Example
an
input hie for the
program BRIDE
Donnai
ti
1
S
1,
%"% ####51
M*"*f
ff
fil
IPS Hi
ta!
/lit
II
'
:/,' ,'
tel?:!^*}~^,
-'
*
\
IMUMiiilu
V^. 'd. A V'
I* in m ImJ ill ,
hi* Hnwr
Figure
9.7.:
Step
3.
The cables
are
dimensioned.
The
change
of stiffness of the cables
makes the conditional loads out of date.
&
j,
'
%% ##,##Si
M#"f
f *
ff
h,
T555535
larfaUff*^
T^r
^j
Figure
9.8.:
Step
4.
Another linear
conditional loads
44
are
stage-by-stage iteration is performed in

updated to the new stiffness of the cables.
which the
Monitoring
9.2.
the
bridge
erection
by keeping
the input hie
up-to-date
HPffiT
gj^*""""",, ,
Figure
9.9.:
Step
5.
With this
model,
after
fT
ITS ft.
having updated
the conditional loads the
cables need to be dimensioned another time, which makes the conditional

loads
*f
*^
H
|i/!*!*I isilf
fila
f IM
s-aai
being
out of date
again.
j|j=gllbibbmbbib^ Hsn "3"= fc-
-y-
aal feais
i* IS- f
rf faaU 11 *
"N \
-^^
fss
Figure
9.10.:
r,
c-ijifu-u.-v-
*>
r-u
A last linear
stage-by-stage iteration is performed in which the

conditional loads are updated to the new cable stiffnesses. Now the cable
sections are the right ones and are no longer changed by the program even
if the button for the cables dimensioning is pressed again.
Step
6.
45
9.
Example
Figure
9.11.:
of
an
Step
input hie for the
7.
The model with the dimensioned cables and the
tional loads is saved
Figure
9.12.:
Step
8.
step 7 is
46
program BRIDE
The
as a new
original file
opened instead.
updated
condi
input file.
is closed with
no
changes
and the file stored in
Monitoring
9.2.
'&
bridge
erection
IIIMf"f
<%*% ###-#s
i,
by keeping
f,s
up-to-date
,;<"
-Tv
the input hie
>
---
the
-.^-^
v T
',
'
TU.! T
If
^-TT-i' i
>
K^t^^
-
-f
ml imn
Figure
^<Tii-^5BiS^ttoHii'iraiilBS'TSr5BSiT
9.13.:
Step
9.
step
The
are
slight
differences with respect to the nodal
due to the
rounding
of numeric values when
displacements
writing
in
the stored
file to disk.
r|El*,-,,,,,..t
1
*
S=ii*JFf^^W*i^TTTT=
r Sim
lIJMf*T]a. allpiitU ISSlha
_<;
IsrSr
Figure
9.14.:
stage-by-stage iteration taking all non-linearities into

account and not updating the conditional loads can be performed. The
biggest displacements can be checked. In fact, they should be in an accept
able range (here a node at the centre of the deck close to the right anchor
Step
10.
Finally,
blocks has been
selected).
47
9.
48
Example
of
an
input hie for the
program BRIDE
10.
In its
Conclusions
present form the program BRIDE has
commercial
to the
use.
Nevertheless,
it
represents
not
yet reached the maturity required for
prototype with which original approaches
important and non-trivial
practical problem of the design, erection and retrofitting

of cable-stayed bridges could be developed.
Three concepts have been introduced and developed which allow one to model a cable-
stayed bridge
in
very effective way:
Chronological objects list: when describing the construction of a cable-stayed bridge

not only the information about which parts are added to the structure, but also
the information about when they are added to the structure is important. The
chronological object list concept allows one to include this information in the input
file, i.e. in a natural way.
Stage-by-stage
iteration: creep and locked-in
the structure's previous deflected
on
influence
the deflected
displacements
shape. Or, for the
also
at
same
to have
given
time
having
reason,
influence
shape, they
going
subsequent deflected shapes. The stage-by-stage iteration allows one to
these phenomena into account in a simple but effective way by modelling
on
are
order and
an
depend
on
an
all
take all
all
con
the effects of creep and
chronological
by retrieving
stages
displacements from the deflected shapes of previous construction stages.
The adoption of the stage-by-stage iteration also allowed the development of quite
innovative ways of taking creep and locked-in displacements into account.
in
struction
locked-in
Conditional loads:
post-tensioning can be modelled as initial displace

ments, but how large should be such initial displacements? The newly introduced
conditional loads are an elegant solution to this problem, in which the user does
specify
not
pre-camber
the absolute
and
intensity of loads explicitly but through
condition to be
and optimize the
post-tensioning
fulfilled.
Since the program BRIDE allows
one
to
accurately plan
only
saving for
the civil
and construction but possibly also in a reduced
number of times the forces in the cables need to be readjusted.
The rather interdisciplinary nature of this project (structural engineering, materials,
information technology) required a very summarizing approach from the author. Now
that the project's skeletal structure has been established, it would be desirable that
schedule,
the added value
from its
deriving
engineer responsible for design
further research and validation work
developed
solution
algorithms
were
on
use
may extend not
to the time
the individual aspects and assumptions of the
performed.
Such research would also be
helpful
for
49
10.
Conclusions
the consolidation of the
effectiveness,
approach developed
still needs from the
user a
in this
project which, despite
good understanding
for the static
its
promising
system's
inner
logic.
The program BRIDE has been
models.
implemented
However, for preliminary design
for the simulation of three-dimensional
two-dimensional
analysis
This restriction would allow to
and the computation
in such
be reached than with
way that
much
simplify the graphic user interface

higher degree of user interaction could
is often sufficient.
three-dimensional program. It is the opinion of the author that the
two-dimensional but
the
analysis
to many
50
of
users.
highly
cable-stayed bridges adopted
interactive program
implementing
the
development of a
original approach to
for the program BRIDE would be welcome
Bibliography
[Anderheggen]
Anderheggen, Lineare Finite-Element-Methoden: eine Ein

fhrung fr Ingenieure, Institut fr Baustatik und Konstruktion,
ETH Zrich, 1993
[CEB/FIP]
Structural
on
mance,
the
E.
Concrete, Textbook
Updated knowledge of
Behaviour, Design
CEB/FIP
and Perfor
Model Code 1990,
Volume 1, International Federation for Structural Concrete
(fib),
1999
[FAQs]
G.
Cline,
M.
Addison-Wesley,
[Gimsing]
Java
[JNI]
R.
Edition,
and
on-line
Tutorial,
design,
Sec
document,
http://java.sun.com/developer/onlineTraining/java3d/
The
URL:
API
3D
Second
1998
Gimsing, Cable supported bridges concept

edition, John Wiley &Sons, 1997
URL:
[Java Tutorial]
Girou, C++ FAQs,
N. J.
ond
[Java 3D]
M.
Lomow,
on-line
Tutorial,
Java
document,
http://java.sun.com/docs/books/tutorial
Gordon,
Essential JNI: Java Native
Interface,
Prentice
Hall,
1998
[Patterns]
E.
Gamma,
R.
Helm,
R.
terns, Elements of Reusable
Wesley,
[Refactoring]
M.
Johnson,
Vlissides, Design Pat

Object-Oriented Software, AddisonJ.
1995
Fowler,
K.
Beck,
J.
Brant,
toring: Improving the Design of
Opdyke, D. Roberts, Refac

Existing Code, Addison-Wesley,
W.
1999
[Schlaich]
M.
Schlaich,
Erection of
Cable-Stayed Bridges Having Composite

Slabs, Journal of Bridge Engineer
Decks with Precast Concrete
ing, September/October
2001
51
Bibliography
[Shaw
and
Whiteman]
S.
Shaw
and
J.
R.
Whiteman,
Some
Partial
Differen
Arising in Viscoelastic1997, URL:

ity, Brunei University, Uxbridge, England,
www.hindawi.dk/books/9775945054/B9775945054000125.pdf
tial
[Stroustrup]
B.
Volterra
Stroustrup,
Equation
The C++
Addison-Wesley,
[Wittmann]
52
F. H.
Problems
Programming Language,
Third
Edition,
1998
Wittmann, Werkstoffe
im
Bauwesen,
ETH
Zrich,
1989
How to define
A.
numerical model
for the program BRIDE

In the former
it has often been mentioned that the numerical models in the
chapters
program BRIDE consist of
but it
was
not
explained
chronological list of model objects specified by the user,

user is supposed to specify them. This appendix closes
how the
this gap.
Every numerical simulation program needs an input produced by the user for de
scribing the model to be analyzed. Depending on the program, this may be specified
interactively through a graphical user interface or in written form as a text file using a
problem-specific input language. Both alternatives have advantages and disadvantages.
A
a
graphical input
is
more
intuitive for the
input which allows the experienced
text
There
also solutions which
are
solutions offer both
user
are
inexperienced
user
user
graphical
re-editing
of
some
For the program BRIDE the text

parency. In
addition,
like constructs
by
text
means
of
defining
construction
input
of
graphical
the main program.
Even if
remains accessible for correctness
input solution has been chosen because of
its trans
input allows the implementation of programming-languagea
preprocessor
it. These
defined in the input file
text
by
mostly
Such
details.
In the program BRIDE to every model
the input file
and text input.
friendliness and transparency and consist
inputs graphically, the generated
checks and for the
but less transparent than
to have full control of the model.
combination of
preprocessor which generates the text input to be read
the
user
can
stage. This has
are
section
object
there
A.2).
corresponds
text instruction in
called model instructions. The type of model

be
free, provided they

important advantages:
alternately
two
(see
all
belong
objects
to the
same
belonging together (for example by

putting a new node used in a new element right before the new element) so that it
is not necessary to search the whole input file for already defined information while
writing it. This principle has been borrowed from modern programming languages
like C++ or Java, which for the same reason allow one to declare a variable where
it is needed in the code and not, like in older programming languages, at the
beginning of a block.
It allows to group
It allows
one
together
to write
an
the information
input file which
contains the information about the
con
chronology just by inserting the different model objects in the input file
in the same sequence in which their physical counterpart are built in the real world.
Such an input file appears like a construction schedule, which is the clearest and
simplest way to specify the complex chronology of the construction process.
struction
53
A. How to dehne
A finite element model needs the
1.
Nodes: each node is defined

node six
2.
by
identifier and
an
displacement parameters
by
are
introduced.
by
an
Elements: each element is defined

section and
information:
following
its
spacial
identifier, by
its
coordinates. In each
physical
attributes
(like
by
material)
("incidence nodes").
and
the identifiers of the two nodes to which the element
is connected
3.
Supports: supports
4.
Loads:
typically
similar to
elements,
but
act
elements
or
are
loads
can
on
2.
An identifier
one
incidence node.
nodes.
this schema:
following
uniquely defining the object consisting of letters, digits and the un

derscore character (e.g. deck-node-1, eastsupport ,...). The first character must
be a letter. The identifier is case- sensitive and may not conflict with the keywords
of the text preprocessor
(see
3.
The information
4.
A semicolon ends the input
building
Blanks and end of lines

In the
In this
project
Internally
object1.
object.
treated
separators.
as
explanation
parts of the railroad
A.2).
of the syntax for the model instructions
diagrams2
which
are
not
The
only
acceleration of 9.81
constant used in the program with
^ acting
been taken into account
user
self-explanatory.
diagrams:
further formalism has been introduced for the railroad
the program does not transform the values found in the
their units.
The
are
section
the content of the
sub-sections the
following
is limited to those
on
just
object (e.g Node, Support, ...).
the type of the
keyword uniquely defining
have
directly
The syntax for the model instructions is constituted

1.
they
in the
during
negative global Z-direction. The
program
is assumed to know the
input file
predefined physical
development
is the
in order to
change
units is the
gravity
set of metric units which has
following:
length
meter
force
kilo newton
kN
mass
ton
stress
kilo Pascal
kPa
physical meaning of the values

accordingly whenever needed.
used to graphically represent a syntax and
he defines in the
input file, and
thus to transform them

2
consist of a line (the track) in

diagrams are
terrupted by stations representing the so-called "tokens" and the switches representing alternatives
routes. To define an object it is necessary to read its diagram from the upper left corner, to copy
to the input file the tokens encountered and to insert the numeric values required along the track.
When a switch is encountered, the user can choose one direction to continue following the diagram.
Railroad
54
A.l.
are
objects
keywords with a predefined semantic meaning (they do not need

completely: one or more characters from the keyword's beginning
long as they unequivocally define it),
enough
as
italic is used for user-defined tokens

or
for the dehnition of model
bold is used for

to be written
Syntax
floating-point notation),
(in
most
the italic text
boxes refer to other railroad
cases
explains
identifiers
or
numbers in
which information is
diagrams which, ideally,
integer
required,
have to be inserted at the
position of the box.
A.l.
Syntax
A. 1.1.
objects
Stage object
In the
chronological objects list the stage object signalizes when a new stage has been
reached and gives it an identifier and a day number. The day number is necessary for
the calculation of time-dependent effects and for the merging of many object lists using
the Restart object (see subsection A. 1.2).
The syntax for
stage object is:
stage_identifier DayNumber
Stage
The
day
(unless
number must be positive and
it is the first
construction
A. 1.2.
day_number
stage coming after
bigger
a
than the
restart
day
object).
It
number of the former stage
specifies
the
day
when the
stage is reached.
Restart
object
The Restart
object allows one to merge many objects lists and belongs at the beginning
of a chronological objects list, being appended to an already defined one. This should
be used for structures consisting of different sub-structures which stand alone before
being joined together at the end of construction (typically the cantilever portions of a
cable-stayed bridge attached to different pylons). It allows one to keep the input files
for the sub-structures in different smaller files first
as long the sub-structures are not
joined together and then to easily merge them.
-
The syntax for

Restart
The
file
Restart
is:
merging of different input files
containing just the
serted
object
Restart
is best done
object(s),
to
specify
the
an
additional short input
the input files for each sub-structure
the preprocessor statement Include
using
objects needed
by writing
(see
subsection
A.2.6)
being
in
and the additional
junction elements between the sub-structures.
55
A. How to dehne
A.1.3.
Remove
The Remove
object
allows
object
to detach
one
single
model
moval from the real structure of temporary members

This is its
Remove
^
Element
element_identifier
Support
support_identifier
NodeLoad
to simulate the
(e.g. formwork), supports
or
re
loads.
node_load_identifier
element
load
identifier
Nodes
Each node is defined
Node
as
jomt_identifier
^-Element Load
A.1.4.
so
self-explanatory syntax:
Joint
*-
objects
node
by
identifier
an
identifier and its
coordinate
The three numeric values
spacial
coordinate
coordinates.
coordinate
Syntax
node
object:
X-Coordmate, Y-Coordmate, and
absolute coordinates of the node's location in the
for
Z-Coordmate
program's arbitrary
are
the
Cartesian
system with the Z axis pointing upward (see figure A.3).
rm.
s less s Usi
I
p
*y
"
srPf |t
fisil illsif mt
|
I
mm a sal mi i wm nana it
****
"%mTk III
ilesfiifi tail*
tm0Mesmu
I
Up
1?
I;;;;:;:;.;;
Il SB
f
KU-
**
j*iii*il i***
Wh*
**"_L
m
*
*"
*
lr
-^
if
"
vMm
ntii isrinipi.1
Figure
gai ni
iii itt* a alio mm m
A.L: A
snapshot
bridge
56
*mi itow*
**** m$
tm
of the program BRIDE
model with their identifiers.
showing
the nodes of
cable-stayed
A.l.
A.1.5.
Syntax
Elements
Elements
three dimensional
are
ement is defined
by
an
straight bar elements connecting two nodes. Each el

identifier, its physical attributes and the identifiers of the two
nodes to which the element is connected. Possible element types
Truss elements:
bending
transmit
same as
the program.
Beam elements:
a
only
normal force
along
are:
the element's axis, with
no
stiffness
Cable elements:
by
objects
normal force.
They
truss
take
but
elements,
tension-only
they
can
be
automatically
re-dimensioned
response and sag into account.
bending and shear forces in two directions, torsion and

They correspond to the classical model for slender beams ignoring
transmit
shear deformations.
Cast beam elements:
beam elements but allow the consideration of locked-
same as
in deformations of cast concrete.
Composite
elements:
same
as
cast beam
the bond between the concrete and the

The
following syntax
-Truss
Cable
Beam
-Nodes
element
is used to define
elements, but
steel profile.
truss,
cable,
or a
allow the consideration of
beam element:
identifier
node
identifier-*
77-node
Êxcentricityty
^ Section
identifier
noae_iaenLirier
YOrientation
The two node identifiers define the nodal incidences.
The three numeric values after the
or
optional "Excentricity" keyword
vector
starting
from the incidence node and
end- point
defining
an
dX
dY
dZ-
Z-
start-
Êxcentricity
identifier
section
ponents of
-y-
dXdYdZ-^
dY dZJ
dX
excentric, rigid
pointing
are
the
com
to the element
connection of the element relative
to the incidence node.
The section identifier after the "Section"

which
are
assumed to be constant
The
optional three
prescribe explicitly
numeric values
is
always
user
on
over
keyword defines the section properties

length of the prismatic element.
the
numeric values after the "YOrientation"
allow
one
to
the orientation of the element's local coordinates system. Their
give the direction of
vector
co-planar to the local x axis,

plane (see figure A.3).
the element's axis, and the local x-y
does not
keyword
prescribe
which
If the
the orientation of the element's local coordinates system
57
A. How to dehne
explicitly the program assumes that

plane and, if the element is vertical,
The
following syntax is used

represented in figure 2.5:
CompositeBeam
(formwork-beam-identifier)
cast/composite beams,
The section
Dz
Section
Dz
coupled
of
profile
section_identifier
to
defined beam element
previously
and
section.
composite
are
do not have to be
so
beam in the local
the
The incidence nodes

for both formwork
same
specified.
specified
coordinate.
cast/composite
is that of the
fact that the section of composite beams is

with their formwork section into
cast beam is
using
defined,
(the
section
section
single
specified
no
A.1.6.
combined section.
If
composite beam
single
This allows
steel,
defined,
composite beam)
excentric
its cast concrete
is
is removed from the model after the activation
one
to take into account the shear connection
always
assumed to be
rigid.
for this
the section definition is
Sections
Several elements often have the
same
section,
so
reason
decoupled from the element definition. A complete section specification

implemented element types even if it makes little sense to define e.g.
the
bending
rigid
merged together
carrying the identifier of the
combined section
which is
is
kept separated
are
shear connection between them but
composite beam, while the formwork beam

between cast concrete and
alone, regardless of the

automatically merged by the program
in the definition of the
with the formwork section in
of the composite beam.
beam
the cast concrete and formwork sections
two different beams with
connection of their incidence nodes.
stiffness for
cable
The syntax to define
is
which is intended to model the structural behaviour of
of the steel
or
cast/composite
Y axis.
Dz is the distance between the axis of the formwork beam and the axis of the
If
global
X-Y
composite beam, whose idealization
or a
and the orientation of the local coordinates system

and
to the
global
Cast and composite beams must be

the formwork
parallel
to the
parallel
-n
-^
^FormworkBeam formwork_beam_identifier
that it is
cast
1er
j- element_identif
vCastBeam
to define
the local y axis is
Section
an
as
the program would
ignore
is needed for all

a
non-vanishing
it in the calculation.
element section is:
section_identifier
SectionGroup
section_group_identifier
A A
It
It
Iy Iy
Iz
Iz~^.
(
v
yMax y_max yMin y_mm
The
z_max
section-group-identifier
section group.
58
zMax
zMin
z_mm
Material
material_identifier
is used to define sets of sections
This is needed when
re-dimensioning
belonging
the cable sections
as
to the
same
the program
A.l.
Syntax
objects
BPfiVP
[ *,*,*,,-,,...,,
Figure
A.2.: A
*-
snapshot
bridge
**"* * ^FJtW
lllf-
n>
t *
1*1
m,
showing
ijt
>
the elements of
the
same
group
and minima
are
as
the section of the cable
needed for the volumetric
element representation and to find the normal stress distribution
As many sections
^
^
along
the local z-axis.
Materials
usually
have the
same
material,
from the section definition. The syntax to define
cable-stayed
model.
only searches among the sections belonging to

being substituted. The local y- and z-maxima
A.1.7.
Material
material_identifier
MaxCreepFactor
the material definition

a
SigMax sigma_max
max_creep_factor DaysForHalfCreep
MaxShrinkageS train
max_shrinkage_stram
was
separated
material is:
G Rho
Rho~*
days_for_half_creep
DaysForHalf Shrinkage
-\
days_for_half_shrmkage
E is the
Young's modulus, SigMax is the value of the maximum tension which is

re-dimensioning, G is the shear modulus used to determine the torsional
stiffness, Rho is the density (mass per unit volume). The development of both creep and
shrinkage over time, being approximated by a hyperbolic tangent, have to be specified
with two values each: the maximum asymptotic value for t
00 and the number of days
needed to reach half the maximum asymptotic value (see the chapters 5 and 6).
used for cable
59
A. How to dehne
^,
vm
Is
**p1i ilK
hiss wem
|
I
JBi
"
% s ^ftf#
of
.,,
"*
l"
Tshii.Tsa.li. ilsrafe
r^^"
''
liW*!
&
-i-
t j -m s i i
ibiIhiI
Jllaii
iiitii i*iiiieiMil lesfl
Ste
ilTHifii-i tail
s-
I* m m
s as
Jt
Mi lai
#*.1,,.
iI(i*|
""M' !% I# #!#!#!
! l
rj
lU.pta.muilp
i
*
IsreaJtiiiHk
ftffiflil ll!JI#s
liliailfll Ifiifpf
poire
>
Ti%
||(HHBIJS<
I
\mmmi^mn^$%
f*l*
III III
.'. L,,,,,,,,
* m S3 fSSiIiMlllHl
H# hlînif
US
toiii islftal A
*#
1S* ?
;^J^^
-'
f#li
tf
\
\
BiSr <tfe*lLiMnt
Nato
HiJlI !:(:(:]( | $$$ tt
i^d
^rf
su
f !lll| ffll' Pft Might
Figure A.3.:
relies* sif
III f miif $*4* 1
^ /S
fill
ii
BI
SS
*f
snapshot
tilever
si H*
(represented by
dashed
showing
line)
an
element
modelling
can
with its support, its local coordinates
system and the global coordinates system.
A.1.8.
Joints
are
(element
to
an
Joints
zero-length
elements used to connect two nodes which
incidence nodes in the
element of
zero
length
same
The syntax for the definition of

-Joint
joint_identifier
Sti ffness
tx
ty
tz
rx
Nodes
ry
location
and infinite
a
are
are
in the
not allowed because
same
they
location
would lead
stiffness).
is:
joint
node_identifier node_identifier 3
rz-
XOrientation
Y Z YOrientation
ZJ
\~
xNoTension
-yNoTension
^
The two node identifiers define the
tx, ty, and
tz
are
rx, ry, and
rz are
nodes, expressed
springs connecting the rotational degrees of
in local coordinates.
An orientation of the local coordinates system

to be defined
60
springs connecting the translational degrees

nodes, expressed in local coordinates.
the stiffnesses of the
freedom of the two
incidences.
of freedom of the two
joint's
zNoTension
explicitly.
differing
from the
global one has

keyword YOrientation
A.l.
give the direction of
beforehand,
The
vector
and the local x-y
Syntax
co-planar
to the local
axis, defined immediately
and zNoTension allow
of the three local
one
directions,
one
to define
in which the nodes
This represents another non-linear effect the program
separate freely.
objects
plane.
keywords xNoTension, yNoTension,
contact connection in
can
can
deal
with.
HPF
Aie #!*%*!# 1
f 4>
i#
# A
li
-#,
ltiiii
mnwiittimm
**l
fiiHliiBitaiitnirtlrnu
tail gl ferps-silflfirfi
Tlijpljiii
fat$
Utii! #sriji*fii|
Figure
%M rttt
*S #r**f itl vigM
A.4.: A
lri
snapshot
bridge
|slagi*r
jb*4$
aaSas
ciftiH teste
*rt
showing
the
joints of
model with their identifiers and the elements in
cable-stayed
wire-frame repre
sentation.
A.1.9.
Supports
Supports
are
similar to
nection between this

static model because
joints but with
single
they anchor
The syntax for the definition of

-Support
support_identifier
Stiffness
tx
ty
tz
single
incidence node:
they represent a con

ground. Supports are necessary in a
making it stable.
incidence node and the
rx
ry
Node
it in space
a
support is:
node_identifier
rz-s
XOrientation
Y Z YOrientation
ZJ
\~
xNoTension
-yNoTension
zNoTension
61
A. How to dehne
The
node identifier defines the
single
tx, ty, and
tz
rx, ry, and
are
freedom of the
springs connecting the rotational degrees of

freedom of the single incidence node to the ground, expressed in local coordinates.
An orientation of the local coordinates system
give
beforehand, and the local
The
in
support
from the
to the local
and zNoTension allow
to define
one
of the three local directions in which the node
one
This represents
ground.
from the
one
can
lift off
of the non-linear effects the program BRIDE
deal with.
i IS1
toUWH
ft
i.
(\
/
f
x-y
co-planar
plane.
keywords xNoTension, yNoTension,
contact
can
vector
global one has

the keyword YOrientation
axis, defined immediately
differing
The three numeric values after
explicitly.
the direction of
rtid
springs connecting the translational degrees of

ground, expressed in local coordinates.
incidence node to the
single
rz are
to be defined
incidence node.
support's single
MM
'
\\
\
'
\
",
<
afsn osi.il il la if 1.1 laifi

lii iisi-luj wmmmm
*wi%. i##i#y lllp'*t!
:V
*,
'-"-'*""
X\,.
\
V
1\\
^x
"
"
XXX
mmmtmdiaiilat Israelii I *|Hk
li
aillll lai.il.
lB*ifl*THBll
Jr-^-.-rri,
J^,
ili i alafiil agitai
//
nisi ii-aftfiip mi hh
i $ftu mwt *ifl #Pii i*iii*i
Figure A.5.:
iiBPi $m m i
snapshot
bridge
showing
the supports of
model with their identifiers and the elements in
cable-stayed
wire-frame repre
sentation.
A.1.10.
Load
In structural
a
cases
engineering

it is usual to
investigate all possible hazards which could involve
structure and to simulate them with sets of forces
62
acting
on
the structure called load
A.l.
Loading
cases.
element
or
have to be
Load
forces
are
defined
any other
as
node. To attribute load forces to
appended
cases
be
can
at the end of the
multiplied by
Syntax
object.
loading
affect
case
just
one
identifiers
subsections A. 1.12 and A. 1.13).
different factors and
rules of the national codes in force.
can
load cases, load
one or more
object (see
force
objects
Those load
superposed according
superpositions
case
are
with the
called load
combinations.
This is the syntax for the definition of load combinations:

7Lioaauase
Any
number of load
load factor and
A. 1.11.
cases
given
ioaa_case_iaentirier
in the load combination
are
ractor\;
multiplied by
the
adjacent
superposed.
The
load
load_history
case
self-equilibrating forces f simulating locked-in displacements, creep and

shrinkage are automatically applied to the model during stage-by-stage iteration, other
user defined loads such as the self-weight or conditional loads (see chapter 7) are not.
To state that a user defined load should be taken into account during the stage-by-stage
iteration it has to be assigned to the load-history load case (see subsections A. 1.12 and
While the initial
A.l.13).
A.1.12.
Node loads
The syntax for
NodeLoad
node load instruction is:
node_load_identifier
Node
node_identifier -y-Force
^Mass
LoadCases
Zt-^
I
jy- ioaa_nistory
case
jyy-|
identifier
condition
\r
-^
The node identifier defines the node in which the force acts.
on
condition
^ load
mass'
keyword
Force define the vector of
force
acting
the node.
The numeric value
mass
defines
is used for the modal
concentrated
mass
attached to the node which
analysis,
analysis,
buckling analysis: when
a static or a buckling analysis is performed the concentrated mass is automatically
multiplied by the acceleration due to gravity of 9.81 ^ resulting in a load in the
negative Z-direction. Rotational nodal masses cannot be taken into account.
A conditional node load

If
the static
can
be defined
by adding
and the
condition
condition is defined the node load has also to be
(see
assigned
subsection A. 1.14).
to the load
case
load-history.
63
A. How to dehne
A.1.13.
Element loads
The syntax for the
Element Load
element load instruction is:
plane
Element
element_load_identifier
^^SelfWeight
WeightFromSection
vDistributedMass
Force
section_identifier
InitialForce
vy
vz
tx
ty
mx
my
tz
rx
mz
7v^-1
condition
condition
7y-ioaa_nistory
case
rz
ry
7^load_history
^ load
7yy-|
identifier
The element identifier defines the element
The element forces defined
The distributed
which the force acts.
The
in the -Z
self-weight
automatically multiplied by the acceleration due to gravity

direction when a static or a buckling analysis is performed.
mass
is
element force does not need any further
ment section and material
on
by selecting the keywords S elf Weight, WeightFrom

Section, and DistributedMass, are defined by a distributed mass which is used for
modal analysis and for defining the weight for static analysis and buckling analysis.
acting
|\r
-'
-\
distributed_mass
InitialDisplacement
^ LoadCases
element_identifier
The
are
self-weight element force

own
one specified with the
This has been introduced to define the weight of liquid concrete on the
element of a cast or composite beam to take locked-in displacements into
weight-from-section
formwork
because ele
known.
element force is the
but the section used is not the element's

identifier.
specification
same as
the
but another
account.
The numeric value distributedLmass defines the value of the

mass over
the
length
uniformly
distributed
of the element.
The three-dimensional vector after the

tributed force's vector
keyword
(force components
Force defines the
per unit
length) acting
uniformly
on
dis
the element's
axis.
The six values after the
keyword InitialDisplacement
translation and the three rotation components of the
respectively, the three

initial displacement of the
are,
second incidence node in the element's local coordinates system, if it

move
in any direction and if the first incidence node
With the
keyword
InitialForce
an
in the element's incidence nodes
64
initial
can
were
were
free to
group of forces
acting
clamped.
self-equilibrating
be defined. The six numeric values after the
A.l.
Syntax
objects
keyword InitialForce are, respectively, the normal force, the shear forces in local y
and z directions, the torsion and the bending moments in local y and z directions
acting on the first incidence node. The normal forces and moments acting on the
second incidence node are the same as the specified ones but with opposite sign.
If shear forces are defined (i.e. they are not equal to zero) the related bending
moments acting on the second incidence node are changed according with the well
known equilibrium relations between bending moments and shear forces.
The element forces defined
tialDisplacement
tion A.
load
IKoil
*JM
only
by selecting
1.14).
case
If
the
keywords Force, InitialForce

buckling analysis.
and Ini
used for static and
A conditional element load
are
be defined
can
by adding
condition
condition is defined the element load has to be
(see
subsec
assigned
to the
load-history.
ilf il IJ alillsis III i.
afiliil Si irt H ils Sal ilssifs

isisl aiitsst SalfiHSil
liiwtif#i mMm
lraii,,htt*f
iitaHiiiJ'iiAg
i*MirtiaMsiBigin
Figure A.6.:
|i* peitf
snapshot
m$m
stayed bridge
A.1.14.
mm m
showing
element loads
acting
on a
cable-
model.
Conditions
The syntax for conditions
prescribing
displacement
or a
force is:
65
A. How to dehne
| condition|
-Translation
rransis
-y-Global
-y-
^ Rotaticion
^-vMoment r-Local
^
Force
Value
y zÊlement
element
identifier
Normed
normed
x-
'
z-
value
Portion
val
port ion
Of
ue
element
identifi
Normed
er
normed
,J)
(
^
StageForCondition
stage_identifier
The
arbitrary intensity load is defined in the load object the condition has been
assigned to and the activation stage is implicitly defined by the position of the
load object in the chronological objects list. The fixed load is defined by all load
objects belonging to the load case load-history active in the condition stage.
The first line of the rail-road
which
or a
the
be
translation,
coordinate
allow
one
(i.e. going
or
parameter
in another
values for
an
one
relatively (as
absolutely
in the program
to define the condition
combination of inner
from 0 at the element's
If it is
beginning
input reflects the
absolute
displacement
subsequent
StageForCondition
structure).
service
are
parameter.
can
one
as
are
in most
also fulfilled
are
The syntax used to define
only
expressed
same
kind of
set to 0.
the activation stage
its identifier has to be
computed during
cases
the
or
subsequent
given after the keyword
same as
the activation stage.
stage-by-stage iteration, hence

case load-history
conditional loads must be attributed to the standard load

their conditions
be
geometry of the model defined
otherwise it is assumed to be the
Conditional factors
to 1 at the element's
geometry of the cable-stayed bridge, the
condition
same
If the
moments
normed value for
portion of the value assumed by the
point of the
parameter
bending
to define the condition value which
The condition stage may be the

one.
rotation,
one
to define the location of the condition
The second line allows
allows
combination of section forces. The element identifier and
end)
can
diagram
for this load
the
and
case.
"same factor as" condition:
condition
SameFactorAs
conditional
load
The condition is defined

have the
load.
66
same
conditional
identifier
by specifying the identifier of the conditional

factor, the condition stage is retrieved from
load which must

that conditional
A.2.
A.1.15.
Cable
dimensioning object
The syntax to define
cable
RedimensionCables
is:
y.
load_combmation_identifier
consists of
object
dimensioning object
jcable_identifier
^j LoadCombination
The
Text-preprocessor
stage_identifier
Stage
list of cables to be dimensioned and
r;
stage/load-
list of
combination pairs which the cables have to resist.
A.1.16.
Notes
Notes
be inserted at any point of the
can
Note
note
such notes
text
A.2.
list
using this syntax:
-unlikely
appear in the
objects
to preprocessor comments
input files stored by the
(see
subsection
A.2.7)-
are
stored and
program.
Text-preprocessor
powerful text-preprocessor developed by Dr. Guido Sartoris has been integrated in

in the input file
of parameters,
checking system. This allows the use
for-loops, if-tests, mathematical expressions, and standard mathematical functions.
the syntax
The text-preprocessor has its

which
disseminated all
are
syntax checking system, the
own
along
keywords
and constructs for text
the input file.
When
they
are
manipulations
found
by
BRIDE's
preprocessor evaluates them on-line and substitutes them
with the result of their evaluation.

The text-preprocessor features combined with the loose sequence of input
allow
loop's
control variable
can
and the
be used for both the definition of the numeric values
through
e.g.
objects
can
mathematical expressions and the definition of the

preprocessor variable value in its
A.2.1.
identifier which may contain
object
string.
are
similar to the variables of
be used in innumerable ways, e.g.
variable in
3the
generated
in
Variables
The preprocessor variables

can
be
objects
for-loop
powerful input generation,
mam
loops
for the
differences
after
being
more
than
are
declared
once
as
programming language3 and
parameters in expressions
or
generation of coordinates and identifiers. Their
that the preprocessor variables have just
they remain visible

there is always only one
to the end of the
floating point
input and that
instance associated to
variable
even
as
name
control
gives
number type, that
if
they
are
declared
name
67
A. How to dehne
A. 7.: Two different models both obtained with the
Figure
variable "number_of_masts" first to 2
meaning
to the value
they
store
and, depending
(left)
on
same
input file by setting
and then to 6
its
(right).
how the input file has been written,
can change the whole input file (see figure A. 7).

be declared and defined in two ways, either with the following
their initialization
changing
Variables
can
object-like
definition:
Define
or
as
variajble_name
variajble_value
by using and initializing the new variable directly where it is needed with a "=",
following example where the variable Susanna is declared, initialized with the
in the
value 4 and used to express the X coordinate:
Node
mast_l Susanna =400;
Node
mast_2 Susanna 0 3
The former
objects
Node
mast_l 4 0 0
Node
mast_2 4 0 3
Variable
names
would be transformed
(e.g. Susanna)
and the underscore character
A.2.2.
can
as
follows
be constituted
beginning
with
the preprocessor:
by
by
any sequence of
Expressions
Expressions
can
be used at any
place
of the input file in which
numeric value is
The syntax is the standard expression syntax also used in programs such
programs. Parentheses
begin
can
with
e.g.
by just
68
minus
(in
can
such
contain variables which
itself,
letters, digits
letter.
the
one
example
variable
be used to force evaluation

a case
can
the whole expression has to be put in
be defined and initialized
in the former
delivering
the
paragraph (there
coordinate).
required.
spreadsheet
precedences. Expressions may not
previously
or
as
parentheses)
in the
the expressions
are
and
expression
constituted
The operators allowed in expressions
A.2.
Operator
Associativity
Description
To Left
Conditional
?;
Right
||
Left to
&&
Left to
Right
!=
Left to
Right
Left to
Right
AND
Inequality
Right
Equality
<
Left to
Right
Less than
<=
Left to
Right
Right
>
Left to
>=
Left to
Right
Left to
Less than
or
equal
to
Greater than
Greater than
or
equal
Right
Addition
Left to
Right
Subtraction
Left to
Right
Multiplication
Left to
Right
Division
Left to
Right
Modulo Division
Table A.L: The
operator's precedence in increasing

precedence are grouped together.
are
OR
Logical
Logical
Text-preprocessor
order.
to
with the
Operators
same
summarized in the table A.l.
The preprocessor also allows the
use
of the
following
standard mathematical functions
parameter:
needing
exp(...), log(...), sin(...), asin(...), cos(...), acos(...), tan(...),
atari (
) abs (...), int (...), sqrt (...)
one
or
two
parameters:
min(...,...), max(...,...), pow(...,...), atan2(...,...).

The definitions of all these functions
A.2.3.
Brackets
[expression] integer
are
evaluation
A.2.4.
used for
replaces
are
the
same as
in the
programming language C.
evaluation
integer evaluations of expressions, where the rounded result of the
the brackets. This statement is very useful to generate identifiers.
For-loop
The For-loop makes BRIDE's syntax checking system re-read

given number of times. This is its syntax:
given
text
portion
For controLvariable From initiaLvalue To finaLvalue

text
porhon read repeatedly
}
The first time the syntax
the control variable
assumes
control variable is increased
checking system
the initial value.
by
one
(this
reads the text portion in
By
means
any further
reading
curly
brackets
the value of the
that the initial value has to be smaller
69
A. How to dehne
than the final
one).
When the control variable
assumes
the final value the text portion
is read for the last time.
This is
an
example showing
the statement used to generate
some
identifiers and
coor
dinates:
For
To
mast_[i]
From
Node
3*i
The former instructions would be
Node
mast_0 4 0 0
Node
mast_l 4 0 3
Node
mast_2 4 0 6
Node
mast
A.2.5.
3 4
as
follows
by
the preprocessor:
If...Else...
This construct allows
pending
equal zero
expanded
construct
to switch
one
on
and off portions of text inside the input file de
the evaluation of the expression in parentheses: if the expression doesn't

or if it is a true logical expression the first block of text in curly brackets is
on
read otherwise the second
This is the construct's syntax:
(expression) {
If
text read
one.
if
expression
evaluation^ 0
or
true
if
expression evaluations 0
or
false
Else
text read
The statement
is
an
For
example
From
To
==
Else
Node
If
0)
can
also finish
right
before the Else, if
used in combination with
no
alternative is needed. Here
by
the preprocessor:
For-loop:
{
Node
mast_
[i]
deck_0 4 0 3*i
4
3*i
The former instructions would be
70
Node
deck_0 4 0 0
Node
mast_l 4 0 3
expanded
as
follows
A.2.
Node
mast_2 4 0 6;
Node
mast_3 4 0 9;
A.2.6.
Text-preprocessor
Include statement
When the preprocessor encounters an Include statement it replaces it with the content
of the included file. In this example a file named "my_bride_libraries/sections.brd" has
been included in the input file:
Include
my_bride_libraries/sections.brd
A.2.7.
Comments
Portions of the
input file
can
be commented out with
(* at the
beginning
and
*) at
the end of the commentary.
A.2.8.
Finish
The very last word of the outermost input file

to be the
keyword
(the
file not included
by
any
other)
has
Finish.
71
A. How to dehne
72
Graphie
B.
interface of the
user
program BRIDE
The program BRIDE has been conceived
know how such simulations
the
graphic
the
user
user
interface
and the inner
are
as
performed.
(GUI)
workings
of
simulation program for
engineers who
Therefore the
reducing
as
much
of the program. This
goal was set for the design of

possible the distance between
has been attempted by choosing
as
combination of GUI components which
and its state
(see figure
B.l and the
reproduce the inner structure of

program's snapshots in appendix A).
static
analysis panel
modal
BRIDE'S window
analysis panel
buckling analysis panel
tjt,
menu
3D
model state
control panel
the program
3D
tool-bar
3D output
control panel
canvas
text
bar
navigation
panel
Html tool-bar
Figure
B.l.: GUI structure of the program BRIDE.
The program BRIDE appears in

window there is
file appear.
the "file"
window when it is started. At the very top of this
title bar in which the
name
of the program and the
Inside the window below the title bar there is the
menu
with which the input file
automatically generated by
the program
can
can
be
opened
be saved.
stage-by-stage
iteration and the cable
options, the stage and the load
analysis panels
model and the
the
case
Each
right.
analysis's results
on
can
Below the
panel"
re-dimensioning can be
be
set)
menu
on
opened
containing
contains
bar there
(with
are
which
started and the model
and the tabbed

"3D
menu
the left
panel with the three

canvas", in which the
analysis panel
represented three-dimensionally,
are
of the
bar
and closed and the input file
the two main GUI components: the "model state control

the
path
and
"3D output
73
B.
Graphie
control
user
the
user
formatted text
with the
mouse.
tool-bar" with which the
user
containing snapshots of the
In this section
describing
represented
"text
panel"
graphic object
the three-dimensional
canvas
and of the text
the GUI
instinctively
Button:
region of the GUI firing
border and
are
defined.
The
use
of standard GUI
on
same as a
expressed through
Check-box: the
an
event if clicked with the
icon
an
again
darker
same as a
or a
to be deselected.
rendering
region of the GUI showing
in
bar
opening
on
one
if the
delimited
by
selection is
toggle-buttons
region.
the selection state is showed

button
item at
region.
time selected
by
the
user
out of
list of items to be selected when clicked. Menus appear
usually
program's
functionality they
clicks
by
region.
only
mouse
once on
window and carry
(e.g. "file", "edit", "help",.--)
kind of program
mostly
the combo-box
the top of the
in all programs
mouse
In
of the button
toggle-button except that

a darker rendering of the
list of items which appear
Menu: button
operate
string.
check-mark instead of
Combo-box:
to
button except that when it is selected it remains selected
and needs to be clicked

a
user
his experience with other programs.
containing either
the
Toggle-button:
selected
panel.
components is very important for user-friendliness because it allows the

based
canvas
in which
standard GUI components
some
standard GUI components
some
in the 3D
At the very bottom of the window there is the "Html

can save the computation's results in an Html document
3D
Definition of
B.l.
choose what should be
user can
the appropriate check boxes. At the bottom there is
there appears
by
in which the
panel"
by selecting
thus
the
typically
allowing the user
same names
to guess which
contain.
Bar, tool-bar: stripe-shaped region of the GUI containing
more
components with
re
lated functions.
Panel:
rectangular region
of the GUI
containing
one or more
components with related
functions.
Tabbed-panel: GUI component allowing one to show only one panel at a time out of a
group of panels. The presence of the other panels is expressed through selectable
tabs on the side of the panel being shown.
3D
74
canvas:
special panel allowing
the
rendering
of three-dimensional models.
B.2. Menu bar
B.2.
The
Menu bar
(see figure B.2)
keeping
open it
save
bar
menu
an
contains the file
the current output options
input file reflecting the
which allows
one
to open
file,
re
slightly modified),
(i.e. considering the
the input file has been
content of the actual loaded model
cable sections and the
possibly changed
(if
menu
conditional
computed
loads),
close the model
and exit the program.
Figure
B.3.
B.2.: The file
Model state control
menu.
panel
panel (see figure B.3) contains the GUI components needed for
the following computation options: second order, cable sag, tension/compression only
(in elements such as cables, supports and joints) and shrinkage.
The model state control
In the middle
on
the left side
execution of the
vertical
toggle-button
is used to start and
stop the
right are the check-boxes for

the option related only to such an iteration: locked-in displacements, creep, conditional
loads, the slider to set the error feed-back factor a explained in the subsection 7.3.2
1 for the last stage-by-stage iteration as explained in
and the check-box to set a
section 7.3.3. The chart in the white rectangle represents the error in the fulfilment of
the conditions of conditional loads prescribing a given absolute translation and is used
to show the user if convergence is taking place during the stage-by-stage iteration and to
help him operate the slider for the feed-back factor a. A label right above the chart gives
the error in the last iteration and the darker lines in the chart's background represent a
repeated
stage-by-stage
iteration.
On its
power of 10.
The button
user can
section
carrying the label "dimension cables"
read the
areas
weight
in tons per cubic
(the
by
the
combo-box)
user
in the
on
and
one
input file. The
cases
(in
tons if the
in square metres and the material
user can
analysis panels.
to select the combo-box item
are an
above the button
The
density
the lower end contain the identifiers of all construction stages
and of all load
is desired to view in the
allow
expressed
right
dimensioning.
metre).
The two combo-boxes

first
of all cables in the label
of the cables have been
starts the cable
(the
select whatever construction

The buttons
coming before
one) defined
stage/load pair
second
carrying the labels "+" and

or
after the
one
being
"-"
selected
alternative in addition to the standard combo-box selection mode.
75
Graphie
B.
user
toil ortir
j cable
j
j
lai
tension/compreision-only
shrinkage
[3 locked-in displacements
residual addition factor
m
=
|H
.2
.4
.8
croap
H update conditional laais

final rasiiual subtraction
error in
prescribed translations:
5.J0E-4
construction
stage
final
load
case or
load combination
load.history
Figure
B.4.
Elements
B.3.: The model state control
common
Figure
In the program BRIDE
to all three
B.4.: Tabs of the
panel.
analysis panels
tabbed-panel.
specific "analysis panel" with a 3D canvas, a 3D navigation

panel has been dedicated to each kind of analysis
the program BRIDE can perform: static, buckling and modal analysis. These three
panels have been stacked by means of a tabbed-panel GUI component (see figure B.4).
When the user selects a tab with the mouse, the corresponding analysis is performed or,
respectively, the corresponding panel appears on the top of the panel stack.
tool bar and
The 3D
76
3D output control
canvas
dimensional
can
looks the
objects
are
in the three
analysis panels:
Snapshots of the picture
document selecting a button of
same
rendered.
be included in the Html
rectangle
in which three-
rendered in the 3D
the Html tool bar.
canvas
B.4. Elements
The 3D
navigation tool bar
is
in its construction and merits
1^
I^^Bs^^^H
li
\*
LS
common
IS
I Lb
Figure
The
navigation
defined
through
closer examination
1 lib
vl\
only
t/"
B.5.: 3D
at
one
to all three
analysis panels but

(see figure B.5).
\f
'\
analysis panels
is
quite complex
V"
t^mm*.
*js
navigation tool-bar.
points: the "eye" point
the "center" point to which the eye is

be selected
to the three
in the three-dimensional world
two
common
represented
inside the 3D
canvas
is
in which the eye of the viewer is located and
looking
time and define the
at. The first three buttons

mouse
on
the left
behaviour inside the 3D
can
canvas.
If
the first button is
selected, by clicking any three-dimensional object a formatted text

describing
object (which can be either a model object or the representation of
a result) appears in the text panel. If the second button is selected the viewing direction
of the user in the three-dimensional world represented in the 3D canvas can be changed
the clicked
by clicking
which it
was
by clicking
the three-dimensional representation: the point of the representation
on
clicked becomes the
any three-dimensional
view direction is not
allows
one
to
zoom
seventh, eight
changed,
in, the fifth
image.
and to center the
object
to define the eye
to choose
The tenth allows
one
or
one
perspective.
change
speed
toggle-button
The last but
to define if the three-dimensional

a
pop-up
of the eye:
The
viewing direction [-1,-
arrow
image changing
translatory.
out.
main axis
to view the three-dimensional
moving behaviour of the subsequent four
The last button makes
The fourth button
collaterally.
The twelfth button is
which allows
if it is
buttons:
the
viewing direction,
one
button is
toggle
representation should be iso-axial
menu
define the coordinates of the eye and of the center

the
one
is selected
the clicked point but the
to choose the
image. The eleventh allows
if it is not selected the eye movement is
allowing
on
image and the sixth to zoom

viewing direction parallel to a
selected the eye is moved around the center of the

button
picture. If the third button
the center is set
to center the
one
representation from the other side.

one
center of the
i.e. the eye is also moved
and ninth allow
and to center the
1,-1]
new
on
pop-up in which the
manually.
user
The slider allows
can
one
to
if the slider is moved toward the rabbit the eye becomes
faster and if the slider is moved toward the snail the eye becomes slower.
The 3D output definition
analysis panels also have common ele

panels
organized in three columns,
the two first columns contain check boxes and the third column labels (see figure B.6).
The first column, denoted by an icon representing an I-profile, contains the check boxes
which allow one to require the representation of the objects specified in the label on
the same row. The second column, denoted by an icon representing a "T", contains the
ments:
in all 3D
panels
output definition
in the three
the upper side is
check-boxes for the representation of text elements such

and maxima
(for
the
results)
as
identifiers
(for
model
objects)
relative to the
objects specified in the label. On the bottom

of the three columns there are three sliders, the first allowing one to adjust the scale
of some three-dimensional objects not having a constraining physical dimension (nodes,
joints and supports), the second one is used to set the number of points in the element
77
B.
Graphie
user
[i
[j
ii
elements
elements thick.
D D SIJPPr*s
|j |j joints
[j [j
rj rj
q |j
axis
ielei
element loads
q |j initial forcas/displs,
nodal loais
p ij displacements
thick
[j [j displacements
nodes
[j [j iisplacei
[j [j
ri
ri
reactai
il
[j [j
VI
r-j i|
MX
D EI] **^
D [J
[j [j
^z
normal tension
keep actual scale

sealiBi
*
subdivision
lamina
size
Z^LZ
red
Figure
B.6.: The 3D output definition
in which the results have to be
allows
one
B.5.
raw
panel
in the static
analysis panel.
interpolated during post-processing
and the third slider
to set the size of the maxima to which the results have to be scaled.
Specific analysis panel components of
the 3D
output definition panels

B.5.1.
Static
analysis panel
By selecting the "elements" check-box (see figure B.6) the elements are represented as
wire-frames, by selecting the "elements thick" check-box they are represented as boxes
having the width and the height specified in the section object with the maxima and
the minima in the local y and z directions. The difference between "displacements" and
"displacements thick" is the same as between "elements" and "elements thick": selecting
"displacements" check-box the displaced shape of the structure is shown as a wire
frame spline where the elements in traction are coloured in red and those in compression
in blue, while selecting the "displacements thick" check-box the spline is represented by
boxes between the interpolation points having the same cross section as the boxes used
the
for the thick elements.
78
If the "normal tension" check-box is selected the
same
appears
B.5.
with "elements
as
Specihc analysis panel components

but if such
thick",
an
element is
of the 3D output dehnition
selected,
the normal tension in its
representing
interpolation points along

appears in the text
panel. If the toggle-button "keep actual scale" is selected the scale for the amplification
of the results representation is kept unchanged; this is very useful for comparing results.
deck
deck
1 0/3
the local
line of charts
panels
deck
1 1/3
axes
deck
12/3
mm
Mt&
1 15 E3
15/3
/[1 1CE3
l 28E3
^L3A^^.
Figure
B.7.: Text
panel
in the
with the charts of the normal tension
interpolation points
of
composite element
along the local z axis

carrying the identifier
"deck_8_l".
B.5.2.
Buckling analysis panel
g [j element!
[j [j elements thick
rj r-j iupports
rj [j joints
[j [j
aie!
[j [j
nodes
[j |j
element loads
[j [j
initial forees/displs.
q |j nndal loads
q p displacements
thick
rj r-j iiipiaesmanti
ici
D q iiipiacrf
lac Ulli factor; 7. FiE3
Figure
The
Components
panel.
B.8.:
only
factor.
has
no
Modal
buckling analysis
The
not
physical meaning.
analysis panel
self-explanatory components
with the label "+" and

resonance
in the
panel (see figure B.8) is the label showing the buckling

displacements represent the buckling shape and are normed, i.e. their
B.5.3.
only
specification panel
characteristic of this
The
magnitude
of the 3D output
frequency
"-"
in this
panel (see figure B.9)
are
the buttons
to choose the natural oscillation and the label with the
in Hertz.
79
B.
Graphie
user
p*
HD
elements
D D
elements thick
D D supports
D D joints
ax
D D
D D nodes
D
lenient loads
initiai
forces/displs.
noial loads
II D
D D iispiacamant
thick
[j displacements
!
III D displaced
elganmade number
resonance
Figure
B.9.:
Components
panel.
Ifepippfi* ?i.fi*|iMiffiliMTTill
tri me all
tfMjfti
itou** **i*
Figure
B.6.
Text
In the text
f rafla nef; 1.60E-1
of the 3D output
ei
specification panel
in the modal
analysis
mm
B.10.: The Html tool-bar.
panel
there appear
panel
formatted text
description
of the three-dimensional
graphic objects
(see
figure B.7). Snapshots of the text panel content can be added to the Html document
using a button of the Html tool bar. At the beginning of the Html document an index
constituted by links leading to the different snapshots is inserted.
selected with the
B.7.
Html tool-bar
(see figure B.10)
The Html tool-bar

constituted
by snapshots
the first button

text
panel,
canvas
the
on
error
with the
of the 3D
the left it is
third,
allows the
canvas
possible
and the cable
with the last button it
B.8.
can
user
to
generate
and of the text
panel
an
weight
with the second
button,
generated
Html1 document
and to store it. With
to add to the Html document
with the fourth button the
snapshot
snapshot
document
can
of the
of the 3D
be saved and
be reset.
Syntax highlighting
Syntax highlighting
is
very useful feature not
BRIDE but in most text editors used

1The Hyper Text Markup Language
all web
80
and the charts of the normal tensions
mouse
browsers,
but also
by
by
directly implemented
programmers.
To custom
is the format of most Internet pages and
the main word processors.
can
in the program
an
editor for
be read not
only by
B.8.
Syntax highlighting
EBB
File
Edit
Search
Folding
M_arfcers
Utilities
View
Plugins
Macros
P|
QaS
Ik
Define
10P-
Include /home/pi etro/bn de/source/smpl s/hal f_demo, brd
Joint
dl_c2
Nodes
dl_0
c2_0
Stiffness
1E10
1E10
1E10
1E10
1E10
1E10
Joint
cl_d2
Nodes
cl_0
d2_0
Stiffness
1E10
1E10
1E10
1E10
1E10
1E10
Joint
el_e2
Bodes
el_0
e2_0
Stiffness
lElO
1E10
1E10
1E10
1E10
1E10
Remove
Joint
cl_9_tmp
Remove
Joint
dl_9_tmp
;.
Remove
Joint
c2_9_tmp
;.
Remove
Joint
d2_9_tmp
cl_9
Nodes
al_l
cl_9
Stiffness
dl_9
Nodes
bl_l
dl_9
Stiffness
Joint
c2_9
Nodes
a2_l
c2_9
Stiffness
1E10
Joint
d2_9
Nodes
b2_i
d2_9
Stiffness
lEio
ieo
Joint
Joint
Remove
Remove
Sypport
Support
Support
Support
cl_17
d2_17
Con sol
1E10
1E10
1E10
Node
ci_i7
Stiffness
1E30
d2_i7_fnnal
Mode
di_i7
Stiffness
I Error Li
it
ie=o
0;
0
0
o
ieo
o
(dayNunber-dayNunber-t-2)
o
o
o
o
bride
Figure
to enter the
0
0
1E30
4 40
special syntax,
1E10
ci_i7_fnnal
Stag last_tiut_one DayNumber
49,5
juttton between the tw bridge hills
lui
XI
Help
such
as
Syntax highlighting
B.ll.:
150-8859-1
snapshot
of the editor Jedit
syntax of
an
in
some
editor's
(downloadable
[single [ins
17Mb
with the editor Jedit.
the input syntax of the program
syntax keywords
none
BRIDE,
configuration
for free at
file.
usually necessary
Figure B.ll shows a
it is
www.jedit.org) highlighting
the
input file for the program BRIDE.
81
B.
82
Graphie
user
C.
Information
develop
to
the program BRIDE
The program BRIDE has been written

Java and
can
Windows.
that
be
compiled
tools used
technology
using both programming languages C++ and

on the operative systems Linux and
without any modification
Mac OS has not been tested but the code should be
operating system
fully compatible
with
too.
In the program BRIDE the
reading of the input file and all computations are per

formed by a shared object (or a "dynamic link library" in the Windows jargon) written
in C++ (see [Stroustrup]) and compiled with the compiler from the GNU foundation
(see www.gnu.org). While the operating system Linux is the natural one for the GNU
compiler, for the Windows operating system the GNU compiler embedded in the Minimal
(see www.mingw.org) has been used.

The graphic user interface (GUI) has been written in Java using the GUI components
of the Swing package (see [Java Tutorial]). The three-dimensional renderings have been
implemented with the Java 3D library (see [Java 3D]), which is a Java interface to the
GNU Windows environment
3D native libraries
Open
GL and DirectX. The GUI has been
compiled using the Java

started, it loads
compiler by
Microsystems (www.java.sun.com).
object and (according to the wishes of the user) begins sending him queries
and receiving back the information needed to represent the results. The communication
between the GUI and the C++ has been implemented using the Java Native Interface
Sun
When the GUI is
the shared
(JNI,
All
see
[JNI]).
compilers and libraries used

During the development of the
the
can
be downloaded for free
program BRIDE it
object-oriented programming paradigm

source code under control;
a few books on
(see [FAQs, Refactoring, Patterns]).
to
the
keep
was
the
subject
on
very
the Internet.
important
to stick to
increasing complexity of the

are
listed in the
bibliography
83
C. Information
84
technology
tools used to
develop
the program BRIDE
D.
Input files discussed
D.I.
Main
(*
1*)
(*
2*)
(*
3*)
(*
4*)
(*
5*)
(*
6*)
(*
7*)
(*
8*)
(*
9*)
(*
10*)
(*
11*)
(*
12*)
(*
13*)
(*
14*)
(*
15*)
(*
16*)
(*
17*)
(*
18*)
(*
19*)
(* 20*)
Note
materials
Section
Section
Section
Section
2.2 zMin
zMax
(-1)
yMin
zMax
5.1
(-.8)
yMin
zMax
.3
It
2.53
yMin
4.765
Section
Section
yMin
(-2.75) zMax
14.09
Iy
Iy
yMin (-.75)
zMax
Iz
6.88
.96
It
2.38
1.075 zMin
It
.05
Iz
.03
20.5
1.125
(-0.125)
.012
Iy
Material
It
(-1.075)
(-1.075)
2.03
concrete_45_35 ;
(-0.125)
zMin
zMin
.2
Iy
Material
A
Iz
concrete_45_35 ;
27
Iy
Material
It
.023
Iy
Material
slab_cross_end SectionGroup notACable A 3.225 It

.75
13.38
7.90
Material
notACable
0.125
Iz
It
4.52
slab_cross SectionGroup notACable A

2.75
Iy
(-4.765) zMax 1.075
yMin
concrete_45_35 ;
concrete_45_35 ;
0.125 zMin
slab_long_end SectionGroup
8.38
concrete_45_35 ;
Material
notACable
(-4.765) zMax
Iz
Material
(-2.2)
(-.3)
zMin
24.89
Iy
Material
(-1.9)
zMin
1.9
slab_long SectionGroup
Section
yMax
(-2.5)
zMin
2.5
SectionGroup not_a_cable
4.765
yMax
zMax
cd_9 SectionGroup not_a_cable A

.8
yMax
(-1)
yMin (-1.5)
1_1
1
yMax
SectionGroup not_a_cable
yMin
1.5
yMax
(-2.75)
2.75 zMin
zMax
300
2_2 SectionGroup not_a_cable A 5.72 It
Section
(* 23*)
yMin (-1.5)
ab_l
1
yMax
40
3_3 SectionGroup not_a_cable A 6.60 It

1.5
yMax
=====================================
DaysForHalfShrinkage
0.0005
MaxShrinkageStrain
yMax
sections
DaysForHalfCreep
MaxCreepFactor
yMax
and
concrete_45_35 E 30E6 SigMax 0 G 11E6 Rho 2.5
Material
(* 21*)
(* 22*)
input file
==========
Section
chapter 9
in
18.032
7.89
Iz
155
concrete_45_35
.006
Iz
1.898
concrete_45_35 ;
1.4
Material
Iz
concrete_45_35
1.24
Iy
Iz
.60
concrete_45_35 ;
(* 24*)
(* 25*)
Material
(* 26*)
(* 27*)
(* 28*)
steel_360 E 195E6
MaxCreepFactor
G
0
MaxShrinkageStrain
Section
(* 29*)
yMax
SigMax
DaysForHalfCreep
yMin (-.25)
I_profile SectionGroup
.25
140E3*0.7 Rho
zMax
notACable
.95
0.14
(-.95)
zMin
7.85
It
1E-4
Material
Iy
.067
steel_360
Iz
.001
(* 30*)
(* 31*)
Material
0.6_strands_steel E 195E6 SigMax 1770E3*0.7 G 0 Rho 7.85
(* 32*)
MaxCreepFactor
(* 33*)
MaxShrinkageStrain
(* 34*)
For
(* 35*)
From
yMax
(* 37*)
To
100
DaysForHalfCreep
0
dc
{
cables
sqrt(150E-6*i/3.14) yMin (-de)
Material
[i]_strands SectionGroup
Section
(* 36*)
(* 38*)
0.6_strands_steel
150E-6*i
zMax
dc
It
zMin
Iy
Iz
(-de)
>
(* 39*)
(* 40*)
Note
==========
one
bridge
half
(* 41*)
(* 42*)
Define
85
D.
Input hies discussed
(* 43*)
in
chapter
/home/pietro/bride/source/smpls/half_demo_numbered.brd
Include
(* 44*)
(* 45*)
Note
==========
other
bridge
half
=============================
(* 46*)
(* 47*)
Restart
(* 48*)
Define
(* 49*)
Include
;
s
/home/pietro/bride/source/smpls/half_demo_numbered.brd
(* 50*)
(* 51*)
Note
==========
between the
juction
halfs
bridge
two
===============
(* 52*)
(* 53*)
Joint
(* 54*)
Joint
(* 55*)
Joint
dl_c2 Nodes dl_0 c2_0 Stiffness 1E10 1E10

cl_d2
Nodes
el
Nodes
e2
1E10
1E10
1E10
1E10
cl_0 d2_0
Stiffness
1E10
IEIO
IEIO
IEIO
IEIO
1E10
el
Stiffness
IEIO
IEIO
IEIO
IEIO
IEIO
1E10
e2
(* 56*)
(* 57*)
Remove
Joint
cl_9_tmp
(* 58*)
Remove
Joint
dl_9_tmp
(* 59*)
Remove
Joint
c2_9_tmp
(* 60*)
Remove
Joint
d2_9_tmp
(* 61*)
Joint
cl_9 Nodes al_l
(* 62*)
Joint
dl_9
Nodes
(* 63*)
Joint
c2_9 Nodes a2_l
c2_9 Stiffness 0
IEIO
(* 64*)
Joint
d2
d2
Nodes
cl_9 Stiffness 0
bl_l dl_9
b2
Stiffness
Stiffness
IEIO
0
1E10
1E10
0
0
1E10
1E10
0
0
0
0
0
;
0
0
0
;
(* 65*)
(* 66*)
Remove
Support cl_17
(* 67*)
Remove
Support d2_17
(* 68*)
Support cl_17_final
(* 69*)
Support d2_17_final Node dl_17 Stiffness 0 1E30 1E30 0 0 0
;
;
Node
cl_17 Stiffness 1E30 1E30 1E30 0 0 0
(* 70*)
(* 71*)
Stage last_but_one DayNumber (dayNumber=dayNumber+2)
(* 72*)
(* 73*)
CompositeBeam cl_0_conc
(* 74*)
ElementLoad
FormworkBeam
cl_0 Dz
1.075
cl_0_conc_l Element cl_0_conc Force 0 0
Section
slab_long
(-27.22)
LoadCase
live
(* 75*)
(* 76*)
CompositeBeam dl_0_conc
(* 77*)
ElementLoad dl_0_conc_l Element
FormworkBeam dl_0
Dz
1.075
dl_0_conc Force 0 0
Section
slab_long
(-27.22)
LoadCase
live
(* 78*)
(* 79*)
CompositeBeam el_0_conc
FormworkBeam
(* 80*)
ElementLoad el_0_conc_l
Element
(* 82*)
CompositeBeam c2_0_conc
FormworkBeam
(* 83*)
ElementLoad
el_0 Dz
1.075
el_0_conc Force 0 0
Section
slab_long
(-27.22)
LoadCase
live
(* 81*)
c2_0 Dz
1.075
c2_0_conc_l Element c2_0_conc Force 0 0
Section
slab_long
(-27.22)
LoadCase
live
(* 84*)
(* 85*)
CompositeBeam d2_0_conc
(* 86*)
ElementLoad d2_0_conc_l Element
FormworkBeam d2_0
Dz
1.075
d2_0_conc Force 0 0
Section
slab_long
(-27.22)
LoadCase
live
(* 87*)
(* 88*)
CompositeBeam e2_0_conc
FormworkBeam
(* 89*)
ElementLoad e2_0_conc_l
Element
e2_0 Dz
1.075
e2_0_conc Force 0 0
Section
slab_long
(-27.22)
LoadCase
live
(* 90*)
(* 91*)
Stage
last
DayNumber (dayNumber=dayNumber+l)
(* 92*)
(* 93*)
(* 94*)
86
Note
==========
cable
dimensioning
stuff
==================================
D.I. Main
(* 95*)
history_and_live_l.7
LoadCombination
(* 96*)
LoadCase
input hie
load_history
1.7
LoadCase
live
1.7
(* 97*)
(* 98*)
RedimensionCables
(* 99*)
For
(*100*)
From
For
(*101*)
From
To
To
0
c[a]_[8-i]_cab d[a]_[8-i]_cab c[a]_[10+i]_cab d[a]_[10+i]_cab
(*102*)
(*103*)
(*104*)
LoadCombination
history_and_live_l.7 Stage
last
what is
input file.
(*105*)
(*106*)
Line 1:
Finish
note
describing
coming
next in the
Lines 3-38: definition of materials and sections needed in the model.

Lines 34-38: definition of
to choose from
a
some
cable sections not used
the cables
during
preprocessor's for-loop.
dimensioning.
Lines 42-43: definition of the first half of the
directly
in the model but needed
The sections
are
generated using
bridge.
Line 42: the variable "s" used inside the file "half_demo_numbered.brd" to define the
sign of the
X and Y coordinates is set to 1.
Line 43: the file "half_demo_numbered.brd" listed in section D.2 is included
using the
pre-processor command "Include".
Lines 47-49: definition of the second half of the

Line 47: since the two halves of the
bridge
are
bridge.
erected
is needed before the definition of the second half
simultaneously
of the bridge.
restart
object
Line 48: the variable "s" used inside the file "half_demo_numbered.brd" to define the
sign of the
X and Y coordinates is set to 2.
Line 49: the file "half_demo_numbered.brd" listed in section D.2 is included
using the
pre-processor command "Include".
Lines 53-91: definition of the
objects joining together
joints joining together the deck
Lines 57-64: substitution of
straining
ones
some
the two halves of the
bridge.
in the middle.
temporary joints used during erection with less
con
for the final state.
Lines 57-60: removal of the temporary

Lines 61-64: definition of the final
joints.
joints.
87
D.
in
chapter
Lines 66-69: substitution of
constraining
ones
temporary supports used during erection with less
some
for the final state.
Lines 66-67: removal of the temporary supports.

Lines 68-69: definition of the final supports.
Line 71: stage
object definition,
the variable
is defined inside the included
"dayNumber"
file.
Lines 73-89: definition of the composite beams cast
gether
in the middle of the
over
the steel
profiles joined
to
mid-span.
Lines 95-96: definition of the load combination used to dimension the cables.
Lines 98-104: definition of
be dimensioned
are
dimensioning object. The identifiers

generated with a preprocessor's for-loop.
a
cable
of the cables to
Line 106: "Finish" command needed at the end of every outermost input file
included
D.2.
(*
1*)
(*
2*)
by
any other
Input file
Note
3*)
Define
sf
(*
4*)
Define
dayNumber
(*
5*)
Define
tmpx
(*
6*)
(*
7*)
(*
8*)
(*
9*)
(* 10*)
(* 11*)
(s-l)*2-l
some
in the main
input file
variables
s/2-.5
Note
==========
Node
s[s].foundation
mast
========================
80*sf
Support s[s].foundation
Stiffness
(twice)
definition of
(*
not
input file).
included
==========
(i.e.
1E30
1E30
Node
1E30
s[s].foundation
1E30
1E30
1E30
(* 12*)
(* 13*)
Node
a[s]_l
80*sf
(-9.53)*sf
(* 14*)
Beam
a[s]_l
Nodes
s[s] .foundation Excentricity
(* 15*)
a[s]_l Excentricity
(* 16*)
ElementLoad
(* 17*)
ElementLoad
(* 18*)
LoadCase
(* 19*)
Normed_x 1
(-1.5*sf)
a[s]_l_s
Element
a[s]_l_p
Element
load_history
Value
13.25
(-2.35)
Section
(-10.7)*sf
3_3 YOrientation 0 sf
a[s]_l SelfWeight LoadCase load_history

a[s]_l InitialDisplacement .01 0 0 0 0 0
Translation Global
Element
StageForCondition half[s]_last
a[s]_l
(* 20*)
(* 21*)
Node
(* 22*)
Beam
(* 23*)
b[s]_l
80*sf
b[s]_l
Nodes
13.25
s[s] .foundation Excentricity 0 10.7*sf 0

b[s]_l Excentricity 0 1.5*sf (-2.35) Section 3_3 YOrientation
(* 24*)
ElementLoad
(* 25*)
ElementLoad
(* 26*)
LoadCase
(* 27*)
Value
88
9.53*sf
b[s]_l_s
Element
b[s]_l SelfWeight
b[s]_l_p
Element
b[s]_l InitialDisplacement
load_history
Translation Global
LoadCase
Element
sf
load_history
.01
b[s]_l Normed_x
Input hie included (twice)
D.2.
ab[s]_l
Beam
a[s]_l Excentricity
Nodes
b[s]_l Excentricity
ElementLoad
ElementLoad
LoadCase
(-1.5*sf)
(-2.35)
1.5*sf
Section
ab[s]_l_s
Element
ab[s]_l SelfWeight
ab[s]_l_p
Element
ab[s]_l InitialDisplacement .01 0 0

Global 0 10 Element ab[s]_l
load_history
LoadCase
a[s]_2
80*sf
(-9.53)*sf
Beam
a[s]_2
Nodes
a[s]_l Excentricity
2_2 YOrientation 0
35.3
sf
(-1.5*sf)
(-2.35)
Element
a[s]_2 SelfWeight
a[s]_2_p
Element
a[s]_2 InitialDisplacement
Translation
Global
Node
b[s]_2
80*sf
9.53*sf
Beam
b[s]_2
Nodes
b[s]_l Excentricity
Section
35.3
2_2 YOrientation 0
LoadCase
Element
sf
ElementLoad
b[s]_2_p
Element
b[s]_2 InitialDisplacement
LoadCase
For
From
Element
load_history
ab[s]_2_pl
ElementLoad
Value
a[s]_2 b[s]_2
Nodes
ab[s]_2_s
ElementLoad
LoadCase
Translation
Global
ab[s]_2
Beam
To
Section
LoadCase
Element
load_history
.01
Global
0
1
10
Element
.01
ab[s]_2 Normed_x
{
80*sf
(-9.53)*sf 47+i*.9
a[s]_[3+i]
Nodes
a[s]_[3+i-l]
sf
a[s]_[3+i]
Section
a[s]_[3+i]_s
Element
a[s]_[3+i] SelfWeight
ElementLoad
a[s]_[3+i]_p
Element
a[s]_[3+i]
LoadCase
load_history
Normed_x
Value
Translation Global
Node
b[s]_[3+i]
80*sf
9.53*sf
Beam
b[s]_[3+i]
Nodes
b[s]_[3+i-l] b[s]_[3+i]
sf
47+i*.9
LoadCase
InitialDisplacement
1
Element
1_1
ElementLoad
Section
1_1
Element
b[s]_[3+i]
SelfWeight
b[s]_[3+i]_p
Element
b[s]_[3+i]
InitialDisplacement
LoadCase
load_history
Normed_x
Translation Global
LoadCase
Element
Stage mast[s] DayNumber (dayNumber=dayNumber+l)

deck
b[s]_[3+i]_s
.01
a[s]_[3+i]
ElementLoad
Value
load_history
ElementLoad
==========
b[s]_2 Normed_x
ab[s]_2 InitialDisplacement
Translation
Beam
Note
a[s]_[3+i]
YOrientation 0
a[s]_2 Normed_x
1_1 YOrientation sf
Node
YOrientation 0
0
ab[s]_2 SelfWeight
Element
load_history
.01
b[s]_2 SelfWeight
load_history
load_history
(-2.35) b[s]_2
1.5*sf
Element
a[s]_2
b[s]_2_s
Value
ElementLoad
LoadCase
a[s]_2_s
load_history
;
;
ElementLoad
ElementLoad
Value
load_history
Translation
Node
LoadCase
input hie
(-2.35)
ab_l YOrientation sf
Normed_x 0 Value 0 StageForCondition half[s]_last
Section
in the main
load_history
.01
b[s]_[3+i]
============================
89
D.
in
chapter
(* 80*)
(* 81*)
Node
(* 82*)
Joint
c[s]_9
(-9.53)*sf
80*sf
c[s]_9_tmp
13.25
a[s]_l c[s]_9
Nodes
Stiffness
1E10
IEIO
IEIO
IEIO
IEIO
1E10
Stiffness
IEIO
IEIO
IEIO
IEIO
IEIO
1E10
(* 83*)
(* 84*)
Node
(* 85*)
Joint
d[s]_9
80*sf
d[s]_9_tmp
9.53*sf
13.25
b[s]_l d[s]_9
Nodes
(* 86*)
(* 87*)
Node
e[s]_9
(* 88*)
Beam
ce[s]_9
(* 89*)
ElementLoad
ce[s]_9_s
(* 90*)
ElementLoad
ce[s]_9_cast
(* 91*)
LoadCase
80*sf
13.25
c[s]_9 e[s]_9
Nodes
Element
cd_9 YOrientation sf
ce[s]_9 SelfWeight
Element
load_history
Section
LoadCase
load_history
ce[s]_9 WeightFromSection slab_cross
(* 92*)
(* 93*)
Beam
(* 94*)
ElementLoad
de[s]_9_s
(* 95*)
ElementLoad
de[s]_9_cast
(* 96*)
de[s]_9
LoadCase
e[s]_9 d[s]_9
Nodes
Element
load_history
Section
cd_9 YOrientation sf
de[s]_9 SelfWeight
Element
LoadCase
load_history
de[s]_9 WeightFromSection slab_cross
(* 97*)
(* 98*)
Stage platform[s] DayNumber (dayNumber=dayNumber+l)
(* 99*)
(*100*)
CompositeBeam ce[s]_9_conc
FormworkBeam
ce[s]_9
Dz
1.075
Section
slab_cross
(*101*)
CompositeBeam de[s]_9_conc
FormworkBeam
de[s]_9
Dz
1.075
Section
slab_cross
(*102*)
(*103*)
Stage platform[s]_hard DayNumber (dayNumber=dayNumber+l)
(*104*)
(*105*)
For
From
To
(*106*)
Node
c[s]_[8-i]
(80-(6.5+9.8*i))*sf
(*107*)
Beam
c[s]_[8-i]
Nodes
(*108*)
YOrientation 0
sf
c[s]_[8-i+l]
ElementLoad
c[s]_[8-i]_s
(*110*)
ElementLoad
c[s]_[8-i]_cast
(*112*)
Element
c[s]_[8-i]_p
(*113*)
LoadCase
load_history
(*114*)
Normed_x
Value
c[s]_[8-i]
Element
WeightFromSection slab_long
ElementLoad
13.25
I_profile
LoadCase
Element
SelfWeight
LoadCase
load_history
c[s]_[8-i]
load_history
c[s]_[8-i]
Translation Global
InitialDisplacement
10
Element
Section
(*109*)
(*111*)
(-9.53)*sf
c[s]_[8-i]
.01
c[s]_[8-i]_conc
(*115*)
(*116*)
Node
d[s]_[8-i]
(80-(6.5+9.8*i))*sf
(*117*)
Beam
d[s]_[8-i]
Nodes
(*118*)
YOrientation 0
sf
ElementLoad
d[s]_[8-i]_s
(*120*)
ElementLoad
d[s]_[8-i]_cast
(*122*)
Element
d[s]_[8-i]_p
(*123*)
LoadCase
load_history
(*124*)
Normed_x
Value
d[s]_[8-i]
Element
ElementLoad
13.25
Section
LoadCase
Element
SelfWeight
LoadCase
load_history
d[s]_[8-i]
load_history
d[s]_[8-i]
Translation Global
InitialDisplacement
10
Element
I_profile
(*119*)
(*121*)
9.53*sf
d[s]_[8-i+l] d[s]_[8-i]
.01
d[s]_[8-i]_conc
(*125*)
(*126*)
Node
e[s]_[8-i]
(80-(6.5+9.8*i))*sf
(*127*)
Beam
e[s]_[8-i]
Nodes
e[s]_[8-i+l]
(*128*)
Section
(*129*)
ElementLoad
e[s]_[8-i]_s
(*130*)
ElementLoad
e[s] _[8-i]_cast
(*131*)
90
I_profile
YOrientation 0
Element
13.25
sf
e[s]_[8-i]
Element
e[s]_[8-i]
SelfWeight
e[s]_[8-i]
LoadCase
load_history
LoadCase
load_history
D.2.
Beam
ce[s]_[8-i]
ElementLoad
LoadCase
Beam
ce[s]_[8-i]_s
de[s]_[8-i]
ElementLoad
LoadCase
e[s]_[8-i] d[s]_[8-i]
ElementLoad
LoadCase
c[s]_[8-i]
d[s]_[8-i]_cab
LoadCase
ElementLoad
.01
a[s]_[3+i]
0.770
c[s] _[8-i]_cab
Element
LoadCase
Value
d[s]_[8-i]
Nodes
load_history
c[s] _[8-i]_cab Normed_x
Element
Excentricity
b[s]_[3+i]
0.770
d[s]_[8-i]_cab_s
load_history
InitialDisplacement
0
d[s]_[8-i]_cab SelfWeight
Element
d[s]_[8-i]_cab
Translation Global
c[s] _[8-i]_cab SelfWeight
Element
ElementLoad
Excentricity
c[s]_[8-i]_cab
l_strands
SelfWeight
c[s]_[8-i]_cab_s
Translation Global
Section
I_profile
InitialDisplacement
Cable
de[s]_[8-i]
Element
Nodes
load_history
ElementLoad
Section
load_history
l_strands
I_profile
ce[s]_[8-i] SelfWeight
Element
de[s]_[8-i]_s
c[s]_[8-i]_cab
Section
Section
Nodes
0
e[s]_[8-i]
input hie
load_history
YOrientation sf
Cable
c[s]_[8-i]
Nodes
YOrientation sf
in the main
.01
0
d[s]_[8-i]_cab
Element
0
Element
LoadCase
load_history
d[s]_[8-i]_cab Normed_x
Value
Stage mid_deck[s]_ [i] DayNumber (dayNumber=dayNumber+l)
CompositeBeam c[s]_[8-i]_conc
FormworkBeam
ElementLoad
LoadCase
c[s]_[8-i]
Dz
1.075
c[s]_[8-i]_conc_l
live
Section
Element
slab_long
c[s]_[8-i]_conc
Force
(-27.22)
Force
(-27.22)
Force
(-27.22)
CompositeBeam d[s]_[8-i]_conc
FormworkBeam
ElementLoad
LoadCase
d[s]_[8-i]
Dz
1.075
d[s]_[8-i]_conc_l
live
Section
Element
slab_long
d[s]_[8-i]_conc
CompositeBeam e[s]_[8-i]_conc
FormworkBeam
ElementLoad
LoadCase
e[s]_[8-i]
Dz
1.075
e[s]_[8-i]_conc_l
live
Section
Element
slab_long
e[s]_[8-i]_conc
CompositeBeam ce[s]_[8-i]_conc
FormworkBeam
ce[s]_[8-i]
Dz
1.075
Section
slab_cross
1.075
Section
slab_cross
CompositeBeam de[s]_[8-i]_conc
FormworkBeam
de[s]_[8-i]
Dz
91
D.
in
chapter
Stage mid_deck[s]_[i]_hard DayNumber (dayNumber=dayNumber+l)

If
(10+i<15) {
Node
c[s]_[10+i] tmpx=(80+(6.5+9.8*i))*sf (-9.53)*sf
13.25
>
If
(10+i==15)
Node
c[s]_[10+i] tmpx=(80+(6.5+9.8*i)-l.5)*sf (-9.53)*sf
Support c[s]_[10+i]
c[s]_[10+i]
Node
Stiffness
1E30
13.25
0
;
;
>
If
(10+i==16)
Node
c[s]_[10+i] tmpx=(80+(6.5+9.8*i)-9.8)*sf (-9.53)*sf
Support c[s]_[10+i]
c[s]_[10+i]
Node
Stiffness
1E30
13.25
0
;
;
>
If
(10+i==17)
Node
c[s]_[10+i] tmpx=(80+(6.5+9.8*i)-18.l)*sf (-9.53)*sf
Support c[s]_[10+i]
c[s]_[10+i]
Node
Stiffness
1E30
13.25
0
>
c[s]_[10+i]
Nodes
c[s]_[10+i-l]
YOrientation 0
(-sf)
Beam
LoadCase
load_history
Element
c[s]_[10+i]_p
.01
InitialDisplacement
Translation Global
Normed_x
Node
If
Value
d[s]_[10+i]
(10+i==15)
c[s]_[10+i] SelfWeight
10
c[s]_[10+i]
Element
ElementLoad
LoadCase
load_history
Element
c[s]_[10+i]
LoadCase
load_history
c[s] _[10+i]_conc
Element
9.53*sf
tmpx
I_profile
c[s]_[10+i]_cast
ElementLoad
Section
c[s]_[10+i]_s
ElementLoad
c[s]_[10+i]
13.25
Support d[s]_[10+i]
Node
d[s]_[10+i]
Stiffness
1E30
Node
d[s]_[10+i]
Stiffness
1E30
Node
d[s]_[10+i]
Stiffness
1E30
>
If
(10+i==16)
Support d[s]_[10+i]
>
If
(10+i==17)
Support d[s]_[10+i]
>
d[s]_[10+i]
Nodes
d[s]_[10+i-l] d[s]_[10+i]
YOrientation 0
(-sf)
Beam
d[s]_[10+i]_s
ElementLoad
LoadCase
load_history
Element
Element
d[s]_[10+i]_p
ElementLoad
InitialDisplacement
Translation Global
Normed_x
Node
92
Value
e[s]_[10+i] tmpx
.01
10
d[s]_[10+i]
SelfWeight
d[s]_[10+i]_cast
ElementLoad
Section
d[s]_[10+i]
LoadCase
load_history
Element
d[s]_[10+i]
Element
LoadCase
load_history
d[s]_[10+i]_conc
0
13.25
I_profile
D.2.
e[s]_[10+i]
Nodes
e[s]_ [10+i-l]
YOrientation 0
(-sf)
Beam
LoadCase
load_history
ElementLoad
Element
e[s] _[10+i]_cast
ce[s]_[10+i]
Section
ElementLoad
LoadCase
de[s]_[10+i]
Beam
Section
LoadCase
ElementLoad
Cable
LoadCase
l_strands
ElementLoad
Cable
LoadCase
l_strands
ElementLoad
If
de[s]_[10+i]
Element
c[s]_[10+i] Excentricity
Nodes
Element
load_history
0.770
a[s]_[3+i]
0.770
b[s]_[3+i]
c[s]_[10+i]_cab
d[s]_[10+i]
Nodes
Excentricity
d[s]_[10+i]_cab_s
LoadCase
SelfWeight
d[s]_[10+i]_cab
Section
e[s]_[10+i] d[s]_[10+i]
c[s]_[10+i]_cab_s
SelfWeight
load_history
c[s]_[10+i]_cab
Section
YOrientation sf
de[s]_[10+i]_s
SelfWeight
load_history
ce[s]_[10+i]
Element
load_history
Nodes
I_profile
SelfWeight
e[s]_[10+i]
Element
YOrientation sf
ce[s]_[10+i]_s
SelfWeight
e[s]_[10+i]
c[s]_[10+i] e[s]_[10+i]
Nodes
I_profile
I_profile
Beam
Section
input hie
e[s]_[10+i]_s
ElementLoad
e[s]_[10+i]
in the main
Element
load_history
d[s]_[10+i]_cab
(10+i<15) {
c[s]_[10+i]_cab
ElementLoad
.01
InitialDisplacement
Translation Global
0
1
Element
c[s]_[10+i]_cab
LoadCase
load_history
c[s] _[10+i]_cab
Element
d[s]_[10+i]_cab
ElementLoad
.01
InitialDisplacement
Translation Global
Element
0
1
d[s]_[10+i]_cab
LoadCase
load_history
d[s] _[10+i]_cab
Element
>
If
(10+i==15)
{
c[s]_[10+i]_cab
ElementLoad
InitialDisplacement
Translation Global
.01
10
0
0
Element
c[s]_[10+i]_cab
LoadCase
d[s]_[10+i]_cab
InitialDisplacement
Translation Global
Value
.01
10
Value
Element
load_history
c[s]_[10+i]_cab Normed_x
Element
ElementLoad
d[s]_[10+i]_cab
LoadCase
load_history
d[s] _[10+i]_cab Normed_x
Element
>
If
(10+i==16)
{
c[s]_[10+i]_cab
ElementLoad
InitialDisplacement
Translation Global
Normed_x
Value
.01
10
0
0
Element
c[s]_[10+i]_cab
LoadCase
Element
load_history
c[s] _[10+i]_cab
93
D.
ElementLoad
in
chapter
d[s]_[10+i]_cab
.01
InitialDisplacement
LoadCase
load_history
d[s] _[10+i]_cab
Translation Global
10
Normed_x 1
Value
d[s]_[10+i]_cab
Element
Element
y
if
(10+i==17)
ElementLoad
{
c[s] _[10+i]_cab
.01
InitialDisplacement
c[s] _[10+i]_cab
LoadCase
load_history
c[s] _[10+i]_cab
Translation Global
10
Normed_x 1
ElementLoad
Value
Element
Element
d[s]_[10+i]_cab
.01
InitialDisplacement
d[s]_[10+i]_cab
LoadCase
load_history
d[s]_[10+i]_cab
10
Normed_x 1
Element
Translation Global
Value
Element
Stage side_deck[s]_ [i] DayNumber (dayNumber=dayNumber+l)

If
(10+i<=15)
CompositeBeam c[s]_[10+i]_conc
Section
slab_long
ElementLoad
Force
Force
(-27.22)
LoadCase
Element
live
Force
(-27.22)
LoadCase
Element
live
slab_cross
LoadCase
Else
slab_cross
Dz
1.075
d[s]_[10+i]_conc
Element
live
e[s]_[10+i]
e[s] _[10+i]_conc
FormworkBeam
ce[s]_[10+i]
Dz
1.075
FormworkBeam
de[s]_[10+i]
Dz
1.075
CompositeBeam de[s]_[10+i]_conc
Section
1.075
CompositeBeam ce[s]_[10+i]_conc
Section
d[s]_[10+i]
FormworkBeam
e[s]_[10+i]_conc_l
(-27.22)
Dz
d[s]_[10+i]_conc_l
slab_long
ElementLoad
1.075
c[s] _[10+i]_conc
FormworkBeam
CompositeBeam e[s]_[10+i]_conc
Section
Dz
c[s]_[10+i]_conc_l
slab_long
ElementLoad
c[s]_[10+i]
FormworkBeam
CompositeBeam d[s]_[10+i]_conc
Section
CompositeBeam c[s] _[10+i]_conc

FormworkBeam
ElementLoad
Force
c[s]_[10+i]
Dz
c[s]_[10+i]_conc_l
(-27.22)
LoadCase
.125
Section
Element
slab_long_end
c[s]_[10+i]_conc
live
.125
Section
CompositeBeam d[s]_[10+i]_conc
FormworkBeam
ElementLoad
Force
d[s]_[10+i]
Dz
d[s]_[10+i]_conc_l
(-27.22)
LoadCase
Element
slab_long_end
d[s]_[10+i]_conc
live
.125
Section
CompositeBeam e[s]_[10+i]_conc
FormworkBeam
94
e[s]_[10+i]
Dz
slab_long_end
D.2.
(*340*
ElementLoad
(*341*
Force
e[s]_[10+i]_conc_l
(-27.22)
LoadCase
Element
live
in the main
input hie
e[s]_[10+i]_conc
(*342*
(*343*
CompositeBeam ce[s]_[10+i]_conc
(*344*
Section
FormworkBeam
ce[s]_[10+i]
Dz
.125
FormworkBeam
de[s]_[10+i]
Dz
.125
slab_cross_end ;
(*345*
(*346*
CompositeBeam de[s]_[10+i]_conc
(*347*
Section
slab_cross_end
(*348*
(*349*
(*350*
Stage side_deck[s]_[i]_hard DayNumber (dayNumber=dayNumber+l)
(*351*
(*352*
(*353*
Node
c[s]_0
(*354*
Beam
c[s]_0
Nodes
(*355*
ElementLoad
c[s]_0_s
(*356*
ElementLoad
c[s]_0_cast
(*357*
LoadCase
(-9.53)*sf
13.25
c[s]_0 c[s]_l
Section
I_profile
YOrientation 0
(-sf)
c[s]_0 SelfWeight
load_history
c[s]_0 WeightFromSection slab_long
Element
LoadCase
Element
load_history
(*358*
(*359*
Node
d[s]_0
(*360*
Beam
d[s]_0
Nodes
(*361*
ElementLoad
d[s]_0_s
(*362*
ElementLoad
d[s]_0_cast
(*363*
LoadCase
9.53*sf
13.25
d[s]_0 d[s]_l
Element
Section
YOrientation 0
LoadCase
(-sf)
load_history
d[s]_0 WeightFromSection slab_long
Element
load_history
I_profile
d[s]_0 SelfWeight
(*364*
(*365*
Node
e[s]_0
(*366*
Beam
e[s]_0
Nodes
(*367*
ElementLoad
e[s]_0_s
(*368*
ElementLoad
e[s]_0_cast
(*369*
LoadCase
13.25
e[s]_0 e[s]_l
Element
Section
YOrientation 0
LoadCase
(-sf)
load_history
e[s]_0 WeightFromSection slab_long
Element
load_history
I_profile
e[s]_0 SelfWeight
(*370*
(*371*
Stage half[s]_last DayNumber (dayNumber=dayNumber+l)
Line 3: definition of the variable "sf

in the main
input file listed
"
which
assumes
the value 1 if the variable "s" defined
in section D.l is set to 2 and 1 if "s" is set to 1.
Line 4: definition of the variable
assigned
to the
"dayNumber" needed to generated the day numbers

stage objects. This variable is also initialized using the variable
"s".
Line 5: definition of the variable
"tmpx"
used to generate X coordinates.
Lines 9-77: definition of the mast.

Lines 17-19: first definition of
Lines 59-75:
related
generation with
objects (nodes
Line 77: first definition of
and
conditional load.
pre-processor
for-loop
of
some
mast elements with their
loads).
stage object.
Lines 105-371: definition of the model
objects
related to the deck and the cables.
95
D.
in
chapter

Line 100: first definition of
Lines 105-351:
using
the
junction between the
composite beam element.
for-loop in which the following

loop's control variable "i":
deck segment
the
the cables
its cast concrete slab
the cables
its cast concrete slab.
on
mast and the deck.
mid-span
model
objects
are
generated eight
times
side of the mast,
supporting it,
deck segment
on
(taken
the
supporting
into account with
side-span
it
composite elements),
side of the mast,
and,
mid-span
deck segment with its cables and cast
con
side-span
deck segment with its cables and cast
con
crete.

crete.
Lines 186-188: first of many if-checks needed to take into account the
deck at the supports while
Lines 353-371: Model
cable
96
anchoring
generating the model objects
in the
changes
for-loop.
objects simulating the deck (without concrete)

mid-span.
to the center of the
of the
from the last
Curriculum Vitae
Personal data
first
Pietro
name
name
Pedrozzi
born
November
email
pietro.pedrozzi @ alumni.ethz.ch
21st,
1975
Education
03/1999
06/2004
Dissertation at the Institute of Structural

Federal Institute of
10/1994
09/1990
Engineering
Technology (ETH) Zurich,
Diploma
06/1994
Matura at the Liceo Cantonale di Bellinzona.
engineering
Switzerland.
at the ETH Zurich.
02/1999
in civil
of the Swiss
Professional activities
03/1999
06/2004
Assistant at the Institute of Structural
Engineering
of the ETH
Zurich.
07/1997
08/1997
Internship
at the
civil
engineering
firm Passera & Pedretti in
Lugano.
10/1996
03/1997
Assistant at the Institute for Geotechnical

Zurich.
Engineering
of the ETH

CSB Analysis Tool

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

CSB Analysis Tool

Uploaded by

Copyright:

Available Formats

Dissertation ETH No.

A Software Tool for the Analysis, Design

A dissertation submitted to the

Doctor of Sciences ETH

born 21 November 1975

Prof. Dr. Edoardo

input parameterization. The

then opens the input file from the program

opened while it is being written to check its correctness). All cal

Snapshots with both graphics and text showing model

main scope of this research work

study original approaches

weiten zwischen 200 und 900 Metern.

weltweit sehr beliebt fr

Solche Brcken werden meistens in Frei-Vorbau

und mssen deshalb in allen Bauzustnden unter vielen Lastfllen untersucht

den. Zudem mssen die

ermittelt und eventuell whrend des Baus

Lage der Decke und des Mastes im

die erwnschte verschobene

Anfangsverschiebungen in Verbunddecken bercksichtigt werden.

Bau arbeitet der Benutzer immer

gleichen Eingabedatei und,

entweder manuell oder automatisch.

Schlussfolgerungen dieser Dissertation erwhnt, soll das Programm BRI

Use of standard structural

Technical progress of computers

The program BRIDE and its

User friendliness: BRIDE

Structure of this dissertation

its local stiffness matrix and

of the erection process

Cast beams to simulate locked-in

elastic formwork without shear connection

beams to simulate locked-in

elastic formwork with shear connection

creep in structures with

The excentric local stiffness matrices kca and

cable-stayed bridge design problem

A tour of the program BRIDE

needed to take concrete

The influence matrix

Convergence of the influence matrix algorithm

the conditional loads

file for the program BRIDE

the input file

numerical model for the program BRIDE

for the definition of model

with non-linear models

Second order effects and cable sag

and load combinations

A.1.13. Element loads

interface of the program BRIDE

Model state control

Specific analysis panel components

the program BRIDE

Input file included (twice)

cable-stayed bridge design problem

fixed to the top

from the deck

along a mast protruding

regular distances by cables

normal force introduced

they can be built with very large